This paper should be cited as: Maina, B. 2005. Monitoring and evaluation of support to decentralisation and local governance: Kenya Case Study. (ECDPM Discussion Paper 61). Maastricht: ECDPM.

Contents

Foreword
Acronyms
Executive summary
Chapter 1 Introduction
Chapter 2 Decentralisation in Kenya
Chapter 3 External assistance to decentralisation in Kenya
Chapter 4 Decentralisation, M&E and local accountability in Kenya
Chapter 5 Conclusion: Options for strengthening M&E and local accountability systems
Persons interviewed

Foreword

In the course of the democratisation and institution reform efforts of the 1990s, support to democratic decentralisation and local governance, has become an important area for European development co-operation.

Decentralisation processes in Africa have been attracting support from a variety of European actors in development co-operation. Alongside bilateral aid agencies, the European Community and NGOs, sub-national and local governments in Europe have become increasingly active in supporting democratic decentralisation and experimentation with new approaches to local governance.

In the last years, there has been an increasing interest in the performance of development co-operation. However, as seminar reports and publications illustrate, assessing the effects and impacts of support to decentralisation on development still constitutes a considerable operational challenge for aid managers and their partners.

The ECDPM found it worthwhile to have a closer look at this issue and conducted research on approaches and tools for monitoring and evaluating support to decentralisation.The following paper analyses the present practice in Kenya, a country that has engaged repeatedly in decentralisation efforts and received support from a number of European donors.

The Kenya case study has been prepared by Betty Maina, a Kenyan national and independent consultant with a long experience of work in different projects and programmes supporting decentralisation and local governance in Kenya.

This paper forms part of a series of case studies exploring the approaches and experiences different actors in development have made with monitoring and evaluating development cooperation in the field of decentralisation. With these publications, the ECDPM aims to contribute to stimulate discussion and reflection on present practices and scope for innovation.

The ECDPM is most grateful to the Swedish Ministry of Foreign Affairs, who co-financed this publication. We would also like to thank the all those who have contributed their time and insights during interviews and consultations with the author. Last but not least, we would like to express our gratitude to our former colleague Charlotte Ornemark, who has planned and set up the research in Kenya and freely provided many valuable comments in the course of the process.

Christiane Loquai
Programme Officer

Acronyms

ALGAK Association of Local Government Authorities of Kenya
CBO community-based organisation
CDD community-driven development
CDTF Community Development Trust Fund
DDC District Development Committee
DDO District Development Officer
DFID Department for International Development
DFRD District Focus for Rural Development
DHMB District Health Management Boards
EC European Commission
EU European Union
GIS geographic information systems
GTZ German Agency for Technical Cooperation
HWM Household Welfare Monitoring
IFAD International Fund for Agricultural Development
ITDG Intermediate Technology Development Group
KADU Kenya African Democratic Union
KANU Kenya African National Union
KLRG Kenya Local Government Reform Programme
LA local authority
LASDAP Local Authority Service Delivery Action Plan
LATF Local Authority Transfer Fund
LUO Local Urban Observatory
M&E monitoring and evaluation
MoLG Ministry of Local Government
NGO non-governmental organisations
NUPP Nairobi Urban Poverty Project
PAMNUP Partnership Approach to Meeting the Needs of the Urban Poor
PRCF Poverty Reduction Co-Financing Fund
PROLOGS Poverty Reduction through Optimising Local Governance Systems
SIDA Swedish International Development Agency
UN-HABITAT United Nations Programme on Human Settlements
VDC village development committee

Executive summary

In Kenya, as in much of the developing world, moves towards democratisation and participation have been accompanied by popular demand for decentralisation and attendant power sharing between the centre and sub-units. This recently culminated in proposals for full devolution to the district level and to two other levels - the regions and locations - as set out in the draft constitution now under discussion in Kenya's parliament.

This is not the first time that Kenya faces decisions on decentralisation. The country attained independence with a quasi-federal constitution that devolved power and responsibility to the regions (the present-day provinces). This system did not develop to full maturity, however, as the independence constitution was abrogated within three years of the nation's birth. Civil servants and other central government bureaucrats frustrated the provisions that would have ensured more decision making at the lower levels. However these same civil servants understood the benefit of local decision making and the need for services to respond to local needs. They therefore undertook initiatives to promote local participation in development, alongside their efforts to reduce local political decision making. De-concentration of service delivery was a main means of promoting participation, with systems for public service delivery staffed by central and provincial government servants accountable to line ministries in the capital.

External development partners and the local communities involved with these systems have routinely been frustrated by the unresponsiveness of decisions made to local needs. They call for revision of this approach to reflect more democratic control at the local level. However, these calls have not yet led to the development of alternative frameworks. Instead, in furtherance of the move towards democratisation and calls for political decentralisation, the government, external partners, NGOs and community groups have promoted approaches that would enhance participation and local accountability. These have been undertaken in parallel with reforms intended to revamp the local government system. All of these innovations can be seen as incremental steps towards harmonised decentralisation in Kenya.

Monitoring and evaluation (M&E) is critical for drawing lessons on deepening democracy in Kenya. A review of the approaches used for M&E in these programmes reveals that much information is collected in monitoring exercises and then transmitted to central government offices and to programme sponsors for review. Yet there is little analysis and use of the information at the local level for decision making. Neither do central government civil servants systematically utilise information for learning and design purposes. Rather, they seem to treat such information as a "conditionality" for extending more support. However, there are also examples where communities have acted on the findings of a monitoring exercise or where monitoring has spurred community initiatives, e.g. among local neighbourhood associations. Nonetheless, it is clear that no nationwide process or approach to utilise M&E results "on the ground" has as yet developed.

Decentralisation ought to result in a stronger local government responsive to the needs of an active citizenry. However, since Kenya does not yet have a comprehensive decentralisation plan or policy, decentralisation initiatives in the country vary. Communities have achieved many positive development outcomes, particularly given the weak local government system. But external agencies that support government (mainly through sectoral initiatives) as well as other forms of delivery (e.g. community development through social funds) have inadvertently contributed to further weakening of local government. Often the institutions of community participation that are supported are not rooted in any legal structures and therefore are quite insecure.

For greater benefits and development of democracy, it would be useful if further support to decentralisation in Kenya were anchored in the development of democratic local government. To promote accountability, M&E systems must be designed capable of providing actors at the local level - both ordinary citizens and elected representatives - with feedback for decision making and claim making.

Experience gained from initiatives over the past 30 years provide a rich fabric from which lessons can be derived to build a robust decentralisation framework nested in comprehensive public sector reform in Kenya. The lessons enumerated in this report relate mainly to the following:

• the need for an unambiguous policy framework to support and build on;
• the importance of trial-and-error and learning by doing approaches;
• the need for donors, NGOs and governments to commit to programmes that build habits and institutions that support democratic decentralisation;
• the need to support rather than bypass local governments;
• the need for caution in promoting participation, since local voice may be captured by elites or special interests;
• the need for vigilance in sector-wide programmes to avoid shoring up decision making at the centre at the expense of the locations.

Introduction

1.1 Background

Throughout much of Africa and the ACP there is popular demand for power sharing between the centre and sub-units in the national territorial system. This is reflected in proposals for decentralisation and in the development of sub-national units with a degree of autonomy. The move towards decentralisation is also linked to the current democratisation movement, which is concerned with bringing government closer to the people governed so that they can influence the manner in which decisions affecting them are made.

Democratic decentralisation can be defined as a meaningful authority devolved to local units of governance that are accessible and accountable to the local citizenry, who enjoy full political rights and liberty. It thus differs from the vast majority of earlier efforts at decentralisation … which were largely initiatives in public administration without any serious democratic component (Blair 2000: 21, cited in Johnson 2001: 4).

In the 1990s, Kenya witnessed a growing link between the external support it received and demands for greater democratisation and observance of human rights. Development assistance was extended to civil society organisations and to groups that sought to expand democratic space. In recent years, this has been accompanied by support towards greater local autonomy, decentralised service delivery and promotion of decentralisation in the country. Kenya currently employs several approaches to local service delivery: through the local authority systems, by de-concentration of central government functions to lower levels via line ministries, and through a system that extends direct oversight by the executive (in this case the presidency) in parallel with the supervision/coordination provided by the provincial administration (a carry over from colonial times). However, the government has not yet put in place a legal and institutional framework for harmonised decentralisation. Programmes that have the implicit goal of promoting decentralisation are therefore often stand-alone initiatives, and many are quite recent.

Types of decentralisation Political decentralisation: Giving citizens and elected representatives more power in decision making. Administrative decentralisation: Distribution of authority, responsibility and financial resources for the provision of services among different levels of government. Responsibility for planning and financial management of some government functions is transferred from the central level to the local level. Administrative decentralisation has three main forms, namely de-concentration, delegation and devolution. Fiscal decentralisation: Transfer of authority for taxation and expenditure to sub-national organisations. Different forms include self-financing, co-financing or co-production arrangements, expansion of locally collected taxes and levies, intergovernmental transfers of tax revenues from the central government, and authority for local government to borrow and mobiliseresources through loans and guarantees.

Decentralisation is a complex process that in many respects must be guided by a "learning by doing" approach. There is no clear-cut anticipated path for its evolution or for the responses of the societies and stakeholders involved. Hence there is a need to develop and work with systems that promote learning and can respond quickly. Robust methodologies are called for, both for monitoring and evaluation (M&E) and for the attendant analysis and re-design of programmes to respond to any changes required.

Systematic collection of information, analysis and reporting of results therefore appear critical to the decentralisation process. M&E is needed at the local level to inform residents and encourage public participation. At the central level M&E is needed to monitor and supervise local activities and to provide information for policy development and other institutional responses. M&E is particularly relevant in a context of trial and error and stand-alone initiatives undertaken by disparate actors.

Kenya is in the process of developing a new legal framework for decentralisation. The call for devolution is in fact a key feature of the draft constitution that emerged from the National Constitutional Conference concluded in March 2004. While devolution remains a contentious issue among the main political parties, particularly with regard to the specific responsibilities and revenue control to be assumed by the devolved units, there is no doubt that Kenya will establish a national framework for decentralisation. Experiences from the recent past are therefore quite useful for informing the design and institutional set-up of decentralisation in the country.

A discussion of current approaches to M&E of external assistance to decentralisation and promotion of local accountability is therefore timely. Now, as in the past, Kenya receives support for increased decentralisation. Both government and external donors and their partners within Kenya require an understanding of the impact of their support, to enable them to channel assistance to promote local accountability and learning within the institutions and at the levels that are key to deepening decentralisation.

1.2 Decentralisation, oversight and local accountability

The improvement of local governance in terms of efficiency, effectiveness and accountability is increasingly recognised as a primary vehicle for influencing governance at all levels: central, regional and international. This also extends to civil society and the private sector. Local governance is argued to provide a direct mechanism for people to participate in government, forming a framework that enables the interests of communities to be represented in decision-making structures. Scholars argue that improvement of governance at the local level is the most effective means of building this relationship, enabling human and financial resources to be directly and effectively mobilised in support of improved governance at all levels.

To this end, many governments and cooperation agencies have devoted resources to decentralisation. The anticipation is that decentralisation will lead to improvements in service delivery and local accountability. These improvements are foreseen to emerge from the implementation of programmes in a manner responsive to local needs in terms of relevance, innovation and cost. Such programmes are expected to leverage local knowledge and, through regular elections and other avenues of participation, remain true and responsive to local priorities.

Definition of terms
Different agencies (UNDP, UNICEF, Habitat, SIDA, DFID, etc.) have developed various definitions for terms used in this paper. From a review of their definitions, composite definitions have been developed.

Monitoring. The purpose of monitoring is to systematically assess performance and progress in the implementation of programmes and interventions. Information from monitoring provides the basis for decisions made during the life of a project or intervention. In this case, monitoring supports immediate decision-making needs and becomes a tool for learning.

Monitoring tools in frequent use are regular progress reports (monthly, quarterly, biannual or annual, baseline, end-of-project reports and others), field visits (often to validate reports) and engagement with programme partners (e.g. in focus groups or stakeholder meetings and in participatory M&E).

Evaluation. The purpose of evaluation is to get information that would lead to conclusions on whether objectives have been attained. Evaluation is key to ensuring accountability, as it confirms the credibility of results and validates monitoring reports. Independent actors usually conduct evaluations, which are usually planned for the end of a programme or halfway through it (the mid-term evaluation). Emphasis in evaluation is on reaching judgements regarding performance of programmes (and in laying a relationship with intended outcomes and targets set at the inception of the programme) as well as on programme management and administration.

While monitoring focuses on programme implementation, evaluation focuses on outcomes, results, effects and impacts of interventions. The two together, usually referred to as M&E, provide a critical learning tool for those who design and implement programmes.

Accountability. A dictionary definition of accountable is (adjective) 'responsible to someone or for some action' (Binkerhoff 2001: 2). "Accountability" is the noun form of this adjective. In typical usage accountability refers to answerability; being accountable means having the obligation to answer questions regarding decisions and actions. This has two dimensions: provision of information and provision of explanations and justification for actions (ibid.). Accountability carries the implicit assumption of sanctions that could be imposed by those accounted to, that is, the ability of an overseeing actor to impose punishment on the accountable actor. Such sanctions are specified in codified instruments such as laws and regulations (ibid.).

The location of overseeing actors in a particular governance system is key to accountability. Overseeing actors could be within the state system that is represented by the classical separation of powers and include entities that have responsibility for oversight over other state agencies. They may also be located outside the state but nonetheless play a key role in holding state actors accountable, e.g. through periodic elections, media reports and claim-making forums.

Local accountability. The term "local accountability" has gained currency in recent years along with discussions around where accountability should be located. Institutions and actors in close proximity to the accountable actors are said to be best placed to oversee the accountable actors. Therefore, in the case of public service delivery, local state institutions should exercise oversight over performance and receive responses from those who are answerable (horizontal accountability). Similarly, local consumers of public services should exercise oversight over service delivery and hold state institutions accountable.

Decentralisation and promotion of local accountability
As stated earlier there is popular demand for decentralisation in much of Africa and the ACP. The move towards decentralisation is linked to the democratisation movement, a movement concerned with bringing decision-makers under effective popular control of the people they govern as a way of influencing the manner in which the leaders make decisions affecting the people.

Advocates contend that decentralisation is one way to foster local accountability. Devolution of formal political authority, they maintain, can enhance transparency, responsiveness and accountability at the lower levels of government. These advocates argue that devolved authorities have a better understanding of the views and needs of local people and that periodic elections ensure government responsiveness and accountability. Devolution is said to result in policies, programmes and, more importantly, public expenditure decisions that respond to the needs of the intended beneficiaries.

Nonetheless, several conditions must be met for decentralisation to be effective in fostering government effectiveness and accountability:

• Elected bodies must have adequate funds.
• Elected bodies must have adequate powers.
• There must be reliable mechanisms to ensure accountability of elected representatives to citizens and
of bureaucrats to elected representatives (Manor 1999, cited in Johnson 2001: p. 10).

This implies a system in which elected local representatives have authority to raise and spend resources at the local level. Citizens pay taxes and hence maintain vigilance over the elected representatives and hold the threat of the vote. However, this fails to take sufficient note of the character of policymaking and the influence of interest groups. Some public expenditure decisions do not reflect the strength of the beneficiary constituency in terms of votes.

1.3 Objective, focus and key questions of this paper

This paper contributes to the ongoing debate on methods and current practice of M&E in support of decentralisation in Kenya. At the same time, it contributes to the more general discussion regarding practice in the M&E field. Chapter 2 provides an overview of the changing Kenyan context for external support to decentralisation, drawing on experiences from the past. Chapter 3 then reviews current donor practice in M&E of programmes related to decentralisation and local governance initiatives. The subsequent chapter identifies the main approaches to M&E and to promote local accountability and draws some lessons from these. A final chapter presents options for strengthening M&E and local accountability systems and provides input to the debate on ACP-EU cooperation in the field of decentralisation and local governance.

A literature review and interviews with key informants provided the main information sources for this paper. Decentralisation in Kenya is still evolving, and the country's history of decentralisation is unique. There is thus quite a body of literature on Kenya's experiences and recent steps to restore more political decentralisation and promote local accountability and participation. The key informants interviewed represent both the government and their external support agencies and civil society agencies that have supported decentralised service delivery and the local government system. A list of persons interviewed is provided at the end of the report.

During the past few years, Kenya has engaged in a participatory process of constitution making. During this period the author was involved in public debates in which devolution was one of the subjects that occupied much attention. These public forums also served as an information source.

The period of this study coincides with the Local Authority Service Delivery Action Plan cycle. This planning cycle is a promising development that seeks to link residents and their elected representatives in creative relationships towards better and more accountable service delivery. The author attended two such planning meetings and was able to discuss with residents and civic leaders their views on the process.

2. Decentralisation in Kenya

The Government of Kenya has long maintained a rhetorical commitment to decentralisation. The country attained independence with a constitution¹ that emphasised a devolved system of government. Before independence local governments were quite strong, providing a range of important services such as education, health, roads, water and housing. However this changed drastically over the next decade, with local authorities losing most of their functions and attendant sources of revenue through re-centralisation. This has since been replaced by "de-concentration" of central government functions. Ministry representatives have established a presence at the lower levels of provinces, districts, divisions and locations. These local ministry outposts run in parallel with the provincial governments, which though their mandate is administrative, are often a key player in coordinating local level activities.

2.1 History

In the run up to independence the matter of local government became a key bone of contention, partly because Kenya's pre-colonial inheritance was one of strong localism. It is from this strong tradition, reinforced by vibrant local authorities with responsibility for a range of services, that the tradition of decentralised power grew. The colonial administration encouraged the formation of political associations along regional-ethnic lines, thereby rendering nationalist political movements organisationally weak. The colonialists forbade nascent political parties from organising nationwide, confining such organisation to provinces instead.

Moreover and more importantly, tribal interests based upon land and prospects for future economic position created a polarity in the nationalist movement between smaller tribes and the bigger ones. The Luo and Kikuyu, whether in or out of power, have particularly dominated political debates in Kenya. The alignment of tribal and economic interests eventually led the nationalist movement to split into the Kenya African Democratic Union (KADU), which brought together many of the "small tribes", and the Kenya African National Union (KANU), which brought together the two "large tribes" (the Luo and Kikuyu). KADU demanded a quasi-federal division of power, with strong local government that would leave an African majority government (controlled by the large tribes) less omnipotent than its colonial predecessor. This demand was driven principally by the desire to reduce the control of Kenya's larger ethnic groups.

Negotiations of these parties with the British government produced a quasi-federal constitution, the so-called Majimbo constitution, that gave significant powers to elected regional assemblies. Elected local authorities (councils), with their own local civil servants and bureaucrats, were to deliver most public services, such as education and health. But full implementation of the Majimbo constitution was derailed soon after independence. The winner of the first election, KANU, had never shown any sincere commitment to it and had accepted proposals within the constitution only so as not to delay independence. There was also insufficient support for it among the populace and leaders of the opposition party. After the elections KADU leaders dissolved their party - which had been a strong proponent of devolved government - and took up cabinet positions in a new "unity" government.

The colonial system of government had been quite centralised. The decentralised political structure that the Majimbo constitution aimed to put in place was thus one that Kenyans had not been exposed to throughout the colonial period. Besides the apparent lukewarm commitment by political proponents, perhaps the deadliest blow to this constitution came from the civil service. Central government officials resisted the implementation of the constitution and did little to hide their contempt for "decision making from below". The central civil service bureaucracy successfully maintained a stranglehold on regional development planning and obstructed the transfer of functions. The most telling evidence in this respect was the central government's procrastination in devolving financial powers to the regions as provided for in the constitution. In some cases it even deliberately starved KADU regional strongholds of development funding.

The Majimbo constitution itself was replete with ambiguities, giving central government civil servants leeway to interpret it as they pleased. For instance, provisions were quite vague about the extent and purpose of federalism, procedures for participation and the roles of officials at various administrative levels. This was in addition to the unwillingness of central ministries to provide technical and managerial assistance to local government units.

Moreover, while much of the rhetoric accompanying Majimbo implied that the government was seeking to promote widespread participation, the policy implemented created new arrangements for the central bureaucracy to control local affairs. Inadvertently too, the multiple levels of review and approval through which local plans had to pass created delays that discouraged enthusiastic participation by citizens. This reinforced the bureaucracy's power to veto or modify proposals and created uncertainty and perplexity among rural peoples.

The eight regional governments were abolished in 1966. Nonetheless, Kenya maintained a system of local government operational at the district level and in urban centres, albeit made much weaker by the centralisation of functions and revenue control. Table 1 presents a chronology of events around decentralisation in Kenya. Attempts in 1967 to strengthen these local authorities, to mitigate against the effects and criticism of abrogation of the Majimbo constitution, had little chance to take root as centralisation was accomplished through the Transfer of Functions Act in 1969, by which major services such as primary education, health and road maintenance, were transferred back to the central government. This same Act also transferred the Graduated Personal Tax (GPT), the major source of revenue for local government, to the central government.

Over time other provisions introduced through circulars and directives further weakened local authorities. At one point local authorities were expected to hand over significant revenue sources, e.g. the crop cess, to the district administration. Also, in the past councillors had been part-time - ordinary citizens serving in local government. Amendments to the Act required councillors to be full-time, which lessened the calibre of individuals attracted to the job. In fact, one of the main criticisms levelled against councils is the mediocrity of its cadre of councillors!

Although devolution objectives in the constitution had been repealed by the mid-sixties, the government remained committed to some form of decentralisation, albeit not political. It maintained policy formulation and oversight at the centre but attempted to de-concentrate service delivery. In practice this resulted in central government service providers controlled and funded directly from Nairobi working side by side with elected local governments starved of resources.

Table 1. Chronology of decentralisation in Kenya

Time	Event
1895	Establishment of the East African Protectorate of which Kenya was part.Establishment of the East African Protectorate of which Kenya was part.
1902	Passing of Village Headmen Ordinance guiding establishment of local government. Provincial administration provided power to appoint "headmen" responsible for a village or groups of villages. Duties were mainly maintenance of law and order and assistance in collection of taxes and rural road maintenance.
1920s	The Local Authority Ordinance enacted in 1921 and the Native Authority Ordinance of 1924 saw local native councils introduced. These were local political forums and the first institutions that enabled Africans to sit on colonial government councils. They were not independent however. They were controlled by colonial officers, who were ex officio members and who appointed the native representative on the council, in addition to missionaries and European residents.
1937	Amendment to Native Authority Ordinance provided for election of councillors by the people.
1950	The Local Government Ordinance replaced the Native Authority Ordinance and established African District Councils (for African dominated areas). Europeans had their own councils. Elections were held in 1958. They could not be held earlier because of the state of emergency declared from 1952 to 1957.
1961-1962	Sessional Paper2 No. 2 of 1961 on reconstitution of local authorities (following mergers of European and African governed areas) was followed by the publication of a local government bill and local government regulations in 1962. Local government operated on the basis of the 1962 law and regulations until 1977 when they were replaced by the Local Government Act, Chapter 265 of the Laws of Kenya.
1963	Independence under the Majimbo constitution. Emphasis on strong regional government responsible for service delivery and revenue mobilisation. The constitution provided for elected regional assemblies and governments. Kenya was divided into Nairobi and seven regions. The constitution also provided for local governments within the regions designated as either county or municipal councils. Regional government had oversight over local governments in their area.
1964	KADU (the party that had won votes on a regionalism mandate) dissolved itself and its leaders accepted cabinet posts. This enabled KANU, the ruling party, to change the constitutions with minimum resistance.
1963-1965	Regionalism was dismantled through constitutional amendments, including repeal of regions' power to levy taxes and raise own revenues, abolition of concurrent functions and transfer or supervision of local authorities from regional assemblies to parliament. Other amendments redesignated elected regional "presidents" as simply "chairmen". The Ministry for Finance bypassed the regional finance and establishment committees and dealt with appointed executive officers. Regions became dependent on grants from the centre. Other provisions renamed the regions as "provinces", a colonial term that was used to distinguish between different administrative areas. Other amendments extinguished the regions' exclusive legislative and executive powers.
1965	Publication of Sessional Paper No. 10 "African Socialism and its Application to Planning in Kenya". In this the government expressed commitment to decentralised planning.
From 1966	Application of decentralised development management as embodied in the Special Rural Development Programme, formation of regional development authorities from the early seventies.
1969	Transfer of Functions Act passed. Several major services, including primary education, health and road maintenance were transferred back to the central government. The Graduated Personal Tax (poll tax) collections were passed from local to central government. The GPT was eventually repealed.
1977	Enactment of the Local Government Act embracing the regulations of 1962.
1983	Adoption of the "district focus for rural development" strategy to improve coordination of district level development activities.
1995	Revision of the district focus strategy to embrace participation and a focus on poverty reduction.
2002	Conference on amendments to the Local Government Act, Chapter 265. Amendments not yet presented to parliament because of ongoing constitutional reform.
2004	After 10 months of deliberations, the National Constitutional Conference in March adopted a draft constitution that contains provisions for extensive devolution. This must now be enacted by parliament. However, disagreements in the ruling coalition have ensued regarding provisions on executive power and devolution.

2.2 Decentralised development management

Several initiatives, which continue today, bear testimony to the Kenyan government's interest in and commitment to some form of administrative decentralisation since independence. Yet even today the government and civil service appear disinclined to pursue full-scale decentralisation. The government's foremost strategy statement on economic development, Sessional Paper No. 10 of 1965, emphasises the importance of decentralised planning and extends planning functions to the provinces, districts and municipalities to ensure that progress towards development is made in each administrative unit.

Having successfully thwarted the Majimbo constitution (both politically and bureaucratically) the post-independence government moved quickly to show its continued commitment to local planning and development management by more subtle schemes, such as the establishment of development committees at the district and provincial levels. These committees were to facilitate, among others, coordination of development activities and provide assistance in decision making (Republic of Kenya 1966). Decentralisation of development management reached its apogee at this time with the promulgation of the Special Rural Development Programme (SRDP).

Under this programme decisions regarding project identification and implementation were to be locally made with central government ministries playing advisory roles. Divisional development committees were created to support district development committees. Central government used these structures to exert greater control over district affairs. Indeed, it has even been suggested that the greatest achievement of this rural development programme was to precipitate the decision to move to district level planning which 14 years later came to be known as the District Focus for Rural Development strategy.

In addition to the SRDP, the government established regional authorities in the mid-seventies that covered Kenya's river basins and were to serve as an avenue for development planning at multi-district levels. Conceived in the tradition of the Tennessee Valley Authority of the United States, these authorities were expected to promote coordination of activities within river basins. Activities included hydroelectric power generation and the management of dam siltation.

The District Focus for Rural Development strategy adopted in 1983 was another attempt to improve coordination of development activities and services at the district level. Though much discredited within Kenya by scholars, civil society and political leaders (usually those critical of government), this strategy remains the key approach to planning and implementing development activities by most government institutions. The main objectives of the strategy are four:

• shifting responsibility for development planning and implementation from headquarters to the districts;
• shifting qualified personnel from government departments to the districts;
• institutionalising district planning, including budgeting, in each district;
• establishing a nationwide responsive delivery mechanism made up of various development committees from the sub-locations.

Responsibility for "multi-district" planning and national programmes was to be placed in the hands of the head ministries, while responsibility for the operational aspects of the district-specific rural development projects was to be delegated to the district. Activities of local authorities and non-state actors were to be integrated into district programmes. For this, these local actors were invited to join District Development Committees (DDCs), which formed the cornerstone of the new strategy. Membership of the DDC was to include a district commissioner (DC) as chairperson, a district development officer (DDO), heads of ministerial departments, area members of parliament, the chairperson of local authorities and officials of KANU (the then-ruling party).³

In theory therefore, the District Focus for Rural Development was a strategy for decentralised development planning and coordination, project implementation, management, procurement of goods and services, management of personnel and provision of information. Local citizens were to be involved in project identification, design, implementation and management. Most importantly, resources were to be shared more equitably by directing them to the most needy areas (Republic of Kenya 1983). Thus in theory the programme offered a strong measure of participatory development.

In practice, however, as with most decentralisation schemes in Africa, and all of Kenya's previous efforts in this regard, since the strategy was state sponsored, it was by definition, an integral part of state politics. This naturally affected the way it was implemented. Its implementation hiccups reflect the challenges to effective decentralisation in Kenya.

Though the DDCs were meant to harmonise plans and proposals from committees at subordinate administrative levels, these subordinate committees were in fact not established in many places. Divisional officers, chiefs and sub-chiefs hold development meetings infrequently, if at all. Where such meetings are held, priorities reflect the wishes of the local elite and supporters of public officials. Furthermore, ordinary members of the subordinate committees have no means of checking that proposals and plans are actually forwarded for DDC approval.

Neither have mechanisms for consulting local citizens been institutionalised or regularised. Politicians often outsource projects or raise funds for development plans that they themselves have conceived. Though they may have local input (often from their constituents), they are not subject to oversight by the DDC. Indeed, overwhelming evidence from observations in the country suggests that the more senior the local politicians are in the national hierarchy, the more local development institutions and initiatives are subject to their will as part of their patronage network. Many community projects are sponsored directly by senior politicians; some even carry their names. This has extended to the bizarre level where projects are abandoned or left incomplete when the sponsoring member of parliament loses the seat, the winners preferring instead to develop their own projects.

Hence whereas the DDCs were supposed to be organs of popular participation, senior politicians have become key to what they do or do not achieve. Senior civil servants are also often drawn into the patronage networks of senior politicians. Senior civil servants often accompany politicians on community visits or fund raisers. During the days of KANU (the former ruling party), presidential functions attracted the most senior civil servants, despite the political nature of these appointments (e.g. party meetings or recruitment).

As a strategy for decentralised development management, the District Focus for Rural Development approach has not met its formal objective of democratising local development. The structures set in place do not promote meaningful citizen participation and do not empower local people since citizens have no way of sanctioning the DDC, which is made up mostly of central government-appointed civil servants.

2.3 Elected local government

The main legal framework governing decentralisation today is contained in the national constitution and the Local Government Act, Chapter 265, of the Laws of Kenya enacted in 1977. Furthermore, the current Local Government Act and Kenya's system of local authorities trace many of their features from colonial times as well as from the local government regulations of 1963.

The Local Government Act confers the Minister for Local Government powers over the life and death of local authorities. In principle, the minister can create, amalgamate or alter the territorial limits of local authorities. The minister can dissolve an insolvent local authority or elevate the status of such an authority. Further, the Local Government Act empowers the minister to control the composition of the local authorities and make rules to guide local government elections. In particular, it is the minister who determines the number of councillors elected to each local authority and the minister also determines eligibility to vote or to contest in the local government elections. In effect, the minister exercises more control over recruitment to local government councils than do local residents. The minister also appoints chief officers of local authorities and controls local authorities' financial affairs, as well as enjoying a wide range of administrative powers, such as making rules on the implementation of the Act and operations of councils generally.

This diminishes the role of local authorities as institutions of popular local participation. Councils are not autonomous. Proposals for more autonomous local government and devolution in Kenya reflect concerns and frustrations brought about by weak structures of "local accountability". It also reveals local frustration with decentralised development management as represented by the District Focus for Rural Development strategy, which emphasises de-concentration of central government service delivery.

The limited scope for policymaking and financial decision making has reduced local government's appeal for ambitious individuals. Many councils consist of individuals without great academic or entrepreneurial achievement. The limited space for autonomy reduces their scope for innovative action that responds to the electorate and instead seems to breed self-centredness. Local councils in Kenya are usually characterised by indebtedness arising from bad spending decisions (usually over-employment at the lower cadres to satisfy councillors' need to reward electorates), disposal of council property (land and buildings) to themselves and little regard for public service.

As noted here, the concept and practice of local government originates from a desire to see the citizens of a locality participating in decision making on local affairs. Central government is too far removed and few other effective mechanisms of accountability exist, e.g. independent audits, parliament and the judiciary. The intention is that local governments would play a foremost role in local development. Specifically, they are expected to provide basic services to residents and serve as avenues for popular representation and local self-rule. How has the Local Government Act facilitated or hindered this?

The Local Government Act, first, ensures that every citizen of Kenya (and every inch of Kenyan soil) falls within the purview of a local government authority. Secondly, the Act provides for citizens to participate in electing councillors in all local authorities every five years. Once elected, councillors constitute an electoral college, whereupon they elect a mayor and deputy mayor (in the case of the city council of Nairobi and all other municipal councils) or a chairperson and vice chairperson (in the case of county and township councils). Additional councillors are appointed to the local councils by various political parties depending on the parties' respective numerical strength in council elections nationwide.⁴

Although citizens participate directly in electing their councillors, the performance of local government authorities as an instrument for facilitating popular participation in local governance, development and service provision is abysmal. This is partly due to the lack of clear provisions regularising public participation in the Local Government Act. As a result, public participation in local authorities has been largely superficial, irregular and ad hoc. The councillors, who are the key channel through which citizens express their desires and grievances with respect to development and service provision, are often unavailable and there is no formal institutionalised space for the exchange of views and provision of feedback to the electorate. Councillors are usually local people. But despite the geographic proximity, few councils have instituted public forums for the interaction with the electorate. Councillors hardly convene public meetings, preferring instead to address meetings called by other local leaders, notably by the chiefs and parliamentary representatives.

Early on the local councillors were drawn from among professionals and the business elite and they served on a part-time basis. However, as mentioned earlier, the administrative changes introduced in the mid-seventies required councillors to devote full-time attention to council work - despite the meagre remuneration provided. As a result, those successful in other occupations no longer offered their time for council work. The councils began to attract personalities who viewed representation as a livelihood rather than a calling. Many of these individuals are more interested in meeting personal economic objectives than in becoming conversant with the operations of government (local and national), its responsibilities, rules and procedures. The ability of councillors to articulate the needs of the people they represent was thus severely curtailed. Indeed, many chief officers - who themselves are usually professional civil servants or former mayors - complain about the calibre of the councillors. In the course of this study one recommendation made by a former mayor pertained to the need for councillors to have basic academic credentials. This is, needless to say, unpopular among aspirants to the office!

Their weakness has diminished the councils' potential as effective local representative institutions providing a strong base for organising local development and political discourse. This has helped tilt centre-local power relations hugely in favour of parliament.

But the councillors do not shoulder the whole blame for the low level of popular participation in local government. Ideally, the Local Government Act ought to have been the legislation that resolved the power balance between the central government, the local authorities and the client (citizens). By its very nature, however, the Act fails to provide an effective link between local people and national-level political institutions, such as political parties and parliament.

The Local Government Act not only lacks provisions that would enable citizens to hold councillors to account, it also confers enormous supervisory and executive powers over local authorities to the minister responsible for local government. As a result, local authorities have increasingly assumed the character of appendages of the central government, rather than institutions fed by and facilitating local participation. Indeed, the Minister for Local Government is the only member of cabinet with executive functions. Every "important" decision made by local authorities must be approved by this minister.

All of these provisions mean that the councils are not autonomous but are subject to central control. Unsurprisingly, residents and other stakeholders have preferred to make their contributions and claims directly to the Ministry for Local Government.

Slight changes have been made to these practices in the recent past, as exemplified by the requirement for local participation in development of Local Authority Service Delivery Plans (LASDAPs), which are a precondition for local authorities to receive a block grant from the central government under the Local Authority Transfer Fund (LATF)⁵. However, requirements for local participation are not yet automatic, being currently associated only with LASDAP processes.

2.4 Provincial administration

Administratively, a powerful parallel appointive government exists in the form of the provincial administrators who wield more authority than elected officials. No legal framework establishes and governs the provincial administration except with respect to its lowest ranks (chiefs and the administration police). The provincial administration is an important actor in development: NGOs and bilateral programmes operating in the rural areas, for example, are required to register with the district commissioner.⁶

Thus through the provincial administration, the central executive is able to exercise invasive powers of control down to the local level. In this regard the provincial administration can be seen as one of Kenya's most extensive experiences in administrative decentralisation. The coordination role of the provincial administration is accompanied by a security role. So whereas this system serves coordination functions, e.g. chairing the DDCs, it has also weakened the local government system, since provincial officers usually mediate individually on behalf of aggrieved citizens. The Administration Police Act establishes a local police force controlled by the district administration. This is the lowest level of police and is therefore the one with which communities most interact.

Figure 1 depicts these parallel systems for delivery of local services in Kenya.


Figure 1: Parallel systems for delivery of local services in Kenya



2.5 Devolution in the draft constitution

The draft constitution approved by the National Constitutional Conference in March 2004 (but which has not yet been enacted) proposes extensive decentralisation of services and resource control as well as delegation of decision-making authority to lower levels of government, where the district is the principal unit.

Under this arrangement elected district councils with an executive head will be responsible for many services including primary education and health, road maintenance, land use management and planning. For this, the councils will raise local resources and receive subventions from the central government to fulfil their responsibilities.

While the anticipated devolution respects and maintains Kenya as a unitary state, it delegates responsibilities and authority to three levels of devolved units: the region, the district (as the principal level of devolution) and the locations. The nature of power sharing between the centre and localities is usually quite a contentious design issue when it comes to political decentralisation in a unitary state. Kenya's draft constitution assigns responsibilities to the districts and other sub-national units without providing guidance on the pace of devolution. This is to be addressed in legislation governing devolution's implementation. However, the failure of the draft constitution to provide for phased implementation of devolution - so as to build capacity within the district councils - is one reason given by those opposed to the constitution for the delay in its enactment as it now stands. Another reason given is the constitution's lack of guidelines governing disputes between the centre and devolved units or for preventing the centre from delaying or mismanaging the transition to devolution.

The draft constitution also leaves unclear the motivation behind devolution. This revives an ongoing debate about the objectives of devolution and the role of local government: whether it is principally about deepening democracy or only an instrument for service provision. Literature on democratic decentralisation argues for a comprehensive form in which service delivery reflects local priorities and control. Decentralised development management, which Kenya has pursued since the mid-sixties, emphasises the efficiency of service delivery more than local control. The case study on decentralised health delivery discussed later, in Chapter 3.3, is a good example of this approach. Those opposed to full-scale devolution argue that the councils are incapable of comprehensive service delivery, and therefore there remains a need for a parallel system of decentralised ministry functions. This national debate has not yet been consolidated. The absence of clearly expressed principles in the draft constitution suggests that it may yet be ambitious to laud the devolved units as vehicles for local accountability, participation and governance for the pursuit of a pro-poor policy agenda.

The draft constitution intends to repeal the provincial administration in its current form and to bring under the control of the district government all public services within district boundaries. As such, district government is to absorb the structures of decentralised development management. There is hereby a need to rationalise the current partnership system in the field, in which almost all government departments and ministries have a presence in the districts (the district health officer, etc.) Previous experiences suggest that harmonisation is needed to alleviate any resistance to such a plan that might arise among central government civil servants. The tendency of Nairobi-based ministries to retain lines of communication with the district heads is expected to remain. This will likely become a source of friction with local councils and ought to be addressed in the devolution legislation as well.

The draft constitution provides for an elected government at the national level, the district level and the location level. A fourth tier of government, regional, between the national and district would be constituted through indirect elections by the district council. The head of this regional government is to be elected as a district administrator or governor. All existing local authorities (e.g. the county councils, which usually cover the whole district, and the urban councils, which are usually localised governance for conurbations of more than 5,000 persons) are to be subsumed within the district government. These governments are expected to abide by the principle of participation and consultation with citizens, as enshrined as a directive principle of state policy.

The council and not central government will make decisions regarding local service delivery, with central government confining itself to developing national policy. The elected district assembly will hold the district government to account in the same manner that parliament holds central government to account. These new developments are expected to foster local accountability.

3. External assistance to decentralisation in Kenya

As discussed, Kenya has pursued different initiatives in administrative decentralisation. Though political decision making has not been sufficiently decentralised, there have been attempts to de-concentrate central government functions.

Efforts to decentralise development activities have benefited from external support. The Special Rural Development Programme (SRDP) of the 1960s received pooled donor support. Regional development authorities also relied on external support for their main initiatives. Hydroelectric power production, for instance, benefited from loans provided by institutions such as the World Bank (which supported the Tana and Athi River Development Authority), the French (which as a bilateral donor supported the development of Turkwell hydroelectric power project undertaken under the ambit of Kerio Valley Development Authority) and lately the Japanese (in support provided to the Lake Basin Development Authority).

Following adoption of the District Focus for Rural Development (DFRD) strategy in the early eighties, the country saw focused external support. The Danes led donors in channelling support through the Rural Development Fund. The International Fund for Agricultural Development (IFAD) also made resources available to the government for programmatic support. The DFRD strategy was also supported by various donors that chose to work within particular districts to promote integrated rural development, which was in vogue in the eighties and early nineties. Examples are the programmes started in the districts of Machakos, Kitui, Keiyo, Laikipia and Kajiado funded by the Danish, Dutch and Swiss governments along with the Arid Lands Resource Management Programme supported by the World Bank. This last project covered most of northern Kenya from Turkana in the north-west to Mandera in the north-east and to the coast. Most of these were districts in the arid and semi-arid zone, and the focused external support was largely the result of concerted attention being paid to these lands following the development of a national policy for them in 1979. Between 1988 and 1993, the European Commission provided resources for a micro-projects programme.

A main feature of these programmes is that they worked within the government's decentralised development structures. They were implemented at the district level and sought, insofar as possible, to use the structures set up under the DFRD strategy and the DDC system. Though touted as a system that would promote local participation and accountability, the membership of the DDCs and control by the provincial administration militated against this objective. The DDC system subsequently became discredited and associated with political manipulation by the central government, intimidation of vocal community representatives and abuse of external resources channelled to rural development and other public funds.

There was an attempt to revamp this system in 1995 to introduce greater community participation in project planning and implementation, especially following the commencement of the safety net programmes, such as the Social Dimensions of Development Programme (SDDP), which aimed at alleviating the impact of structural adjustment on the citizenry. However, consistent with previous governments' record of compliance with reform promises, not much came out of the initiative.

The last years of the nineties thus saw attempts and innovations intended to promote local participation and accountability within the established frameworks of decentralisation. Some have sought to advance the democracy movement by initiating institutions and practices for more local participation and decision making. This has been promoted both by local organisations (both community-based and non-governmental) and by large international NGOs such as Action Aid, Oxfam and CARE, as well as in the context of bilateral assistance.

It is therefore possible to identify a distinct category of external support: that focused on innovations and reforms towards deeper decentralisation. This includes support to decentralise management of services that were under the control of central government (e.g. health, education and roads) and assistance to to local authorities in service delivery.

Nonetheless, the Kenyan government has not yet articulated a comprehensive policy framework on decentralisation in a clear policy paper, though the draft constitution contains provisions on devolution. In 2003, the Ministry of Planning launched a process to formulate a decentralisation policy, but it has not moved as quickly as anticipated due to the slow pace of the constitutional reform talks. The Ministry for Local Government began a review of the Local Government Act and prepared and discussed a new draft bill with stakeholders in 2002. However, this project too was shelved to await final outcomes of the constitutional reform.

In this context, external assistance to decentralisation has taken the form of stand-alone projects, unconnected with one another and lacking clear direction from government. Experiences from these initiatives, however, can inform the policy development process. The wide variety of initiatives found on the ground in Kenya demonstrates how the absence of a harmonised framework can lead to a wide variety of approaches to decentralisation, from which lessons can be drawn with which to consolidate a clear policy position.

The various types of support can be grouped into three main categories:

• support to the local government system;
• improvements to decentralised service delivery;
• support to democratic community-driven development.

The following sections delve into these types of support and the external donors providing it. The programmes reviewed are recent, from the last ten years or so. The review focuses on innovations and is intended to produce lessons for further application and replication in the country. The implicit objective is to inform future decentralisation policy and designs. Critical in this regard is the capacity of programmes to facilitate learning within them and from them. This chapter, therefore, reviews not just the programme objectives but also the designs for learning within them.

3.1 EU external actors supporting decentralisation in Kenya

A range of actors have been involved in decentralised service delivery within the main government frameworks discussed above. Local government improvement within the Kenya Local Government Reform Programme received support from the World Bank and the UK's Department for International Development (DFID). DFID, which has also supported specific initiatives in Nairobi and Mombasa (the two largest cities in Kenya), is now supporting Poverty Reduction through Optimising Local Governance Systems (PROLOGS). PROLOGS will document lessons from previous support to local government through the Kenya Local Government Reform Programme and especially the results of the Local Authority Service Delivery Action Plan (LASDAP) and the Local Authority Transfer Fund (LATF). The programme has a budget of some GB£ 5.8 million (about EUR 8.5 million) excluding management costs. The EC Delegation has provided direct support to local government with its micro-projects programme and, more recently, through a "social fund", the Kenya Community Development Trust Fund (CDTF). Support from the 9th European Development Fund is channelled through this fund to improve local government in several rural councils at an estimated cost of EUR 21 million.

Other actors have supported decentralised service delivery by line ministries. The Swedish government, for example, through the Swedish International Development Agency (SIDA), supports decentralised health care delivery and improvements in line with a national strategy developed by the Ministry of Health at a budget of SEK 75 million (about EUR 8.2 million). The Danish government previously supported the same ministry - prior to full-scale decentralisation - but withdrew some five years ago citing high levels of corruption. The Danish government has also been involved in the water sector. In 2003, Kenya started implementing the Water Act, which decentralises water management to regional boards. The Danish government continues to support this decentralisation effort, though the modalities of the effort are still being worked out.

Other support is directed to communities. In this regard, many initiatives around integrated development programmes and arid and semi-arid lands development have received support from EU partners, including Denmark, Switzerland, the Netherlands and Germany. These programmes have worked mainly through the DDC system.

A high level of support is provided to community-driven development, following frustrations with the governmental system and associated calls for more participation and involvement of local communities. Germany, for example, supported the Kilifi Development Programme and the Semi-Arid Rural Development Programme was being implemented by the Netherlands Development Organisation (SNV), which was operating in three areas of Kenya. Though it appeared promising, this programme wound down in 2002 due to a shortage of funding and change of strategy by SNV towards support for institutional development at meso levels of government. The German-supported programme in Kilifi has become one of the often-quoted examples of community-driven development. A review of the experience is useful in exploring the development potentials of lower levels of local government and other community-driven initiatives. NGOs are also quite active. Action Aid and Oxfam, which derive some support from European governments, are notable players here.

Few EU donors had large portfolios in Kenya during the nineties. This low level of activity coincided with a period in which credibility and integrity problems associated with the previous government were highest. The support that was provided was characterised by extensive direct involvement of the external partners in project implementation on the ground. Following the change of government in 2002 several partners began developing new initiatives, and there are indications of significant increases in portfolios. However, few of these programmes have reached fruition - they are works in progress.

The initiatives of the nineties are also characterised by attempts to minimise the role of government; hence the deliberate attempts to build other community-driven structures. Some of these structures have taken root and provide useful lessons for the design of comprehensive decentralisation, both political and administrative. A number of these programmes is discussed in more detail below.

3.2 Support to local government improvement

The Kenya Local Government Reform Programme
The Kenya Local Government Reform Programme seeks to enhance the 'capacity, power, and responsibility of locally elected councils' so as to improve 'service delivery, enhance economic governance and alleviate poverty' (World Bank/DFID 2002: p. 9). Though conceptualisation of the programme was initiated in 1991, operations really took off in 1996. The World Bank and the United Kingdom's Department for International Development (DFID) provided the main support for the effort.

The programme has achieved several milestones: rationalisation of the business licensing process so as to reduce the administrative burden on local authorities and development of an integrated financial management system. However, the programme's foremost achievement has been establishment of the Local Authority Transfer Fund (LATF) (see also box). The centralisation of functions in 1967 and the gradual erosion of independent sources of revenue weakened the financial base of many local authorities. The LATF was established to counter this, to provide local authorities a discretionary block grant to use for service delivery and organisational improvements (World Bank/DfID 2002: p. 10). To access the grant, local authorities must submit four items:

• a service delivery action plan prepared through a participatory process;
• timely budgets, with at least half of the grant amount allocated to capital projects;
• a statement from debtors and creditors and a debt reduction plan;
• a revenue enhancement plan.

Resources for the fund are provided by five percent of all income tax collected in the country. This amount was anticipated to gradually rise to twenty percent, but has as yet remained static.

A review of LATF implementation published in its annual report for 2001-02 suggests that the Fund has been successful in providing both additional resources and incentives to local authorities to improve service delivery, increase participation in governance and enhance financial management and accountability. Some Kshs 1.5 billion (EUR 15.3 million) was identified as the aggregate "resource envelope", which amounts to approximately 15 percent of locally available resources. The Fund has also generated a steady flow of local-level planning and financial information for policy analysis and monitoring purposes. As a result of the Fund, up-to-date abstracts of accounts are now available to the controller and auditor-general for all local authorities. Debt reduction plans have been signed between local authorities and their statutory creditors. Debts are being repaid, and debt control mechanisms put in place. Local authorities were estimated to have spent Kshs 1.1 billion (EUR 11.2 million) on 1,366 capital projects in financial year 2001-02. A substantial proportion (31%) of this expenditure was on vehicles and council premises. Twenty percent of the capital expenditure was allocated for roads and 14 percent for water and sanitation.

To help ensure that expenditures are targeted in a responsive and accountable manner, the local authorities are now preparing revenue enhancement plans and consulting with their citizens through a participatory process - the earlier mentioned Local Authority Service Delivery Action Plan (LASDAP) process - to identify local priorities to be included in the annual budget. The first service delivery action plans vary in quality, but in general they represent a step forward in mobilising citizen input in local-level planning. During the first year, more than 27,900 individuals participated in more than 900 meetings through the process. In the action planning process, citizens identified more than 3,000 projects and activities, which are certain to influence the 2002-03 annual budget preparation. The challenge is to sustain this participatory approach, broaden and deepen the participation, and expand the role of the community in designing, implementing and monitoring local services and programmes. Indeed, interviews with citizen groups (resident associations) and civic leaders suggest a need to further institutionalise participation and expand it to encompass implementation of projects and services as well.

In the action planning process, the local authorities have convened meetings to identify projects. But they have not yet developed forums and spaces for citizen engagement during programme implementation. In this regard, many local authorities treat the demand for participatory planning as nothing more than a hurdle to gaining LATF funds. It is instructive, however, that in some local authorities, e.g. in Nandi, communities have formed watchdog committees to monitor budgetary compliance with community proposals.

Even though local authorities are submitting projected and actual information on revenues and expenditures, including abstracts of accounts, they have so far tended to overestimate revenue. This has led to budget deficits and forced the postponement of some planned services and projects.

The local authorities are responsible for M&E of the LATF as well as for monitoring their own revenues, with oversight in the hands of the Ministry for Local Government and the controller and auditor-general. As a block grant, LATF monies are incorporated along with own revenues into the local budget. The Local Government Act (Chp. 265) lays down the rules on accountability for formulation, execution and monitoring of council budgets. The Ministry for Local Government approves the local budget and has the power to conduct "extraordinary inspections" to address any issues that may arise related to local authority budgets and performance. The Local Government Act enables the central government to take disciplinary action in cases of violation and non-compliance. Such actions may include the application of surcharges, suspension and dismissal. Similarly the controller and auditor-general have the power to monitor ongoing execution of local authority budgets and to issue reports and instructions for corrective action as necessary. Further, the controller and auditor-general audit local authority abstracts of accounts.

Local government inspections do take place and reports are fed back to the local authority. This is a main way in which the central government monitors and evaluates service provision at the local level. Nonetheless, in the course of this study, several interviewees pointed out that the inspectorate has insufficient capacity to monitor all of the country's 174 local authorities. In addition, the local authorities have often failed to act on recommendations made by the inspectorate. Yet other commentators pointed out that the inspectors' recommendations were not "redemptive" enough to enable willing councils to undertake remedial measures.

The LOCAL AUTHORITY TRANSFER FUND The Local Authority Transfer Fund (LATF) was established through an Act of Parliament in 1998 to provide resources and incentives to enable local authorities to (i) improve local service delivery, (ii) strengthen financial management and accountability and (iii) eliminate all outstanding current local debts. The LATF was initially capitalized with two percent of the national income tax during its first year of operation (financial year 1999-2000) but has since been receiving five percent of the national income tax. The Fund distributed Kshs 1 billion (EUR 10.2 million) in FY 1999-2000, Kshs 2.3 billion (EUR 23.5 million) in FY 2000-01, and Kshs 3 billion (EUR 30.6 million) in FY 2001-02 to all 174 local authorities based on a simple and objective formula. The LATF is administered in accordance with provisions contained in the LATF Act and Regulations. The Fund is under the responsibility of the Minister for Finance and administered by the Permanent Secretary of the Ministry for Local Government. An independent advisory committee headed by a private sector representative provides advice on the policies and operations of the fund. The Fund operates in a transparent manner. Newspapers regularly publish reports on the allocation criteria and conditions and planned and actual disbursements. In addition, the Fund publishes an annual report that summarises its operations and overviews planned and actual revenues and expenditures for all local authorities. The LATF is structured to provide local authorities with supplementary funds, which are combined with own-source revenues, to meet Fund objectives. Local authorities are given the discretion to allocate the funds and their own-source revenues through the annual budgetary process as stipulated in the Local Government Act of the Laws of Kenya and related financial regulations. LATF monies are subject to the general budget and financial management conditions and not to specific project requirements. Local authorities are accountable for adherence to their budgets, including the use of the LATF monies, under the Local Government Act. LATF allocation criteria are designed to ensure that funds are provided in a predictable, transparent and fair manner. LATF funds in FY 2001-02 were allocated based on the following published formula: • a basic minimum lump sum of Kshs 1.5 million (EUR 15,306) to all local authorities (9%); • a sum allocated based on the relative population of each local authority (60%); • the remainder allocated relative to the proportion urban population (31%). To receive LATF funds, local authorities must submit budget estimates outlining how the funds, in combination with the local own-source revenues, are to be used. There were three general stipulations for FY 2001-02. The local budget must have allocated the equivalent of at least half of the LATF service delivery amount to capital projects (including maintenance of capital) and it must not allocate more than 65 percent of the total budget to personnel (including consultants). In addition, local authorities are required to pay all statutory charges in the year in which they are due. The Fund observes strict deadlines for the submission of all required documentation. In the event of late submissions, the local authority is penalized. In FY 2001-02, 19 authorities submitted their documents late, resulting in a loss of Kshs 45.7 million (EUR 466,327), about 1.5 percent of total LATF allocations. Of this penalty, Nairobi City Council lost Kshs 8 million (EUR 81,630) for its failure to submit the required abstracts of accounts. In accordance with the LATF regulations, an individual report was sent to each local authority detailing the LATF amounts disbursed and explaining any penalties.

The LATF is now in its fourth year and, according to government annual reports, it is helping local authorities respond to the challenge to play a more active role in improved governance, service delivery and poverty alleviation. It is anticipated that various LATF-linked reforms will be gradually institutionalised as Kenya's decentralisation policy - and the institutions that eventually emerge from that process - is finalised.

Poverty Reduction through Optimising Local Governance Systems
PROLOGS (Poverty Reduction through Optimising Local Governance Systems) is a short, ongoing programme intended to consolidate lessons from various initiatives to support local government in Kenya. The idea is to document lessons that can inform decentralisation policy and institutional design. At the same time it explores the capacity of local authorities to undertake pro-poor work and participatory planning.

PROLOGS has four components. The national component is essentially a continuation of the Kenya Local Government Reform Programme, which DFID has been supporting. It covers the development of a targeted investment fund for poverty reduction, formulation of a national decentralisation policy, strengthening of local linkages with macro reforms and sector reform activities and capacity building for LATF-linked reforms. The Nairobi component is a continuation (albeit slimmed down) of the Nairobi Urban Poverty Project (NUPP) funded by DFID as the Kenya Urban Poverty Programme (KUPP). The Mombasa component is a version (also reduced) of the Partnership Approach to Meeting the Needs of the Urban Poor (PAMNUP). Finally, the all-new rural component is a pilot action research exercise focused on just two county councils, those of Kwale and Nandi. Each of these components will benefit from grants targeted at poverty reduction from the planned Poverty Reduction Co-financing Fund (PRCF) to be made available following compliance with conditionalities associated with the LATF.

PROLOGS has been in operation for one year and is scheduled to end in mid-2005. M&E of the programme has included secondment of programme staff to local councils, regular field visits and team meetings. A mid-term review and eventually an end-of-project review are also planned. The emphasis of M&E has reflected PROLOGS' design as a learning programme. Field staff make observations and collate the information during team meetings. Lessons thus synthesised are then communicated to the ministry and to DFID. PROLOGS monitoring activities have already contributed to changes in the Local Authority Service Delivery Action Plan process. These have pertained mainly to making participatory planning a precondition for accessing LATF resources.

In terms of expansion of local accountability, PROLOGS has set a rather limited role for itself, as it interprets its task strictly in terms of documentation and learning. Though it has made some creative suggestions on institutionalising participation in the use of LATF monies nationwide, it does not yet promote feedback from the local level or provision of approaches and advice other than those suggested by the seconded staff. This approach is not as empowering for the local communities and councils willing to work hard and try new approaches.

Support to Urban Councils
The United Nations Programme on Human Settlements (UN-Habitat) supports several initiatives in Kenya which though small nonetheless offer useful lessons for support to local government. The Lake Victoria Cities Programme supported by the Swedish International Development Agency (SIDA) works with the three large municipalities on the edge of Lake Victoria - Kampala (Uganda), Kisumu (Kenya) and Musoma (Tanzania). The programme supports these cities in participatory formulation of city development strategies and in raising resources to implement the strategies. This is undertaken via a citywide consultation process. With this support, for example, the city of Kisumu completed strategic and investment plans to address the difficult issue of solid waste management.

UN-Habitat has not yet worked out a proper M&E framework for its activities. Indeed, during this study the organisation was involved in discussions on indicators for improved urban governance. The traditional approaches of monitoring visits and end-of-project reviews are utilised however.

The Intermediate Technology Development Group (ITDG), an international NGO headquartered in the United Kingdom, has supported urban councils to develop programmes focused on urban poverty reduction. ITDG was a partner in the Kenya Local Government Reform Programme especially active in assisting municipalities in understanding the Local Authority Service Delivery Action Plan process. ITDG has also helped foster community participation in the action planning process through independent community sensitisation and by facilitating community-level meetings around the process.

One often-mentioned problem with improved local authority management is the absence of credible information that can be used in planning. ITDG has assisted one municipal council, that of Nakuru, to develop a "local urban observatory". This "observatory" is a tool for strengthening participation and accountability, providing a city or town a focal point for urban policy development and planning. Collaboration among policymakers, technical experts and representatives of partner groups is fostered within the observatory, where urban statistics and information are gathered and stored for policymaking and consultative purposes. These local urban observatories are part of the Global Urban Observatory initiative, which assists governments, local authorities and civil society organisations in developing 'better information for better cities' (UN-Habitat 2003). Establishment of the local observatories reflects the appreciation that most councils currently have a weak information base for planning purposes and they have few structures for elected leaders to engage with communities.

Working with the Nakuru council, ITDG began to build an information base containing critical poverty-related data, e.g. that on water and sewerage services, education and health facilities, disease incidence, electoral boundaries, crime, waste management, environment, land tenure and land use. The information was analysed and GIS maps were produced. The maps juxtapose several categories of information, such as the location of schools with the location of water pipes and sewers and leaks. The maps have proven to be powerful visual instruments for interested elected leaders and communities to use to persuade councils to locate needed services in needy areas. Such information is crucial for citizen claim making and decision making by councils. Information bases such as the local urban observatory will be invaluable in building robust local authorities in Kenya.

In terms of monitoring, ITDG has utilised the traditional means of field visits and reviews of programme progress. Working with only one council has enabled the NGO to distil lessons from visits and incorporate these quickly into project operations. Once running, the observatories have improved citizen participation and promoted accountability through the information they hold. However, much depends on the level of analysis and access to the information; that is, whether it is in an accessible location and whether interested citizens can use it.

Rural council capacity building
The rural council capacity-building programme is an initiative of the European Commission. Though the Commission had little prior involvement in decentralisation, it has implemented specific programmes that work in a decentralised manner, e.g. one focused on health in Kenya. The Commission has further been involved in community development in Kenya, first through its micro-projects programme within the Ministry for Planning and subsequently using a social fund approach with its Community Development Trust Fund (CDTF).

An evaluation of phase 1 of the CDTF revealed that whereas the programme was more successful than its predecessor in project completion and relevance, few of the projects had forged institutional linkages with local accountability structures. Hence, lessons from the projects could not be applied more broadly. The Commission therefore initiated a new programme, Rural Poverty Reduction through Local Government Support, which seeks to improve accountability and the responsiveness of local authorities in delivering services to the rural poor. This newest programme aims to empower rural communities in the development process as well as build local government capacity in participatory planning and pro-poor service delivery. Each year the programme will select several rural local authorities to support, starting with 15 local authorities and later scaling up to 30 councils.

This is a new programme. It only recently received approval from the government and has yet to start implementation. Once it begins, specific indicators will be developed for M&E. Meanwhile, project sponsors anticipate utilising periodic evaluations, field visits and twice-yearly audits. In addition to monitoring done by the Commission, the project will seek ways to increase the capacity of the Ministry for Local Government to inspect the local councils and respond to council needs following inspections as well as to provide technical support to the provincial local government office.

n terms of promoting local accountability, the programme will encourage the use of the Local Authority Service Delivery Action Plans for accessing Commission-provided resources. Community considerations will thus be reflected in the project activities. The project will also strive to promote participatory M&E by citizens in the participating local authorities.

3.3 Improvements in decentralised service delivery

Besides support for improved local government, external support has been geared towards sectoral improvements. Reforms initiated within key service sectors have responded to the weaknesses of District Focus for Rural Development (DFRD) structures, as well as those within the local authorities themselves. In essence, programmes have attempted to bypass the established systems of the DFRD and local authorities and create new institutional and financing arrangements. This has affected areas including education, water, roads and health.

Decentralised health management
A health policy framework first approved in 1994 and an implementation and action plan developed in 1996 eventually led to Kenya's adoption in 1999 of a national health sector strategic plan (1999-2004). The plan named decentralisation as one of its key strategies for providing quality health care. The government's intention was to decentralise authority for decision making, resource allocation and management of health facilities to the district and facility levels. The Ministry for Health established a secretariat at its headquarters to oversee the decentralisation process.

District health management boards were set up and district health management teams made up of technical personnel established at the district and facility levels. The teams were to represent community interests. The Ministry for Health is currently piloting this approach to decentralisation in 15 districts with the support of SIDA and the World Bank. Its intention is to gradually expand the programme to other districts under the oversight of the ministry and provincial health management teams. The decentralisation strategy has no role for the local authorities.

SIDA supports seven of the 15 pilot districts, with the World Bank supporting the other eight. The assistance provided is in line with the National Health Sector Strategic Plan. The plan emphasises the need for the district health teams to formulate their own health plans before they can access money, drugs and other supplies. The reform secretariat and the district health management teams are the most developed arms of the decentralisation strategy. The district teams are made up of civil servants, however, and most were transferred to the area in which they work. They have therefore exhibited little ownership of the health plans they implement - though they do have a functional relationship to the plans!

The programme is monitored through periodic visits to the districts, quarterly meetings of the district health teams with the sponsors (SIDA and the World Bank) and regular reports and evaluations. At present M&E emphasises collection of health information (morbidity, case loads, medicine tracking) and reports on work flow and finances. Evaluations mention the inadequate ownership of work plans, but effectiveness is viewed from a purely functional perspective, 'by and large the view at the district level was less one of ownership and increased responsibilities and more one of possibly more reliable and timely supply of funds, materials and equipment' (Tamm 2002: p. 22). Monitoring information is not processed at the collection points - the clinics - but rather is passed on to the district health management teams who then pass it on the central office. It is then all processed at the ministerial headquarters. While there are useful lessons here, it is also clear that this mode of decentralisation, which is not anchored in community and local oversight, is fraught with challenges.

3.4 Support to democratic community-driven development

Several NGOs work to support community development in Kenya. Action Aid, CARE, Oxfam, the Aga Khan Foundation are all active, along with bilateral agencies such as the German Agency for Technical Cooperation (GTZ) and the Dutch SNV. While their focus is principally on community development these agencies also emphasise democratic self-governance in their work. In many of the programmes, communities apply for development support, analyse their own problems and situation, explore solutions and form representative bodies to steer the development process.
The Kilifi District Development Programme is a multi-sectoral district-based rural development programme supported by the Kenyan and German governments and implemented by GTZ. It aims to build the development capacities of communities and service providers in Kilifi district and to contribute to overall poverty alleviation.

The programme encourages communities to organise themselves and mobilise resources towards their own development needs. The communities self-organise and elect village development committees made up of clan elders (without whose consent it is difficult to make progress) and younger men and women. These committees have oversight over the development aspirations of their village and are fully responsible for overseeing projects. They receive technical support from the government and the project, as well resources beyond those that can be mobilised from within the community. The Kilifi experience indicates that self-organised communities such as these villages can, with the oversight provided by peer development committees, facilitate village administration, development and a wide array of services, e.g. sanitation and water, maintenance of roads (or paths), village-based health services, education (nursery and primary schools including bursary administration), village markets and social security. This is the range of services expected of well functioning local authorities. The experience has led to advocacy for the institutionalisation of villages as the lowest government level. In the current government structure the lowest level is the ward, which consists of several villages.

Organisations such as Action Aid, which adopt rights-based approaches, increasingly encourage communities to engage with the public institutions around them. Kenya has no institutionalised mechanisms for citizen engagement with public bodies. Though no consultation and oversight processes are built into law, expansion of democratic space over the last decade has been accompanied by pushes for such consultation.⁷

The introduction of the LATF, discussed above, and the attendant expectation that the Local Authority Service Delivery Action Plans would be prepared in a participatory manner, has opened an opportunity for institutionalised participation. Action Aid and other NGOs have mobilised communities to participate in the action plan formulation process and, in several localities, to participate in implementation of the plans as well. Several local authorities, e.g. the Bondo and Embu county councils and the local authority in Nandi, have standing Local Authority Service Delivery Action Plan committees. The action planning process thus seems to have some promise for formally linking communities with public processes in a decentralised system.

In terms of M&E, NGOs utilise the traditional methods of reports, visits and periodic reviews. They are also increasingly using participatory M&E methods, such as citizen scorecards and report cards. NGOs emphasise sharing and analysing feedback at the local level and upward transmission of input for processing at the centre.

Capacity building for decentralised development
SNV, a Dutch NGO that previously worked primarily in direct project implementation, has now shifted its strategy to capacity development (advisory) services for organisations that work at the meso (intermediate) level. These include organisations working at the district or multi-district level and those active within a specific sector or sub-sector. At present SNV provides advisory support to 53 organisations in Kenya, mostly district-based, as well as to several advocacy organisations. The support focuses on strengthening decentralised structures for public institutions (e.g. councils) and for NGOs. Emphasis is on helping partners access and influence decision making at the district and national levels. Local ownership, capacity building and institutional strengthening are the SNV's guiding principles.

SNV uses both a qualitative and quantitative M&E systems, though use of the quantitative tools is more advanced. Quantitative tools record the spread of the work in terms of numbers of organisations reached, advisory days provided and issues tackled. During the period of this study, SNV was finalising development of its qualitative M&E tools. Further, the organisation reviews the findings of M&E with its partner organisations on a quarterly basis, incorporating partners' responses into the reports.

The range of efforts described in this chapter illustrates the variety of innovations within Kenya with regard to decentralisation. As argued earlier, these different innovations have arisen from the absence of a coherent national policy framework and reflect a certain lack of bearing for decentralisation in the country.

4. Decentralisation, M&E and local accountability in Kenya

Decentralisation is a complex process. It is hard to predict with certainty what will or will not work. Systems that facilitate learning are critical in such a context. This chapter turns to the main approaches to M&E and local accountability found in the initiatives described in Chapter 3. It then distils some lessons from those approaches.

4.1 Main approaches to M&E

The programmes reviewed utilise different approaches to M&E. At present it appears that the systems in use promote learning principally by the sponsoring institution. Several supporters/sponsors, e.g. DFID and SIDA, have developed M&E systems that they utilise throughout the work they support. Reports are prepared in a manner that suggests that the donor agency is the primary audience.

The Government of Kenya, while acknowledging the importance of M&E, has not quite developed a structure and harmonised institutional framework for it. During the past year, efforts to document current approaches to M&E reveal that multiple agencies and institutions are responsible for it. However, few programmes have developed M&E guidelines as early as the design stage. Approaches to monitoring instituted in the eighties and nineties, and the location of these activities at the provincial level, have assumed an inspection and policing tone. In addition, there have been few attempts to learn from monitoring reports. The learning that has taken place was stimulated by external partners late in the lifespan of long-term programmes (e.g. the review of DFRD's progress in promoting participation which was incorporated into the strategy papers in 1995, 12 years after commencement of the programme).

Several organisations admit that M&E is not a strong feature of national programmes at the design stage. While this is understandable, it is regrettable since lessons derived from practice would be useful for further development and for the anticipated countrywide roll out of decentralisation efforts. Where they do exist, most approaches to M&E extract information for use at higher levels. Information is seldom analysed at or fed back to the local level. Few programmes factor in corrective measures in response to information and knowledge gathered from monitoring.

Indeed, the predominant forms of M&E emphasise monitoring rather than impact evaluation. Several of these approaches are discussed below.

Standard monitoring: Regular reports, annual reports and validation visits
Many organisations use reports as their main form of M&E. Formats developed by central authorities are completed by project personnel and passed upwards to the ministerial level. The timeframe for this varies from programme to programme, from monthly to quarterly, half yearly and annually. The Kenya Local Government Reform Programme, SIDA support to decentralised health delivery, PROLOGS and SNV all require such reporting. Only SNV has institutionalised analysis at the lower level. A review of M&E systems within the Ministry for Health revealed limited local use of the information collected in regular reports. The health facilities use at least 45 different reporting forms. The data collected is sent upwards to the district level and later to the centre. Few of the forms are utilised for analysis and review, however, and feedback to the health facility is a rarity.

In the case of PROLOGS, a programme management unit analyses reports completed by staff seconded to the four participating local authorities. Thereafter, the reports are shared with the Ministry for Local Government and DFID. PROLOGS staff meet regularly (at least monthly) to collate learning and share findings and experiences.

Several programmes also utilise field visits for monitoring. This is quite a popular approach to M&E among the programmes reviewed. Some of these field visits are announced while others are not. They are intended mainly to verify the regular reports received at the centre.

Within the local government reform programme, the Ministry for Local Government has a unit responsible for inspection and oversight of local authorities, the local government inspectorate. This inspectorate conducts field visits to monitor the operations and finances of the local authorities. The local authorities are responsible for monitoring the use of the LATF monies and financial management of their own local revenues. The local authorities are, however, subject to oversight by the Ministry for Local Government as well as by the controller and auditor-general. As a block grant, LATF resources are subsumed with local resources for expenditure through the local budget, which is approved by the Ministry for Local Government. This ministry has the power to conduct extraordinary inspections from time to time, in addition to the minister's occasional inspection visits (Republic of Kenya 2002).

Though the LATF is combined with other revenue and allocated within the annual budgetary process, as laid out in the Local Government Act and other financial regulations, there is nonetheless a tendency for the Ministry for Local Government and other commentators, including the press, to treat the Fund as if it was distinguishable "project" resources. Many commentators speak of "LATF money" rather than "local authority budget" or "public money", the terms generally used to refer to public resources. This differentiation between LATF and own resources carries an implicit threat of withdrawal should the local authorities be seen to fall short in terms of operations or management. This does not augur well for local authority autonomy.

The inspectorate within the Ministry for Local Government has oversight over the local authorities. However, local authorities observe that inspections take the tone of a financial audit. In addition, little immediate feedback is given. Where such feedback is provided, it is in the form of an audit management letter emphasising administrative improvements that need to be made. Follow-up work from the audits may take place only several years later, suggesting that the department has weak capacity in terms of reach and breadth of its work.

Evaluations: End-of-term and mid-term evaluations
M&E systems often include regular evaluations conducted either during programme implementation or at the end of a project period. End-of-term evaluations focus more on the impact of the activities, while mid-term evaluations emphasise progress towards set goals and targets. The programmes reviewed in the previous chapter are all under implementation and most are new. For these, the dominant form of evaluation identified is the mid-term evaluation, e.g. the Output to Purpose Review of the PROLOGS programme and SIDA's evaluation of its support to decentralised health care. There is as yet no evaluation of the Kenya Local Government Reform Programme as such. However, a review of the programme does exist within the context of a broader survey of local service delivery and local government in Kenya. Evaluation of the social fund (the Community Development Trust Fund) led the European Commission to develop the programme reviewed here, which was recently approved by the Kenyan government but has not yet started implementation. UN-Habitat conducted an end of phase 1 review of its activities and shared the lessons with its partners in the three participating cities.

Community development programmes supported by NGOs and the community development dimension of World Bank support to the arid lands initiative increasingly use participatory M&E methods such as citizen report cards.

The Kenyan government conducts national welfare surveys every five years or so. While not targeted at any of the programmes discussed here, these surveys have provided information that can be used as baseline statistics or reference points to determine the impact of interventions. The last such survey was in 1997 and plans are under way for one in 2004.

4.2 Main approaches to promoting local accountability

As discussed in Chapter 3, several categories of support to decentralisation in Kenya can be identified: support to the local government system, support to decentralised service delivery by line ministries, and support to the communities themselves. Though all of the initiatives reviewed share the underlying objective of supporting local accountability, they have obviously fared differently.

The National Health Sector Strategic Plan, supported by SIDA and the World Bank, establishes district-level health management boards. Yet inadequate attention has been given to the development of these boards, which could constitute an anchor to promote local accountability. Moreover, this decentralisation effort has bypassed the local authority systems by establishing community boards. These boards are the least developed arms of the decentralisation effort. SIDA's evaluation notes that among the various institutions called for in the strategy, the district health management boards have become quite marginalised and their 'constitution, mandate and composition is out of step with a decentralised governance system' (Tamm 2002: p. 25). The boards are composed of civil servants, the chairperson of the county council, three community members (one from the business community, one from the religious community and a social leader who is either a women's group leader or youth representative). The minister appoints these boards upon the advice of the medical officer of health (a civil servant). There are no checks and balances on this selection process, meaning that it can be manipulated by those charged with appointments. In this sense the district level health management boards are quite vulnerable to changes in government, as accountability is to the appointing authority.

Community development approaches - such as those utilised by the European Commission in its social fund, by NGOs and by the World Bank - make use of community development committees. These are representative committees chosen by programme beneficiaries or promoters in the case of social funds. Though they often operate democratically, e.g. as in the Kilifi programme, they are not anchored in legally secure democratically legitimised structures. They are not subject to oversight beyond the project area and it is doubtful whether they can be subjected to other institutionalised oversight systems such as a national audit process.

Several programmes have direct links with local authorities, e.g. the Kenya Local Government Reform Programme, as well as initiatives of UN-Habitat and ITDG. A new programme being undertaken by the European Commission will link directly with local authorities. The Commission programme intends to support better local governance and claim making by citizens who, according to project documents, will be 'active participants in the implementation and monitoring of project investments'.

4.3 Lessons from the current approach to M&E in Kenya

This section synthesises some lessons from current approaches and experiences with M&E in Kenya.

1. Monitoring information is seldom analysed locally.
The methods used for M&E extract information from the local level and transmit it upwards, usually to the centre, for analysis and review. It might then be reflected in altered programme designs at some point in the future. In this regard, knowledge is seldom utilised to build more responsive structures on the ground or to foster local accountability. As such, decision-makers in the field are seldom held accountable to consumers and residents in their sphere of operation.

2. Links between monitoring and learning are weak or nonexistent.
Information gathered during monitoring seldom leads to corrective actions to solve the problems identified. Indeed, monitoring often appears to be used merely to verify reports or to document situations that are already well known at headquarters. For instance, the local government inspectorate
(a department within the Ministry for Local Government responsible for oversight/supervision and support of local authorities), regularly reports that local authorities are under great stress and that few effectively utilise the resources in their budget. Similarly, many cases are documented of misuse of resources. Given the gravity of the situation and the regularity of such observations, it is remarkable that the Ministry for Local Government does not yet have a systematised corrective programme it can offer to local authorities faced with such problems and irregularities. The ministry is, however, developing tools and instruments aimed to link monitoring of the local authorities with subsequent improvements and to improve organisational learning within the ministry itself.

In the case of mid-term and end-of-term evaluations, where these had taken place, those involved observed how difficult it was to change the design of programmes immediately in order to respond satisfactorily to the concerns raised. In the case of PROLOGS, a review revealed the need to clarify some aspects of the programme's design. However, by the time these findings were reviewed it was noted that the timeframe of the programme was too short for a comprehensive response to be implemented. In the case of SIDA's support to the health sector, reviews have identified weaknesses in structures for promoting local accountability and therefore the programme is working on ways to resolve this. In other instances, e.g. programmes of the European Commission and UN-Habitat, by the time the final evaluations are completed, the institutions have already designed follow-up programmes, factoring in only lessons from mid-term evaluations, monitoring and changes in national context.

3. High staff turnover disrupts learning and institutionalisation of reforms.
Programmes aimed at decentralisation of service delivery rely on the staff implementing them, especially those staff at the lowest levels and on the front lines of service delivery. This reliance is particularly strong in early years, before systems are routinised and institutionalised. However, programmes in Kenya have been beset by disrupted learning and institutionalisation of reforms caused by frequent transfers of staff.

4. There is a shortage of staff and competencies for collecting, analysing and reporting M&E information.
In addition to stability and longevity, departments need staff who are competent in information analysis and in preparation of rapid responses. Partly because much of the monitoring information is passed upwards for analysis, few institutions have developed competencies in analysis and response at the lower level. For example, in the SIDA-supported decentralised health delivery system, staff were found to be competent in capturing information, but they were unable to analyse it and prepare responses.

5. Establishment of parallel structures mitigates against local accountability.
A number of the programmes reviewed had established parallel structures and systems that bypass local government. Examples are the district health management boards and the village and community development committees. Though the emergence of such structures can be partly understood as a response to the weaknesses of the local authority systems and the low quality of government councils, it nonetheless shields local activities from the purview of typical social accountability structures.

6. Learning systems within the Kenyan government are weak.
The government seems particularly weak in responding to information from the field. M&E information is seldom incorporated into new or revised plans. Indeed, Kenya does not yet have a comprehensive M&E system. The habit of constant learning is developed though disciplined adherence to such an institutionalised M&E framework. Kenya has no developed structures that would promote systematised M&E. Though the government acknowledges the need for M&E, it is not always built into programmes.

7. Links between monitoring and local accountability are still developing,
Not all is lost however. There are ongoing attempts to link monitoring and local accountability. The Local Authority Service Delivery Action Plan process emphasises participation in planning, but participation has not yet been extended to implementation. In some local authorities, citizens have formed watchdog committees to follow through on the planned activities, e.g. in Nandi and Bondo. Community participation in implementation has also been stimulated by NGOs such as Action Aid and ITDG. ITDG accesses information on the budget approved by the county council and the minister and extracts relevant figures for communities to use in monitoring implementation of the projects in their locations.

8. There is insufficient focus on communities in programme design.
M&E systems are often designed with technical personnel or institutions as their primary users. This is consistent with this study's finding that the systems are intended to extract information for analysis and response at the centre (as opposed to analysis and response at the local level). In the case of the health sector programme, the primary users of the M&E system are the district health managers. M&E of the Kenya Local Government Reform Programme paid little attention to the local authorities involved and district residents. This has implications for the design of systems as well as for the broad sharing of the information thus collected.

NGOs such as Action Aid and ITDG have worked with communities to promote participation in the Local Authority Service Delivery Action Plan process. They observe the lack of provisions for communities to be involved in monitoring councils' fidelity to proposals put forth by the community, in both planning and in implementation. This observation is echoed by the Association of Local Government Authorities (ALGAK) and the Kenya Association of Resident Associations (KARA).

Why is there no institutionalised M&E within the government structures in Kenya? • There is no culture of learning. • Civil servants fear having mistakes revealed and the attendant punishment. • M&E is not seen as part of how the government operates, i.e. not part of "business as usual". It requires systemic change. • Past reactions to M&E have been punitive. Examples were given of previous "monitoring" exercises, particularly at the district level, which punished officers for slow programme implementation or other anomalies discovered. • Most officers consider M&E external to them and therefore do not institutionalise it. • There are no political costs for non-delivery. • There are resource constraints, both human and capital. • The institutional structure prevents learning from M&E. Departments within the ministries do not possess structures and forums for learning from each other. Therefore some departments repeat work done by others or do not make reference to previous work in other units. • It is not clear who and what M&E is for. • There is inadequate planning by spending units on what to spend money on, which makes monitoring hard. • Leadership is unwilling take tough decisions. • There is an absence of brave leadership both nationally and in the districts. Results of Brainstorming by participants at a workshop on Participatory Monitoring and Evaluation held in September 2003.

9. Links with political processes are weak.
All politics is local - so the adage goes. The fact that the current M&E approaches do not promote local learning means that they also do not inform local political discourse. The changes experienced in Kenya over recent years show that empowered communities that are able to make claims against the government can play a significant role in development. Parliament has managed to significantly raise its stature by utilising the reports from the auditor-general as well as from independent analysts to interrogate the executive arm of the government. Yet there has been no parallel development within local government.

10. Do donors factor in promotion of local accountability?
It is not clear that donors factor local accountability into their programme designs. The local authority system in Kenya is quite weak and discredited by the central government. It is deliberately bypassed and parallel structures set up. Donors have aided and abetted the development of two parallel structures for local service delivery. The unintended consequence is a further weakening of local accountability and institutional strengthening. The structures for community accountability and participation in sector reforms, e.g. the health management boards, are quite weak and it is not clear that this problem is being addressed.

11. What is the incentive?
In the SIDA-supported decentralisation of health care delivery, it was noted that civil servants at the district level view the district health plans and other measures that promote local responsiveness as "conditionalities" for gaining more resources. There was no local ownership. Incentives for enhancing ownership and accountability might therefore need to be created. Years of centralisation and politicisation of decision making at the centre have limited incentive among government officials to evaluate their work and improve on implementation. There is virtually no learning culture in government and little reward for effective M&E.

5. Conclusion: Options for strengthening M&E and local accountability systems

5.1 In general

Designs that promote learning
M&E is not always factored into programmes at the design level. Where it is, M&E principally serves the purposes of programme promoters and not necessarily of the communities being served. While M&E is needed on the programme side, it is also useful to factor in learning by communities. In this regard, monitoring information should be made available to elected community representatives for their use in local analysis and decision making.

Further, these communities need links to capacity building for claim making. Residents need the ability to access and understand information on the programmes implemented in their locales and to be able to act vis-à-vis local elected structures.

Decentralisation is about local government
Decentralisation is principally about relocating decision making to levels nearer the consumers of services. Yet these lower levels must have the capacity to analyse and respond to feedback from residents and other constituents in their proximity. M&E and the empowerment of local communities will only lay bare the impotence of local institutions should these be unable to respond to citizen action and have to rely on bureaucrats at the centre who are not subjected to the same level of political control.

Decentralisation support should therefore go hand in hand with development and support to local governance structures. Admittedly Kenya's local authorities are quite weak and often lack integrity. Donors have contributed to establish or support parallel processes, for example, of community development and social funds, to bypass the weak local government system. While their motivations are understandable, experiences from such programmes cannot inform decisions at the local level and do not increase community leverage over elected leaders. Hence, they are not fully empowering. This practice therefore needs to be rethought.

Links between donor support and national policy and reform processes
The experiences of various organisations involved in decentralisation have not been reflected in public policy processes. This signifies a weak linkage between donor support and eventual decision making within the government. Effective linkages between donor support and national policy and reform processes are hampered by the absence of a comprehensive public sector reform framework of which the anticipated decentralisation framework would be key. The absence of this framework has led each sub-sector to follow its own process, leading to a variety of simultaneous stand-alone initiatives. However these initiatives have built up demand for more comprehensive reforms, as reflected in various reports and culminating in the draft constitution.

It would be beneficial to link and channel future support to the development of comprehensive public sector reform and decentralisation frameworks. Disjointed support can only lead to more complications along the way and contribute to resistance against more comprehensive rationalisation.

Documentation and dissemination of experiences
There is extensive knowledge and experience on decentralisation support in Kenya, but this has not yet been documented and made available in an accessible format. It is important to compile this information and make it available so that other institutions can learn from it.

5.2 ACP-EU cooperation

The Cotonou Partnership Agreement anticipates a key role for non-state actors and local authorities in its implementation. Article 6 of the Agreement defines the partners of cooperation to include both state (local, national and regional) organisations and non-state actors (private sector, economic and social partners and civil society in all its forms).

The anticipation is that in the future discussions and programming decisions will be made after dialogue and consultation with these new actors. This is a significant shift that needs to be accompanied by fundamental changes in the way business is conducted. In the past there was principally a relationship between states. The Cotonou Agreement foresees involving a whole new range of partners in programming, including in identification of areas of support and channelling of aid allocations under the 9th EDF, as well as in implementing and evaluating some programmes.

Over the last few years, the ACP and the European Union have worked to create mechanisms for this broad-based dialogue and involvement of these new actors. However, it is as yet unclear how well developed these are, given the prior history of planning ACP-EU cooperation (as a largely bureaucratic exercise with little space or skills for participation).

In Kenya, programming for the 9th EDF has not yet been concluded. Past experiences within the country in promoting participation can therefore be instructive. Kenya does not have a history of institutionalised participation by non-state actors and local government. This does not mean that these actors have not participated in policy formulation in the past: they have, but the process has been rather ad hoc and informal and subject to relationships between specific actors. This was especially true of the period up to 1992. Since 1992, incremental steps have been made in expanding the space for participation by non-state actors. While this space has markedly expanded, it is still not institutionalised. The Cotonou Agreement aims to institutionalise this in ACP-EU cooperation. However, indications are that in the current environment results will be less than satisfactory in the first years.

This discussion paper has outlined features of external support to decentralisation in Kenya and the environment in which this is provided. Kenya does not yet have a comprehensive framework for decentralisation. In the absence of such a framework, initiatives towards decentralisation have fallen into three
categories:

• support to the local government system;
• improvements to decentralised service delivery;
• support to democratic community-driven development.

The experiences overviewed in this study, suggest a number of implications for ACP-EU cooperation.

Defining the policy. It will be hard to realise gains from support to decentralisation without a comprehensive policy on the same. In this regard, it would be advisable to allocate time and resources to government dialogue and policy proposals on decentralisation. The framework thus derived would provide an outline for further programme support. Some of the issues to be covered by such a policy are democratic control of local government and the responsibilities of lower levels of government in service delivery.

Patience and support. Kenya anticipates comprehensive devolution when the new constitution is enacted. However, some time will be required for this to be implemented following enactment of the legislative framework for it. During this period, which should coincide with the implementation period of the Cotonou Partnership Agreement, considerable confusion and absence of clarity are expected on the responsibilities of different levels of government. This calls for a trial-and-error, learning by doing approach and differential treatment for different lower level government departments and agencies.

Innovative programming. In the absence of a comprehensive policy, any programmes initiated ought to reflect a commitment to building habits and institutions that support democratic decentralisation. In this regard, even if the only space available is working within a decentralised ministry structure, care should be taken to build in local accountability structures that are also linked with local political processes.

Support to local governments to engage. Local governments have suffered various setbacks that have weakened them in the past. Many initiatives around decentralised development management and community-driven development have arisen due to this local government weakness. Local governments will require support if they are to play the role anticipated for them in ACP-EU cooperation. This calls for building capacity within forums and spaces for joint action by local governments, e.g. the Association of Local Government Authorities in Kenya (ALGAK). It will also require provision of a constant stream of information on the progress of dialogues on programming. In this regard, care should be taken to confirm that dialogue is indeed taking place and that the whole range of actors is informed. It also calls for dedication of staff and space by the traditional actors - the European Commission and the ACP government counterpart - to promote engagement of local governments.

Democratic decentralisation and poverty reduction. A focus on poverty reduction is one of the Cotonou Agreement's main platforms. Advocates of democratic decentralisation argue that decentralisation promotes participation by interested citizens at the local level and accountability in matching government decisions with people's needs. However, other voices advise caution, arguing that decentralisation might not lead to the intended results for poor people, as the local elite could capture processes. It is important to keep this caution in mind and to re-design programmes and activities to ensure maximum benefits for the poor. In some instances, the more far-reaching results may not emerge from the formal policy process, since this process may not be open to influence by the needs of the poor.

Conflicts with sector-wide approaches. Several EU donors (Sweden, United Kingdom and the Netherlands) have expressed interest in providing programme support to the government through budget support and adoption of sector-wide approaches. Often such support depends on decision making at the centre and reinforces the sector/ministry system. As discussed earlier, this type of system has, in the past, been unable to promote local accountability and links with local governments. Therefore, sector-wide assistance programmes must be designed with vigilance regarding local accountability.

Persons interviewed

1. Augustine Odipo, General Secretary, Association of Local Government Authorities in Kenya (ALGAK)
2. Cecilia Kinuthia, Human Settlements Adviser, United Nations Human Settlements Programme (UN-Habitat)
3. Christine Cornelius, Sector Specialist, World Bank Kenya Country Office
4. Clive Davies, Programme Adviser, Poverty Reduction through Optimising Local Governance Systems
(PROLOGS)
5. Clr. Josiah Magut, former Mayor of Eldoret and former Chairman of ALGAK
6. David Kuria, Programme Officer, Intermediate Technology Development Group (ITDG)
7. Eve Odette, Policy Unit, ActionAid Kenya
8. Guy Jenkinson, EC Delegation
9. Jesse Bokhoven, Country Representative, Netherlands Development Organisation (SNV)
10. Josphat Mwaura, Partner, KPMG PROLOGS Management Consultancy,
11. Martin Oloo, Development Adviser, Department for International Development (DFID), UK
12. Mayors of Nairobi, Eldoret and Nakuru; Deputy Mayors of Kisumu and Mombasa; Treasurers of Eldoret,
Nakuru and Mombasa⁸
13. Mr. Kanake, Kenya Alliance of Resident Associations
14. Grace Otieno, Government of Kenya, Rural Planning Department
15. Per Vickstrom, Programme Officer Health and HIV/AIDS, Swedish International Development Agency (SIDA)
16. Roy Kelly, Resident Adviser, Kenya Local Government Reform Programme

References

Brinkerhoff, Derrick. 2001. Taking account of accountability: a conceptual overview and strategic options. Washington, D.C.: US Agency of International Development.

EC Delegation in Kenya. 2002. Impact assessment of CDP 1. Nairobi: European Commission. (internal report of the EC Delegation).

EC Delegation in Kenya. 2003. Rural poverty reduction through local government support. (Programme Document). Nairobi: European Commission.

Johnson Craig. 2001. Local democracy, democratic decentralisation and rural development: theories, challenges and options for policy. Draft prepared for anticipated publication in Development Policy Review.

Kilifi District Development Programme. n.d. Proposal for piloting the reform of development coordination in Kilifi district. Available on the Internet <http://www.kddp.com/otherdocs_.htm>

Kipngetich, Julius. 2002. Devolution of power and democratisation of the state. Paper presented by the author at a internal workshop for the Constitution of Kenya Reform Commissioners. Workshop organised by the Institute of Economic Affairs. Not published.

Litvak, Jennie & J. Seddon. n.d. Decentralisation briefing notes. Washington, D.C.: World Bank Institute.

Oyugi, Walter. 2003. The search for an appropriate decentralisation design in Kenya: historical and comparative perspective. Paper presented at a conference on devolution in Kenya, organised by the Ministry of Planning. July. Notes later revised for a publication of the Center for Law and Research International (CLARION).

Republic of Kenya. 1983: District Focus for Rural Development. Nairobi: Republic of Kenya.

Republic of Kenya. 1998. Local Authorities Transfer Fund Act. Nairobi: Republic of Kenya.

Republic of Kenya. 1999. Local Authorities Transfer Fund Act (regulations), Legal Notice 142. Nairobi: Ministry of Local Government, Republic of Kenya.

Republic of Kenya. 2001. Local Authorities Transfer Fund annual report FY 2000-2001. Nairobi: Republic of Kenya.

Republic of Kenya. 2002. Local Authorities Transfer Fund annual report FY 2001-2002. Nairobi: Republic of Kenya.

Republic of Kenya. 2003. Concept paper on decentralisation. Nairobi: Planning Department, Republic of Kenya.

Republic of Kenya. 2004. Strengthening district level M&E and health management information systems (HMIS) in Phase 1 districts. Nairobi: Ministry of Health, Republic of Kenya.

Salzer, Walter. 2003. The untapped potential: self-governing communities for public service provision in rural areas, the Kilifi experience. Available on the Internet <http://www.gtz.de/urbanet/pub/salzer.pdf> (accessed July 2004).

Semi-Arid Rural Development Programme. 2002. SARDEP approaches to development: a compendium of experiences. Volume 2: administering support for community-driven development. Available on the Internet <http://www.snvworld.org/Kenya/PublicaMain.htm>.

SIDA. 2003. Field notes of a visit to seven SIDA-supported districts (June 8-20).

SIDA. n.d. Support to health services. (Brochure on SIDA programme). Stockholm: Swedish International Development Agency.

Tamm, Gordon. 2002. Report on rural integrated health services in Kenya. Stockholm: Swedish International Development Agency.

UN-Habitat & SIDA. 2003. An analysis of the development initiatives in the Lake Victoria region (EA): Musoma, Kisumu, Kampala. Nairobi/Stockholm: UN-Habitat/Swedish International Development Agency.
UN-Habitat. 2003. Global Urban Observatory: better information for better cities. Website <http://www.unchs.org/programmes/guo> accessed July 2004.

UN-Habitat. 2003. Lake Victoria Region City development strategies (CDS) for improved urban environment and poverty reduction: report of a sub-regional workshop held in Musoma Municipality. Nairobi: UN-Habitat.

Van Oosterhout, Fon. 2002. Moving targets: towards monitoring democratic decentralisation. (Royal Tropical Institute Bulletin 349). Amsterdam: Royal Tropical Institute (KIT).

Verhoeven, Dolly. Aid, a changing necessity: SNV from volunteers to advisers. The Hague: Netherlands Development Organization.

World Bank (Kenya Country Office) & DFID. 2002. An assessment of local service delivery and local governments In Kenya. Nairobi: World Bank/Department for International Development.
World Bank (Kenya Country Office). 2002. Kenya. Community-driven development: challenges and opportunities. Nairobi: World Bank.

-----------------------------------------------------------------------------------------

Notes

1. Usually referred to as the Majimbo constitution, which was a result of negotiations among the nationalist forces before independence. The colonialist provided logistical support to the negotiators in London.

2. Policy paper issued by government to guide policy developments. Similar to a White Paper.

3. The district commissioner is a presidential appointee with oversight over a single district. The commissioner is thus essentially a representative of the president - a more or less unelected sheriff who oversees all matters - though principally interested in security. The district development officer is an employee of the Ministry of Planning and has oversight over development coordination in the district. Other ministries, e.g. agriculture, health and education, would also have a district head, e.g. the district health officer and the district education officer. The DDC also co-opted representatives of parastatals - government-owned companies and institutions undertaking development work in the district (e.g. the regional development authorities).

4. These councillors do not represent electoral wards, but are nominated by parties in positions originally established to cater for special interests and under-represented groups, but which have since been abused to reward supporters and (usually local mobilisers for) national-level politicians.

5. The LATF is discussed in Chapter 4 under Support to Local Government Improvement.(usually local mobilisers for) national-level politicians.

6. The district commissioner is a rank below the provincial commissioner and is administrative head of a smaller geographical area.

7. However, the constitutional reform process has been the only one with an expectation of institutionalised wide consultation.

8. Attendees at workshop on city management that the author facilitated during the study period. Though the meeting was on a different subject there was a lot of discussion on local government autonomy and the LATF which proved useful in preparing the study.