This paper should be cited as: Watson, D. 2005. Capacity building for decentralised education service delivery in Ethiopia. (ECDPM Discussion Paper 57H). Maastricht: ECDPM. http://www.ecdpm.org/dp57H

David Watson
Accredited Governance Consultant
Department for International Development (DfID)

In association with
Lissane Yohannes, Associate HELM Corporation

The UK Department for International Development (DFID) financed the field work for this paper

A case study prepared for the project ‘Capacity, Change and Performance’

Study of Capacity, Change and Performance: Notes on the methodology

The lack of capacity in low-income countries is one of the main constraints to achieving the Millennium Development Goals. Even practitioners confess to having only a limited understanding of how capacity actually develops. In 2002, the chair of Govnet, the Network on Governance and Capacity Development of the OECD, asked the European Centre for Development Policy Management (ECDPM) in Maastricht, the Netherlands to undertake a study of how organisations and systems, mainly in developing countries, have succeeded in building their capacity and improving performance. The resulting study focuses on the endogenous process of capacity development – the process of change from the perspective of those undergoing the change. The study examines the factors that encourage it, how it differs from one context to another, and why efforts to develop capacity have been more successful in some contexts than in others.

The study consists of about 20 field cases carried out according to a methodological framework with seven components, as follows:

• Capacity: What are the components of capacity, i.e. the capabilities, in a given context?
• Endogenous change and adaptation: How do processes of change take place within an organisation or system?
• Performance: What has the organisation or system accomplished or is it now able to deliver? The focus here is on assessing the effectiveness of the process of capacity development rather than on impact, which will be apparent only in the long term.
• External context: How has the external context – the historical, cultural, political and institutional environment, and the constraints and opportunities they create – influenced the capacity and performance of the organisation or system?
• Stakeholders: What has been the influence of stakeholders such as beneficiaries, suppliers and supporters, and their different interests, expectations, modes of behaviour, resources, interrelationships and intensity of involvement?
• External interventions: How have outsiders influenced the process of change?
• Internal features and key resources: What are the patterns of internal features such as formal and informal roles, structures, resources, culture, strategies and values, and what influence have they had at both the organisational and multi-organisational levels?

The outputs of the study will include about 20 case study reports, an annotated review of the literature, a set of assessment tools, and various thematic papers to stimulate new thinking and practices about capacity development. The synthesis report summarising the results of the case studies will be published in 2005.

The results of the study, interim reports and an elaborated methodology can be consulted at www.capacity.org or www.ecdpm.org. For further information, please contact Ms Heather Baser (hb@ecdpm.org).



Contents

Acronyms
Summary
1 Introduction
2 Devolution policy and the public sector in Ethiopia
3 Previous experiences of capacity building
4 Key capacities in the context of devolution
5 The influence of external contextual factors
6 The influence of external interventions
7 Stakeholders in devolved education service delivery
8 Internal management factors affecting capacity
9 Capacity, change and performance in Ethiopia
Appendix: List of individuals interviewed
References
Notes


Acronyms


BESO Basic Education System Overhaul (USAID)
BoFED Bureau of Finance and Economic Development
CBO capacity building officer
CSC Civil Service Commission
CSO civil society organisation
CSRP Civil Service Reform Programme
DfIDE Department for International Development Ethiopia
DiP Development in Practice (review by DfID)
DSA Devolution Support Activity (USAID)
ECSC Ethiopian Civil Service College
ELIP English Language Improvement Programme
EMI Ethiopian Management Institute
EPRDF Ethiopian Peoples' Revolutionary Democratic Front
ESDP Education Sector Development Programme
E&TB Education and Training Board
GoE government of Ethiopia
GTZ Deutsche Gesellschaft für Technische Zusammenarbeit
HEI higher education institution
HESO Higher Education System Overhaul
HRM human resources management
IST in-service training
JRM joint review mission
MoE Ministry of Education
M&E monitoring and evaluation
MoFED Ministry of Finance and Economic Development
MCB Ministry of Capacity Building
NGO non-governmental organisation
PRSP Poverty Reduction Strategy Paper
PSCAP Public Sector Capacity Building Support Programme
PTA parent-teacher association
RCBB regional capacity building bureau
REB regional education bureau
RMI regional management institute
SDPRP Sustainable Development and Poverty Reduction Programme
Sida Swedish International Development Agency
SNNPR Southern Nations Nationalities and Peoples Region
SWAp sector-wide approach
TA technical assistance
ToR terms of reference
USAID United States Agency for International Development
VSAT very small aperture terminal
WEO woreda education office


Summary

This report examines capacity building for decentralised education service delivery in Ethiopia. It is one of over 20 case studies connected with the project 'Capacity, Change and Performance' being compiled by ECDPM for GovNet, the Network on Governance and Capacity Development of the OECD's Development Assistance Committee. It attempts to apply the methodological framework adopted for the studies to consider the capacity building challenges posed in improving the delivery of devolved education services in Ethiopia. Two other related papers have been produced - an equivalent study in Punjab province of Pakistan, and a comparative analysis of the two cases, in which conclusions are drawn about the relative significance of the issues raised. DfID has sponsored this set of cases within the GovNet brief.

Devolution and education service delivery in Ethiopia
Ethiopia is a low-income country with a GNP per capita of only $100 in 2002, and a Human Development Index ranking of 169 out of 175 countries. Almost 40% of the population are illiterate, and only 24% of pupils complete primary school education. For the present government, which came to power in 1991 after 16 years of bloody civil war, education is now a national development priority: nearly 14% of government budget is allocated to it. The Education Sector Development Programme (ESDP) was introduced in 1997 as a vehicle for implementing the 1994 education policy, which envisaged universal primary education by 2015.

The devolution of service delivery was included in the 1995 Constitution, which provides an unusual degree of autonomy to Ethiopia's 11 regional states. In 1999-2000, however, studies by the World Bank depicted an administration more akin to deconcentration below the regional level. Regional state governments tended to dominate service delivery at the woreda and kebele levels.

After a crisis within the ruling party (EPRDF), which culminated in a policy 'renewal' in 2001, more genuine devolution took place. Full discretion was given to elected woreda councils to allocate unconditional grants from regional state treasuries (initially in the four largest regions). At the same time, a Ministry of Capacity Building was created, charged with implementing the national capacity building plan at all levels of government, and across all sectors, including education.

Capacity building strategy, interpretation and experience
The national capacity building strategy has three elements: human capacity, systems and procedures. It also embraces gemgema, an indigenous system of appraising individual contributions to their organisation's goals, and relationships with colleagues. There have been few major (bilateral) capacity building efforts related to devolved service delivery: one focused on financial management, and another on relieving the chronic capacity constraints facing education planners, teachers, and the thousands of parent-teacher associations. One initiative, the Public Sector Capacity Building Support Programme (PSCAP), backed by a consortium of donors coordinated by the World Bank, is related directly and indirectly to devolved education services.

While there are no impact analyses of these programmes, they appear to have contributed to capacities at various levels, and there is some evidence of positive performance outcomes (especially financial management) due to the systematic approaches adopted. There is no evidence that the results of, and lessons emerging from, these initiatives are being shared. DfID funded an inventory of capacity building challenges, experiences and capacities in Ethiopia and elsewhere, but it was not given the exposure planned. It therefore appears that Ministry of Capacity Building has not yet 'learned from experience'.

Capacity challenges of devolved education services
There are major challenges at all levels, with indications of capacity constraints, functional overload and detachment from local realities at the federal level, particularly in the Ministry of Education. The Ministry is clearly still finding it problematic the make the transition to its policy, planning and evaluation role under devolution. It is still mired in implementing a number of bilaterally funded projects for which it remains responsible. The Ministry of Capacity Building apparently exhibits capacity constraints to a somewhat lesser degree, although many staff are inexperienced.

At the regional and woreda levels, although capacity constraints are still real, the pressures of political accountability are keenly felt, and appear to be acting as a spur to improve performance despite resource constraints. While attempts are clearly being made to make devolution work, there is little evidence of how levels of local government are responding (and performing) under devolution.

There appears to be a reluctance to examine the implications of devolution for federal Ministry roles, structure, staffing and priorities. There are indications of generic (federal civil service) and specific (sector-specific) capacity weaknesses. Information management is emerging as a major issue: disaggregated budget data is lacking, and there is no reliable information on staff turnover especially at woreda level.

Donors
The 'balance sheet' regarding donor engagement has positive (capacity enhancing) and negative (capacity draining) features. On the positive side, USAID has supported two systematic and long-term efforts that have begun to have positive effects on education service delivery. A World Bank-led consortium of donors has helped to prepare a nationwide programme of demand-led capacity building, and DfID has sponsored a study of comparative approaches to local government capacity building in Ethiopia and elsewhere. Increasingly, donors are providing aid in the form of budget support, and encouraging evaluation through programme reviews and the production of 'tools' to help implementation. On the other hand, despite the existence of formal coordination mechanisms (including one devoted to education), and some budget support, donors have retained bilateral 'projects' running parallel to the ESDP, and have retained control of significant funds for TA within the supposedly 'pooled' funding arrangements for the new programme.

Conclusions
It is not feasible to arrive at firm conclusions about the relationship between capacity, change and performance in Ethiopia due to the lack of data about some aspects of the performance of the system. Although devolved arrangements in their present form have been in operation for only three years, it is clear that the operational situation and working environment at the woreda level have changed dramatically since 1999-2000. Reform at regional and woreda levels is being seriously attempted.

Economic deprivation and severe resource constraints impinge negatively on key capacities, and especially on the number of teachers that the country can afford. However, there are significant contextual factors that provide a favourable environment for additional investment over time. These include:

• The autonomous status of the regional states is enshrined in the constitution, and their roles (and those of lower-level governments) provided in detailed guidelines.
• There is a long-standing education policy and broad political consensus that education is important for nation and democracy building, that devolution must be made to work, and that local structures have an important role to play. Education is seen as a central plank in the efforts to build capacities.
• There is a distinct national capacity building strategy and politically empowered administrative structures at all levels to implement it.
• Unconditional fiscal devolution puts resources at the disposal of local governments, and ensures that they are accountable to their constituents for their decisions and actions. There is strong local accountability, with education 'watchdog' bodies at woreda and kebele levels, and a high 'density' of elected council representatives (one for approximately every 200 people).
• Community contributions to the education infrastructure and even supplies directly support the 'physical' capacities. As direct 'investors' in education, communities have a strong stake in ensuring that the state delivers in terms of professional management, high standards of teaching and relevant curricula.

HIV/AIDS is having an increasingly significant, negative impact on efforts to develop and maintain education service capacities at all levels, yet Ethiopia lacks the capacity to combat it effectively.

1 Introduction

Ethiopia is a low-income country with a GNP per capita of only $100 in 2002, and a Human Development Index ranking of 169 out of 175 countries. Almost 40% of the population are illiterate, and only 24% of pupils complete primary school education. For the present government, which came to power in 1991 after 16 years of bloody civil war, education is now a national development priority; nearly 14% of the government budget is allocated to it. The Education Sector Development Programme (ESDP) was introduced in 1997 as a vehicle for implementing the 1994 education policy, which envisaged universal primary education by 2015.

This report examines capacity building for decentralised education service delivery in Ethiopia. It is one of over 20 case studies connected with the project 'Capacity, Change and Performance' being compiled by ECDPM for GovNet, the Network on Governance and Capacity Development of the OECD's Development Assistance Committee. It attempts to apply the methodological framework adopted for the studies (see inside front cover) to consider the capacity building challenges posed in improving the delivery of devolved education services in Ethiopia. Two other papers have been produced, related to the present one: an equivalent study in Punjab province of Pakistan, and a comparative analysis of the two cases, which assesses the significance of the issues raised. DfID has sponsored this set of cases within the GovNet brief.

At a series of meetings in late 2003 ECDPM and DfID Policy Division agreed to look at education service delivery in two decentralising but very different contexts - Pakistan and Ethiopia. This focus corresponded to DfID's concerns about pro-poor service delivery. These cases are the only ones in the DAC sample that analyse public service capacity issues in the context of public service delivery and decentralisation.¹

The approaches adopted in the Ethiopia and Pakistan studies were similar, including the preparation of the ToR with DfID Ethiopia and its counterparts in education services, the recruitment of a national consultant (Lissane Yohannes) with experience in the sector, and arranging meetings and visits by Mr Yohannes. This proved to be far from straightforward, given the autonomous nature of regional and local (woreda) governments, and the degree of persuasion (not direction) required of the federal Ministry of Education with the heads of education in the regional state of Amhara. Mr Yohannes spent three days negotiating with and briefing regional stakeholders before the fieldwork could begin.

The fieldwork was carried out in July 2004. An issues paper was discussed with DfID Ethiopia at the start of this phase. Meetings were held with many individuals at federal, regional and woreda levels (see Appendix). It was decided to visit four woredas, but Mehal Meda - selected as a successful example of decentralised governance - could not be visited due to lack of time, and was replaced by Debresela.




2 Devolution policy and the public sector in Ethiopia

2.1 Historical and political background

This section outlines the main historical 'milestones' in Ethiopia's history that have a strong bearing on the context for public sector capacity building and change (see Polhemus and Yohannes, 2002; GoE, 2004b).

In 1974 Emperor Haile Selassie was deposed by the military Derg regime. During the period of the Derg, an armed resistance movement formed a coalition, the Ethiopian People's Revolutionary Democratic Front (EPRDF), which eventually toppled dictator Mengistu Haile Mariam and his government in 1991, after nearly 16 years of brutal autocratic misgovernment. The interim transitional government paved the way for a constitution of the new state of the Federal Democratic Republic of Ethiopia, which came into being on 21 August 1995.

The constitution provided for nine regional states (plus two urban authorities) with equal rights and legislative, executive and judicial powers, including the right of secession (Article 39.1). The form of government is parliamentary and democratic. The regional states have their own constitutions, and, because they were established on the basis of 'nationality', they vary enormously in size and population. Amhara regional state, in the northwest, has the second largest population (over 17 million out of a total of nearly 70 million), but the third largest in area. It has 105 autonomous local government authorities termed woredas. Amhara and two other regions (Oromia and Southern Nations Nationalities and Peoples' Region, SNNPR) account for over 80% of the population and are each larger than at least 20 of Africa's independent states (GoE, 2004b, ch.2 p.12).

The devolution of power and resources from federal government, and the reform of public sector institutions, has taken place in two phases. The period 1995-2001 saw 'deconcentration' from regions to woredas, although regional governments prepared woreda budgets, and imposed restrictions on how transfers were to be used. Woredas were thus obliged to follow regional government priorities and development programmes. From 2001, a crisis in the ruling coalition resulted in a process termed 'renewal' (Polhemus and Yohannes, 2002), with devolution of power from the regions to the woreda councils, who were then able to set their own development priorities and budgets. They continued to depend on regional governments for well over 80% of their revenue, although since 2001 these transfers have been in the form of unconditional block grants. Below woredas, kebeles are the lowest rung of democratically elected local government (see Figure 2).²




The federal government was reorganised in 2001. An Office for the Coordination of Capacity Building (soon converted into a ministry) was established to initiate capacity building polices, to design and implement related programmes in support of the regions, and to coordinate other related organs of government, including the Ministry of Education.

2.2 The education sector

The Education Sector Development Programme (ESDP I) (1997-2002) was designed to address the very poor conditions and performance of the education system, in the context of widespread poverty.³ The primary school enrolment rate (30%) was the lowest in the world (and less than half the average for the region). It had fallen as low as 19% during the later years of the Derg regime (GoE, 1999). Change was necessary to address chronic problems, particularly in rural areas, where less than 30% of boys and 20% of girls attended primary school. One-third of grade 1 pupils dropped out, and 20% of the remaining pupils repeated classes (taking nine years on average to complete six years of primary education). Inadequate facilities, the lack of textbooks, under-qualified teachers (less than 40% of grade 9-12 teachers were qualified), and inappropriate curricula and examinations were the main causes. In 1992/3 government spending on education was just 2.6% of GDP, rising to 3.8% in 1995/6, which amounted to 13.7% of the budget. Planning and management capacities, particularly at regional and woreda levels, were described as 'weak'.

The ESDP I attempted to put into action the 1994 education policy, which included a sector plan, backed by a consortium of donors, designed to enhance enrolment, particularly in primary schools, decentralisation, and community empowerment. ESDP II was launched in 2002/3 and will run to 2004/5. An annual joint review process evaluates progress against performance milestones. The ESDP Action Plan stressed that 'greater institutional capacity at all levels ... is the first priority of the programme' (GoE, 1999, p.9). Regions considered especially ill-prepared 'will concentrate on capacity building in the initial period of ESDP, especially in the areas of educational administration, procurement, financial management and M&E. Head teachers and officials at all levels will be given training in education management, and a reporting and monitoring system will be put in place to facilitate their accountability. ... In the MoE planning, finance management, implementation and monitoring and evaluation capacities will be improved' (ibid., p.10) The Secretariat of the ESDP is the Planning and Programme Department of the Ministry of Education.

Another major innovation since the launch of the ESDP was the introduction of the Sustainable Development and Poverty Reduction Programme (SDPRP) in 2002, as Ethiopia's version of a Poverty Reduction Strategy Paper (PRSP). Three of its four 'pillars' relate to governance changes (the other is agricultural development-led industrialisation), including reform of the justice system and the civil service, decentralisation, and capacity building in the public and private sectors. Devolution is seen a means to several ends: a socio-economic transition to democracy and better governance, improved service delivery (by shifting decision making closer to the grassroots for improved accountability and responsiveness), and citizens' empowerment and participation in governance.

3 Previous experiences of capacity building

3.1 National capacity building strategy

A national capacity building strategy was developed as part of the Civil Service Reform Programme launched in 1998, and a plan for implementing it was put in place following the 'renewal' of the EPRDF in 2001, to be led by the newly created Ministry of Capacity Building. The strategy's three elements - human capacity, systems and procedures, and organisational structures and interrelationships - are based on the understanding that the country will only achieve its goals if it can coordinate the use of its human resources, institutional capacity and the procedures to carry out these tasks (see Box 1).



3.2 Indigenous approaches: gemgema

An important and widespread feature of capacity building practice in Ethiopia is gemgema, a traditional approach to encourage performance based on regular group feedback to individuals regarding their performance and behaviour in relation to the group and the objectives of the organisation. It was developed during the period of armed struggle (1975-91) against the Derg, first by the Tigray People's Liberation Front and then by other allied groups during the civil war, and has gradually begun to be practiced by all branches of the civil service in Ethiopia.

Although work on formal performance evaluation systems is ongoing in the civil service, gemgema is complementary to them - it does not involve written assessments, it is regular (several times per year) but not arbitrary. Gemgema is based on objectives, tasks accomplished, problems encountered and solved, targets and their achievement, interpersonal communication and attitudes to the group. The objective is to improve the effectiveness and productivity of the group in relation to the organisation's aim or mission. It is widely practiced, and its outcomes are taken seriously by all involved.

To our knowledge there has been no formal assessment of the impact of gemgema. It is, however, symbolic of the perceived importance of the contributions of all individuals to achieve the purpose of the organisation for which they work, and through it, the socio-economic development of the country. This feature of working life is pervasive, transparent and apparently unique.



3.3 Capacity needs analyses

There have seen several major principal analyses of capacity needs and requirements related to education services in Ethiopia include the following:

'Woreda studies' (World Bank, 2001)
One of the most ambitious and wide-ranging diagnostic studies of local government operations and services was undertaken in 1999-2000, with the assistance of the World Bank. It formed the basis of future government/World Bank collaboration in a range of programmes, in particular the Capacity Building for Decentralised Service Delivery project, a precursor to the Public Sector Capacity Building Support Programme (PSCAP, 2004-8).

The study was preceded by a five-month pilot assessment to test methodologies and develop hypotheses. The main work was done in nine woredas in four regions that were selected to highlight the main contrasts in the economic system: urban/rural, food secure/insecure, and agrarian/pastoral, focusing on the processes of governance, and perceptions thereof, at woreda level. Fieldwork was undertaken at all levels of government, from region to sub-kebele, in six sectors: health, education, water, agriculture, finance and planning. The study's findings included:

• the enormous magnitude of the task of providing basic services in the geographical and infrastructural environment of Ethiopia;
• federal and regional governments were still applying a hierarchical approach to development, creating a highly centralised system of governance;
• resource management was firmly controlled by the bureaucracy;
• elected representatives focused solely on local tax collection and contributions of labour; and
• citizens were passive beneficiaries of services, rather than active participants in their management, so that the system did not capture the resourcefulness, resilience and adaptability of local communities.

The report recommended actions to decentralise power progressively to elected bodies, and to enhance community control of development resources.⁴

The Education Sector Development Programme (ESDP I and II) (1997-2002-2005)
The ESDP was preceded by joint design missions that assessed all dimensions of education service delivery capacities, including current deficiencies, gaps and priority actions to bridge them. However, the programme action plan did not include capacity deficiencies, or measures to address them (GoE, 2002a). Only 3% of the indicative Birr 3 billion budget was to be devoted to capacity building activities.

National Capacity Building Programme (1998-2003-2008)
Early experiences with a public sector reform programme introduced by the government in 1997,⁵ and a series of donor assessments, revealed several underlying capacity challenges. These led to the preparation of a national capacity building strategy and plan in 1998 (see Box 1), and the creation of the Ministry of Capacity Building (MCB) to lead its implementation in 2001. The challenges identified were:⁶

• Inefficiencies deriving from unpredictable financial management, poor incentives, and the lack of a strategic or performance orientation. At the federal level problems included the lack of compliance with the financial budget calendar; resource allocation and expenditure frameworks; political ownership of the macroeconomic fiscal framework; and reporting and audit systems. At regional and woreda levels, there were problems connected with bringing accounts up to date, and budgeting and procurement systems.
• Personnel incentives were weak. A 'results-oriented performance evaluation' system had been introduced in the Civil Service Reform Programme (CSRP) in 1996, but there were cultural and operational difficulties with its 'roll-out', pay negotiations had stalled, and the rules for hiring, transfer, promotion and firing staff at the woreda level were ambiguous. There was evidence that qualified managers and IT staff were leaving to work for NGOs and the private sector.
• Empowerment of woredas and municipalities required granting them more fiscal and administrative autonomy. The inefficiencies identified in the woreda assessments (see above) were compounded by over-centralised procurement practices, and weak personnel management.
• Weak checks and balances on the executive, which limited accountability, recourse and redress.

Development in Practice (DiP, 2003)
This comparative analysis of capacity building approaches for local government was commissioned by DfID in order to inform the preparation of the PSCAP (especially its district-level decentralisation component), by reviewing lessons from earlier attempts at local government capacity building in Ethiopia and elsewhere. It pointed out that organisational approaches to training across woreda administrations (providing for guided experience in the workplace, as opposed to off-the-job) had never been employed in Ethiopia, and questioned the providers' ability to meet new needs for generic woreda capacity building. Organisational development approaches had proved effective elsewhere.⁷ It also drew attention to ambiguities in the relationship between regional and woreda governments, in particular to an apparent need for a dedicated bureau at regional level to regulate woredas. It appears that this valuable and comprehensive study was never discussed (as planned and proposed by DfID) among stakeholders in the PSCAP preparation process and other
development partners.

Higher Education System Overhaul (HESO, 2004)
This assessment of higher education (HE) management capacity was one of the most thorough and participatory of the needs analyses reviewed here. The committee of inquiry (including senior staff of HE institutes and the MoE, and internationally experienced (volunteer) staff) focused on leadership, governance and management of the sector. Major foci for change identified were: autonomy and accountability; promoting an enabling culture; management development and training; human resource management; quality of inputs, processes and outcomes; financial and resource management; administrative systems and the challenges of HIV/AIDS. The report made recommendations for action by the MoE, and the boards and senior managers of HE institutes.

3.4 Examples of capacity building efforts in Ethiopia

EDSP I was launched at the same time as the first phase of the Civil Service Reform Programme (CSRP I) in 1997. The latter comprised components concerned with top-level coordination and change management, resource management and control, performance improvement and appraisal, transparency, and service delivery. In the period up to 2001 the CSRP focused on research and preparing for implementation, although this was taken forward only after the 'renewal' in 2001. Since then the major innovation has been the creation of the Ministry of Capacity Building to lead the implementation of the national capacity building strategy.

Within the CSRP's financial management and expenditure control component (the only one active before 2001), one initiative has had a major impact on capacities for financial devolution in the four main regional states. The Decentralisation Support Activity (DSA, 1997-), funded by USAID (and latterly Irish Aid) addresses financial management, accounting and budgeting capacities for decentralisation, and is based in the Ministry of Finance and Economic Development. No recent overview of its activities is available, and the steering committee has apparently not met for two years, but the aggregate results of its efforts are impressive. A total of 23,000 staff at federal, regional and woreda levels have been trained since late 2001, nearly 5000 of them in accountancy, involving eight regional management training institutes and the Ethiopian Civil Service College between 1998 and 2001. The project has produced more than 80 reports and manuals on accounting issues, 120 on budget reform, five on cash management, and 52 on public investment and expenditure reform, as well as policy advisory notes, training manuals, IT-related publications and media reports. The project also facilitates discussions of reform experiences between participating regions, and comments (sometimes critically) on current (allegedly over-hasty) approaches to reform in these fields.⁸

The Basic Education System Overhaul (BESO) is USAID-funded capacity building programme that aims to improve equity of access to education (especially for girls), and the quality of (particularly) primary education. It has three elements:

• The community/government partnership programme focuses on improving links between communities and schools (via parent-teacher associations, PTAs) and includes incentive grants (managed by the PTAs) together with assessment and planning skills training aimed at empowering PTAs to make them 'informed users' of education services.
• Teacher development: 20 teacher training institutes have received support (about Birr 1 million each, plus libraries and computer equipment) with an emphasis on active learning methodologies, communication skills and management of large classes. In-service training for clusters of schools is also provided, focusing on leadership development and tackling HIV/AIDS.
• Planning and information management skills development. In all 11 regions electronic personnel records and systems have been installed; the emphasis has now swung to woreda education offices with the recent completion of a country-wide woreda capacity building programme. In collaboration with Addis Ababa University, a 16-module 300-page workbook, in four languages, has been developed with and for woreda planning officers. Training of trainers for the 'roll-out' of the training was provided for regional education planners. Teams of staff from each of the 611 participating woredaswere invited to bring their own data to subsequent 10-day courses, and they were then guided through the process of drawing uptheir own plan based on real data. This approach was well received by the more than 3000 participants.

A major teacher development programme is under way. This comprises a 'Teacher Education System Overhaul' (TESO) and a management training element, as well as the English Language Improvement Programme (ELIP) for teachers. The latter, supported by a consortium of six donors, is designed to improve the competence of 140,000 teachers of English to children in grades 5-12. It is currently the subject of the only impact evaluation we encountered in Ethiopia.

Alternative Basic Education Programme Amhara. This (rare) example of successful cooperation between an NGO and a regional education bureau is designed to provide informal opportunities for children from poor or isolated communities to obtain an education.

The Area Development Programme in Amhara (supported by Sida) uses grants to encourage community contributions, with positive results.
From sources in the Civil Service Commission and the Minister of Capacity Building we understand that a series of management seminars is being organised for senior federal officials to encourage reform-mindedness. A major effort supported by DfIDE and other donors was to revise the existing ESDP programme implementation manual in early 2004, in response to the perceived need for a reference source for all stakeholders, which would provide clear and easily referenced guidance to policies, structures, funding mechanisms, institutional roles and management procedures underlying the ESDP. This replaced an earlier version of a manual produced in late 1998 ostensibly as a 'living document' (which died).

Based on interviews at the regional and woreda levels, it appears that the regional capacity building bureau (RCBB) in Amhara has taken responsibility for capacity building among education and training boards and PTAs, but that other education-related elements had been left to education bureaux or offices to take forward, with or without federal and/or development partner support (for example under the programmes mentioned above). No documentation was available to assess the extent of these programmes, or their approaches.

3.5 Lessons, themes and contrasts

The main points emerging from the capacity building experiences described above appear to be as follows:

• Devolution to the woreda level is a recent phenomenon; before 2001 the style was more 'deconcentration' from the regions to woredas. Since 2000, when the 'Woreda Studies' were compiled, there have been major changes in the behaviour of regional and woreda government actors.
• All the major capacity building efforts (except the management seminars) have been funded by, and technically supported by donors.
• There have been no formal objective 'third party' evaluations of capacity building practices. There are preparations for an impact analysis of the ELIP, and there are plans to evaluate BESO. It is therefore unclear how the programmes have contributed to the changes in behaviour noted above.
• The relationships between these initiatives and the Ministry of Capacity Building (MCB) or its regional offices (which supervise education bureaux and woreda education offices) have been polite but distant. In the MCB documentation we were able to access there are almost no references to the major programmes described above. In regional and woreda capacity building offices, staff have participated in most of the training. This apparent 'distance' between the MCB/RCBB and the MoE/REB may relate to the reluctance of the two ministries to cooperate - despite the seniority of the Minister of Capacity Building in cabinet, the Minister of Education is a senior in party circles (being a member of the politburo), and is directly accountable to the Prime Minister. It may also relate to the history of the evolution of the PSCAP, in which regional officials were involved in compiling elaborate cross-sectoral capacity building programmes (but which were ultimately unfundable given the eventual scope of the World Bank programme), but apparently did not involve education or health sector staff.

Feedback on the USAID-supported Devolution Support Activity (DSA) and BESO programmes has been positive:⁹

• The programmes' consultants and trainers have long experience of Ethiopian realities, and sound knowledge of the situation facing trainees (based on needs analyses and frequent visits to work sites);
• regional offices are involved in programme design and preparation;
• formal off-the-job training for woreda staff is followed up (at least in the case of the DSA) by periodic on-site assistance from local support offices;
• These are timely interventions, in that they relate to parallel systemic changes brought about or encouraged by the federal government. For example, cost-centre budgeting and double-entry book-keeping have been introduced in Amhara. The training has been relevant to woreda staff, who are under pressure to account more promptly and accurately for development funds.
• The training is practical. For example, the BESO educational planning training invited participants to bring their own data so that they were able to produce something (an action plan) that would be of direct use to them on return to work.
• In both cases, teams of colleagues attend the same training, so that a 'critical mass' of graduates returns to the workplace.

3.6 Capacities for capacity building

The issue of capacities for capacity building has been addressed during the preparation of PSCAP. The DiP report (2003) lists the main (public sector) capacity building centres, together with details of current and past activities, as follows:

• The MCB supervises the Ethiopian Civil Service College (ECSC) and the Ethiopian Management Institute (EMI). The MoE is jointly accountable to the MCB and the Council of Ministers via the Prime Minister. Regional capacity building bureaux supervise regional education offices and management institutes.
• The ECSC provides a wide range of management programmes for federal and regional staff related to the implementation of the CSRP (supported by DfID).
• The EMI trains private and public sector managers (60% from federal or regional governments) - a survey in the four main regions concluded that management and leadership capacities at woreda level are very scarce.
• Neither the ECSC nor the EMI have links with regional management institutes (RMIs), whose staff and capacities are reportedly weak, although the Amhara RMI offered CSRP training for nearly 2500 staff in 2004. The regional government is building a new facility in Amhara at considerable cost.
• There are no facilitators of organisational capacity building at woreda level, nor have any been trained in the past.¹⁰

We are not aware of any organised efforts to develop the capacities of Ethiopian consultants.

The dissemination of a MCB policy on capacity building for civil society and non-governmental organisations (CSOs/NGOs) was imminent but still pending at the time of our review (GoE, 2004d). One component of the national capacity building plan is devoted to civil society (see Box 3).

In an analysis of the impacts of Pact's past efforts in 24 NGOs, Fesseha (2002) found improvements of between 20% and 50% in seven key areas (governance, management, human resources, financial management, service delivery, external relations and sustainability). The impact was greatest in those cases where tailored training was followed by mentoring in the workplace (DiP, 2003, p.21).



4 Key capacities in the context of devolution

4.1 Roles and performance

This section examines the roles assigned to various levels of government (Box 4) in the devolution policy as the basis of our assessment of 'which capacities matter most'. We then explore aspects of the performance of the education system from several perspectives.




Based on our discussions and visits, the main changes introduced in the above levels of government through devolution (especially since 2001) include the following:

1. In August 2002 the Ministry of Education issued a comprehensive set of guidelines for educational management, community participation and finance GoE, 2002b).
2. A significant proportion of donor support (for education and other key sectors) is now delivered via budget support rather than projects, although the EDSP Secretariat still manages 32 bilateral projects, as well as the ESDP.
3. The annual joint review mission (JRM) appears to be a valued as a mechanism for monitoring progress and identifying implementation problems.¹¹
4. Block grants are in operation in the four largest regions. There is evidence that these have acted as a 'capacitator', in that woreda and kebele councils have to distribute scarce budget resources among competing sectors. No funding is earmarked at regional level before transfer to woredas. Compared with other decentralising countries in Africa, this is a rare characteristic (the norm is elaborate earmarking, which partially defeats the objective of devolution). There is evidence, however, that these grants are inadequate to fund much more than teachers' salaries, which accounted for 93% of grants in 2004.¹² Stakeholders at national level complain of the lack of disaggregated data on actual allocations to education at woreda level.
5. Councils, education and training boards at all levels, and school PTAs were active in the areas we visited. Political and community accountability aspects of devolution appeared to be in keeping with policy aspirations. Woreda education management is scrutinised by the E&TBs, while scrutiny of school management has been devolved to community level in the form of PTAs.

These changes have been accompanied by (but have not caused) a favourable trend in enrolment. The major factor influencing enrolment was the decision to scrap primary education fees in 1994; others included policy consensus, changes in parental attitudes, and the improved supply of teachers. Between 1992 and 2002 aggregate enrolment in grades 1-12 increased at an average of 9% per year (World Bank, 2004b), and the primary gross enrolment ratio trebled to 62%. Ethiopia's enrolment rates are now only slightly below those in other countries of sub-Saharan Africa. Enrolments in tertiary and vocational education grew from 3000 in 1995 to 54,000 in 2002.

4.2 Efficiency and performance in the education system

Efficiency and performance parameters relate primarily to the major policy issues raised in the review by the World Bank (2004b). Teacher-pupil ratios are clearly under stress, and have declined from 1:25 in 1992 (grades 1-8) to 1:65 in 2001 (1:52 in secondary schools). The average primary class size increased from 47 in 1994 to 75 in 2001. Conditions are worse in rural areas, where pupil-teacher ratios are 50-60% higher than in urban areas.
Other policy parameters identified in the World Bank study relate to the costs or the cost-effectiveness of the education system. It drew attention to relatively costly elements of the policy and the delivery system in relation to what is currently affordable:

• the length of primary education (8 years compared with 6 years elsewhere);
• the relatively high qualifications required to teach at various levels;
• the unit costs of classrooms (and the related issue of unrealistic building standards);
• the (low) share of expenditures allocated to primary education (50%, compared to a benchmark of 67% in the Education for All fast-track initiative);¹³
• the (high) proportion of expenditures devoted to teachers' salaries (much higher than in comparable countries); and
• the (inflated and unsustainable) per
• capitcosts of secondary and particularly higher education.

Cost factors were also related to geographical and human resource management factors. The distance between pupils' homes and schools was identified as a significant constraint on access. Whether the school could offer a complete teaching programme was another important variable (only 20% of primary schools offer classes up to grade 8), although this is more of a problem in rural than in areas. The cost-effectiveness of the system was also found to be related to the apparent randomness of teacher distribution (higher than in most countries for primary education as a whole, and the most random in a comparison of 21 African countries at grades 1-4).

The report's conclusions on (poor) learning efficiency were based on a MoE baseline assessment in 2000 that analysed grade 4 and 8 test (correct answer) scores: 41% was the average at grade 4, and 48% at grade 8. Thus large numbers of children were not achieving curriculum objectives, due to factors such as inadequate access to textbooks and radios for supplementary instruction. The report noted that at grade 8 'process' (management environment) rather than input variables were the most significant. There was virtually no statistical link between schools' recurrent expenditures and learning outcomes. The report concluded that 'money by itself is insufficient to achieve good results', and that the challenge is 'to discover the underlying support and incentive structures that motivate teachers and school directors, and adopt these ...' (World Bank, 2004b, p.xxv).

Community contributions to or expenditures on education are significant - in 1999/2000 they amounted to 30% of all government contributions, and more than 50% of that for primary education. There is evidence that these expenditures (not fee costs) impacted disproportionately on the poorest. For them the opportunity costs of sending children to school would be highest (in terms of lost labour) and the perceived importance of school least (for parents who are more likely to be illiterate and with no educational experience). These observations on community contributions were corroborated by the 2003 JRM, and by our own visit to Mota woreda in Amhara region where, in 2004, community contributions (in kind, labour and cash) added more than 60% to the public education budget.

4.3 Implications of devolution for 'capacities'

Within the federal MoE and related research institutions involved in educational policy, a number of capacities are clearly crucial, in particular the ability to identify and analyse policy options, costs and trade-offs. Relevant issues include minimising the costs (and maximising revenues) of secondary and higher education, management of higher education institutions (through more devolution), setting minimum teacher qualifications at various levels of the system, and identifying ways to recruit more teachers. Note that the analysis of efficiency and performance in the previous section was derived from a report compiled by the World Bank, not the MoE.
Devolution has clear implications for education resource planning and management capacities at the regional level. The major challenges at this level include developing capabilities for generating and analysing data, curriculum development, teacher distribution across woredas, and allocating resources. Capacity challenges appear to be greatest in 'getting the learning environment right' at school level, such as by providing guidance to woredas on the selection of head teachers.

At the woreda level, capacities at a premium include optimising teacher distribution, overall recruitment, selecting and training head teachers (with help from the region); implementing HIV/AIDS information campaigns (given the significant numbers of teachers absent or lost due to AIDS-related diseases); analysis of the spatial distribution of schools, and of schools offering a full teaching programme. The above role assignments, and the woreda staff we interviewed, emphasised woreda capacities and performance in mobilising community resources. We return to this issue in the discussion of incentives below, given its prominence as a criterion of 'excellence' in regional assessments of the performance woreda administrators.

5 The influence of external contextual factors

In this section we review the significance of various external contextual factors on capacities, learning and performance in the Ethiopian education sector. The apparently high degree of political consensus on the constitutional, legal, organisational and policy frameworks in which education services are delivered - at least since 2001 - has provided a significant and supportive context in which measures to promote the development of capacities could be applied. The principal features of this context include the following.

1. The constitutional arrangements for regional states provide a stable and secure base for building the capacities of institutional actors at regional and lower levels. National development priorities have not fundamentally changed since 1991, but were subject to 'renewal' within the EPRDF in 2001. They have not been challenged by any alternative political party machine or politically organised voice.
2. The dominant party has played a significant role as a source of development ideology and policy, and of mechanisms for the placement of individuals who are mandated to carry it out. This has been manifested in the priority attributed to capacity building and education, the creation of the MCB in 2002, and in ensuring that incumbents of senior posts in regional capacity building bureaux and woreda offices are party members who are committed to the policy. A (deliberately) related factor is the formal accountability of education (and in some regions health) service agencies to respective capacity building offices. Only capacity building heads have positions in regional and woreda executive cabinets.
3. The 1994 education policy states that education is essential for responsible citizenship and a basic human (and constitutional) right; it also refers to the need to include the education of girls. This policy has not been fundamentally changed, although strategies for putting it into practice have been subject to annual review by the government and development partners since the late 1990s.
4. The need for effective devolution was re-emphasised in 2001, and (unusually for Africa) was backed up with flexible financial resources. Although these were inadequate to meet all development needs, they were to be allocated by democratically elected woreda and kebele councils.
5. Since 2001 the emphasis on community empowerment and participation has also been backed up with institutional arrangements within the education sector (education and training boards at local levels, and PTAs for schools), which have 'teeth' (PTAs can recommend disciplinary action against a teacher). This may explain the willingness and (in the better-off regions) the ability of communities to contribute to the costs of education.
6. The combination of a dominant political coalition, the institutional empowerment of local management and 'watchdog' bodies, and the increased awareness of their members appears to have resulted in little or no politically motivated interference in the management of education personnel or of resource allocation among constituencies.
7. HIV/AIDs is having a major impact on the education sector. In its study of higher education, the World Bank (2003, p.50) noted that about 10,000 teachers are HIV positive. About 22% of teacher attrition is due to AIDS, and recruitment needs to increase by 16% annually to achieve the Education for All goals. Also, absenteeism due to AIDS-related diseases is likely to have a greater impact on education than mortality - the equivalent of 1035 teaching years were lost in 2002. These observations accord with our discussions, yet neither the 2003 JRM nor the recent education study (World Bank, 2004b) refers to the problem.
8. The extent of the final factor, corruption, has been the subject of at least two mutually inconsistent analyses.¹⁴ Our impression is that in at least this basic service sector it is not a major problem.

6 The influence of external interventions

In fiscal year 2002/3 external grants amounted to 21% of total government recurrent revenues; external loans and other capital assistance were 44% of the total budget (GoE, 2003b, Table 4, p.43). Over 30% of the 2002/3 education budget came from external sources (GoE, 2003c, Annex 5).

6.1 Donor-supported activities

On the basis of our discussions and review of the literature, we conclude that in the 1990s there was a history of ad-hoc donor-driven 'gap-filling' capacity building approaches in Ethiopia, which the Public Sector Capacity Building Programme (PSCAP) is meant to prevent in future. The Education Sector Development Programme (ESDP) was introduced in 1997 to shape a coherent externally supported programme relevant to capacity building for education services. This remains the principal external intervention in the process of building capacities in the education sector in Ethiopia.

Despite the introduction of budget support, the ESDP Secretariat still manages over 30 development projects, imposing a demanding burden on the MoE's capacities in areas such as procurement, project formulation, management procedures, and reporting on progress and expenditures to individual donors.

We have observed several donor-funded activities that illustrate the role of external agencies in supporting various functions related to building a more rigorous approach to capacity building. Examples of these activities include:

• Evaluation: The planned impact evaluation of in-service teacher-training in the English Language Improvement Programme (ELIP) is a rare example in Ethiopia. The comprehensive bibliography of the Devolution Support Activity (DSA) project, for example, did not refer to a project review or interim evaluation.
• Programme and project formulation, if undertaken in a collaborative manner, and reflects on the effectiveness of capacity building practice so far, can be helpful in building analytical and design skills. The PSCAP preparation process, for example, in which RCBBs were intensively engaged, took over 18 months.¹⁵
• Programme reviews: The annual joint reviews of the ESDP are the best examples of this type of support function. Future missions will focus on specific themes, reflecting the priorities of stakeholders. The 2004/5 review will examine how education budgets are formulated and spent at woreda level, and the problems in relaying data on local budgets to federal level.
• Production of 'tools' to support programme implementation: Examples include the EDSP II project implementation manual and the appendices to the PSCAP project document, which provide model ToRs for consultancies.
• Experiential 'stock-taking', or policy/sector analyses to inform policy dialogue and development. Examples include the studies by the World Bank (2004b) and the DiP (2003) of local government capacity building approaches. The impact of these types of support depend on their timing, the capacities of the principal (GoE) stakeholders to take forward dialogue and action on the study, and the extent to which the study was commissioned in response to initiatives or invitations from interested (GoE) stakeholders. The HESO (2004) report was initiated and primarily sponsored by GoE.¹⁶

6.2 Donor coordination

In the education sector there is coordination among a specialised donor (sub)group of the Development Assistance Committee (DAC), which includes DfID, the European Commission, Italy, the Netherlands Ministry of Foreign Affairs, Sida, UNICEF, USAID, and the World Bank. The ESDP is a sector-wide programme (SWAp)¹⁷ in which donors have agreed to support the policy and programme based on annual joint reviews. The PSCAP is also intended to be a SWAp, but there is evidence that some (bilateral) donors are reticent to commit all their resources to this 'pooling' mechanism.¹⁸

7 Stakeholders in devolved education service delivery

This section discusses the stakeholders in the education sector, in particular their relationship to service delivery capacities, and any apparent questions, achievements, doubts or issues arising in relation to that relationship.

7.1 Federal level

Constitutionally, the civil service is a federal institution. Regional state governments not only are governed by separate constitutions, but have their own independent civil service commissions and policies. The only limitation is that terms and conditions applied by regional governments should be of a comparable standard to federal conditions.

During our research there was little mention of federal-level reform, except for the Civil Service Reform Programme (CSRP). The only component of the CSRP that appears to have been active pre-2001 related to expenditure management and control. We found no indications of reform and restructuring of roles and functions at federal level following the devolution of powers to regions and woredas.¹⁹ All federal ministries have become accustomed to playing a role in the implementation of country programmes and projects. Devolution implies that these tasks are taken on by lower levels of government, yet there is no evidence of transfer of staff, exchange of staffing categories, nor 'downsizing' at the centre to tailor administrative and personnel structures to the more policy-oriented tasks.

Ministry of Education
On the basis of our observations, and those of our informants, there are unresolved weaknesses in capacities at federal level, especially in the Ministry of Education. There is also apparently some resistance to the notion of devolution whereby lower levels of government have decision-making authority and control over operational issues. Some regional observers with direct experience of several ministries' operations since 2001 felt that there was a dearth of knowledge within the MoE about regional and woreda level initiatives and realities, combined with a lack of caring and a 'culture' of service.
The extent and nature of the MoE's reform and restructuring efforts - past, present and future - are unclear, but we understand that a team is working on a plan for the Ministry, due to report in 2005. So far there have been no official consultations with the Ministry's stakeholders (e.g. regional governments and donors).²⁰ It is also unclear whether this review will address the fundamental professional and managerial capacity constraints in the new Higher Education Strategy Institute and Quality and Relevance Assurance Agency.

Ministry of Capacity Building
The MCB is structured along the lines of the national capacity building strategy (see Box 1). Much effort has gone into defining the PSCAP, which will provide financial and technical assistance to all levels of government to analyse deficiencies and build capacities on a 'demand-led' and/or 'challenge fund' basis. According to one experienced observer (outside government), however, there has been no internalisation (yet) of the notion of capacity building, and the understanding of the term is still limited. This gives rise to the question of how bids for capacity building activities will be framed and their merits assessed.

MCB officers see their role as to provide technical support to the regions, to identify (capacity) gaps, carry out relevant studies and disseminate the results. In order to understand the 'realities' at woreda level, the MCB has commissioned studies of the planning and budget system and financial transfer mechanisms, and is currently conducting one on human resource development and training needs. In future this will be extended to include a study of how human resources can be more effectively used at that level. The study will also assess the feasibility of merging woredas, the use of 'pooled' administrative and financial services in woredas, and ways to increase own-source revenues. An overall 'impact assessment' of devolution to date is also planned. The MCB's district-level decentralisation programme awaits completion of these studies before finalising its programme document.

A generic problem that has begun to arise is the lack of routine flows of information from the regions to the MCB. Studies such as those mentioned above are currently the only reliable way for the Ministry officials to learn about local implementation realities. For example, the MCB has no up-to-date data on staff turnover at woreda level (reportedly a problem in some regions, but not apparent from our visits).²¹

The private and NGO sectors
Although the private sector and NGOs are not yet significant suppliers of primary education, there is a small but growing number of private primary schools in the largest urban areas. The number of private secondary schools is also growing - from a larger base - also in urban areas. The MoE's Annual Abstract of Education Statistics (2002-3) lists 681 non-government primary schools out of a total of nearly 12,500 (about 5%); over 13% of senior secondary schools are 'non-government' (mostly private). NGO schools are playing a significant role in improving access to basic education for children in the poorest areas.



for footnotes ²³ and ²⁴ see bottom of page

7.2 Regional state level

The regional state level of government is - at least in the largest and better-resourced regions - an important testing ground for innovation (see Box 5).²² The regional councils demand accountability from the executive 'cabinet'. Regional government expectations of federal ministries, and the extent to which they are met, depend on which sectoral agencies are involved. The relationship between the REB in Amhara and the MoE appeared to be less satisfactory than that between the RCBB and the MCB.

7.3 Woreda level

At woreda level, we noted an apparently positive relationship between, and complementary roles of education and capacity building officers (CBOs). The former is subordinate and accountable to the latter, despite the relatively recent introduction of appointed woreda CBOs who tackle capacity building and orientation for the school/community interface. We witnessed the head of a woreda CBO working alongside the head of the woreda education office at a series of evaluation seminars organised during the school holidays.

There appears to be a functioning relationship between executive offices and elected councils, which is forged by 'cabinets' of heads of offices (led by the woreda administrator), which report to council members.²⁵ Considerable attention is given to budgetary matters because the council decides - on the basis of inputs from the kebele councils - on the distribution of resources (both recurrent and capital) to the woreda. This is a heavy responsibility.
Before our visits we had heard comments about the probable implications of procurement pressures or infrastructure projects in the ESDP. When we raised this capacity challenge with woreda staff, they did not recognise it, since the bulk of recent education infrastructure projects had been funded by community contributions, and had been supervised by communities themselves. They expected this community support to continue.

The other notable feature of woreda operations is the 'pooling' arrangements for administrative and financial staff, who serve several sector offices. This appeared to be a pragmatic move, born of necessity in view of the staff shortages that were clearly apparent.

7.4 Community level

Users/citizens
The Ministry of Education Guidelines (MoE, 2002) provide a comprehensive and ambitious institutional blueprint for woreda, kebele and community level bodies, including education and training boards and PTAs. We heard positive comments about their activities, and the impressive community contributions to education facilities they have helped to generate. The RCBB has taken the lead in mobilising woreda CBOs to begin developing the capacities of these bodies, to be maintained by the education officers thereafter.

Schools: head teachers and teachers
We were impressed by the fact that teachers willingly participated in evaluation seminars and group discussions during their holidays. There were, however, indications of resistance from some teachers, both to the recently introduced curriculum reform (requiring more child/learner-centred approaches), and to the emphasis being put on their role as community mobilisers. They could not see why they had to do more than teach.

These grievances may be related to the near-absence of effective in-service training for teachers. It is worth remembering that, compared to the (accelerated) role changes facing local authority and regional staff, who are regularly reminded of their accountability to locally elected representatives, teachers have had little guidance through the transition from their traditional roles and arrangements and being accountable 'upwards' to head teachers and education offices, to being accountable 'downwards' to parents and communities. They have also been reminded of their obligation to Ethiopian society to mould its future citizens. In order to do that, the GoE clearly assumes they need to know more about national development policies, and expects them to do more to explain them to their pupils.

The significance of head teachers and their leadership of teachers as factors affecting individual and school performance has apparently been recognised, albeit rather late. We understand that this target group and their skills are to be given prominence in future capacity building strategies within the ESDP.

8 Internal management factors affecting capacity

Apparent constraints at federal level
There are indications of inherent constraints in policy implementation capacities at the federal level, possibly caused by dysfunctional civil service structures (staff grading and training) related to routine administration. This observation is based on discussions with technical advisers in the MoE. After being asked by the Minister²⁶ to advise on and take forward the design of a major aspect of policy, they were surprised to find that they had no counterpart professionals to work with, and that there was no one to undertake basic clerical and administrative tasks in their offices.

One senior observer in the civil service (outside the MoE) acknowledged that there were several generic 'people-management' problems, including the lack of ability and willingness to delegate on the part of senior staff, and a lack of internalisation of the concept of planning, and the framing of initiatives in relation to the broad definitions of the functions ascribed to individual posts. Many civil servants are still essentially working reactively, rather than proactively. Based on these and other observations, we conclude that these apparent constraints on capacities at federal level have multiple dimensions, including:

• poor overall organisational leadership;
• the failure to define an organisational mission, leading to limited individual understanding and 'ownership';
• the reluctance to engage fully in the transition to new policy, advisory and evaluation roles under devolution, and the tacit perpetuation of project implementation duties (necessitated by the continuing bilaterally funded projects despite the ESDP);
• poor work assignment and organisation; and
• limited group and individual motivation, and poor attitudes, knowledge and skills.

Of these, only individual knowledge and skills are amenable to conventional training solutions. It is doubtful whether individually oriented 'capacity building' in the form of training courses will make an impact on what appear to be symptoms of the need for organisational development approaches. The DiP (2003) study concluded that not only was this tentative diagnosis broadly applicable, but that there was a dearth of organisational development experience and (therefore) a shortage of facilitators who could take forward this type of capacity building work in future. No capacity building institution in Ethiopia is currently engaged in developing organisational development facilitation skills.

Clear responsibility for capacity building
One positive factor is that the MCB has clear responsibility for coordinating capacity building efforts. As indicated above, the MCB would do well to study systematically the (considerable) capacity building efforts and experience in developing finance- and education-related systems and capabilities through the USAID-supported DSA and BESO programmes. This would help strengthen the assessment modalities and criteria for the selection of proposed capacity building schemes for the three levels of government as the PSCAP gets underway.

Reforms in human resources management
The Civil Service Reform Programme (CSRP) There have been several reforms (some still ongoing) of HRM practices in the CSRP. These include performance-oriented job descriptions, together with more open, and more functional performance appraisal mechanisms (beyond the gemgema system). Innovative criteria are now also used for assessing performance - e.g. the significance attached to community participation/contributions in assessing the performance of woreda administrators in Amhara.

Management of information between levels of government
An underlying, and largely unresolved management issue that is emerging in the relationship between levels of government is access to, and management of information. This may well be addressed by the proposed VSAT satellite data and video exchange system once it is connected. We saw that satellite dishes had been installed (but not yet connected) outside several secondary schools. This ambitious IT component of the national capacity building plan will improve access to and reception of educational programmes (already a feature of the education system in Ethiopia using radio and TV in areas where signals are available). The system will eventually permit video-conferencing, and will as more regular flows of disaggregated budget/expenditure data, school enrolment and performance data between three
levels of government.

Unclear situation on staffing at local level
According to our informants and the literature, staff shortages and turnover, especially at the woreda level, are major problems for local development. We asked about this, but emerged with the impression that they must be more significant issues elsewhere in the country. There appears to be no reliable information about the need for staff of requisite grades and qualifications, staff available in post, or historic turnover statistics.²⁷ It should be noted that in the early days of the decentralisation policy, the World Bank was critical of 'theoretical' woreda staffing structures, labelling them 'unrealistic and unsustainable'.
We conclude that there may be an increasing risk of the public sector losing skilled staff - particularly at the technical and professional levels - as more become trained in commercially valuable skills such as double-entry book-keeping.

9 Capacity, change and performance in Ethiopia

There is a dearth of well researched, objective evidence of what has changed, and why, in public service delivery performance, standards, effectiveness of management of inputs, and effectiveness of capacity building programmes attempted so far. This has impeded our research, and inhibited the drawing of firm conclusions. If not attended to, it will inhibit the process of defining optimal approaches to capacity building within the PSCAP and other programmes planned within ESDP II.

With the above caveat, we put forward the following tentative conclusions on the relationship between capacity, change and performance in the Ethiopian education service delivery system. It is hoped that the analysis will contribute to the deliberations of forthcoming annual joint review mission of the ESDP.

Evidence of change especially at woreda level
There is evidence of considerable (positive) changes in public service delivery since 2000. Reform is being seriously attempted particularly at regional and woreda levels of government. Compared with the situation in the late 1990s (World Bank, 2001), the degree of change is particularly marked in the operational environment at woreda level, and in the attitude of regions towards woredas and communities (and the impact of the fiscal policies introduced in 2001).

Factors supportive of systematic capacity building efforts
What is clear, however, is that a wide range of mutually supportive factors is affecting regional and local organisational performance in the current Ethiopian context. These have created favourable circumstances in which systematic capacity building approaches (such as those used in the DSA and BESO programmes) have been able to deliver positive results. These factors include:

• The decision in 2001 to provide unconditional block grants (covering routine and capital finance) to woreda councils (in the main regions) was significant. This gave woredas unambiguous responsibility for fair and efficient allocation of resources among sectors, uses and geographical areas. These block grants were significant as a symbol of and a tool for woreda empowerment.
• This unusually liberal grant mechanism replicated the federal-regional government grant mechanism that had been in place since 1992. Both mechanisms owe their existence to unusual constitutional features of Ethiopia: each regional state has its own constitution and can secede from the federal republic under certain conditions, and the fact that federal powers over regions are strictly circumscribed.
• The consensus about and coherence of policy direction in relation to devolution, capacity building, and the importance of education since 2001 have been positive factors. The political consensus - that devolution must be made to work, and local accountability structures have an important role to play - is one element in this context. There are now clearer incentives, internal to organisations, combined with external pressures to perform. The capacity building implications of desired performance change are thus becoming clearer.
• This consensus appears to be related to the dominance of a durable coalition of political groupings at all four levels of political representation (federal, regional state, woreda and kebele), and the fact that the coalition is based on a consensus around a national development credo and ethos. This was forged in a bloody, bitterly fought civil war that cemented relationships among allied ex-fighters. The period in the past when this alliance was under strain was successfully and decisively brought to a conclusion, when only those who were loyal to party policy were allowed to continue in the party and government.
• These themes in the political background of Ethiopia have served to minimise competition between politicians at the different levels, and have led to an absence of political interference in resource allocations (in terms of functional use, and geographical distribution), and (at least relative to other neighbouring countries) an absence of financial corruption.
• Pressures from councils of public representatives at regional, woreda and kebele levels, together with large community contributions to education service infrastructure and even recurrent costs, have created considerable and continuously high levels of accountability. This appears to be related to the 'depth' of democracy in terms of elected institutions in Ethiopia (the (low) average numbers of electors per representative at kebele and woredalevels are shown in Figure 2).28 This, in our view, appears to compensate for any (inevitable) deficiencies of technical capacities at local level, and obvious delivery imperfections and systemic stresses due to the chronic shortage of resources.
• Given the limited size of the Ethiopian private sector, the turnover of (trained) staff has not yet been a major threat to local government capacity ... but it could well be in the future, given the considerable capacity building activities planned in the public sector, and the gradual introduction of 'marketable' skills.
• Decentralised human resources management arrangements in the civil service (with regions having their own civil service commissions) appear to encourage innovative measures to enhance organisational as well as individual performance. In Amhara, innovative initiatives to increase the performance orientation of staff focus first on organisational and then individual performance, as an incentive to improve teamwork and cohesion.

The modest impact of the Ministry of Capacity Building
The impact of the Ministry has not yet been fully felt, probably due to its novelty (it was created only in 2002). Its role has been limited to the preparation of the major (forthcoming) PSCAP (which will impinge indirectly on education service delivery); handling the community and parent-teacher mobilisation aspects of the ESDP (with very positive results, based on our consultations). It appears, however, that the Ministry has not kept abreast of, or contributed to, some very significant and innovative capacity building experiences in the education sector, particularly those sponsored by USAID.

The biggest capacity gaps are at the federal level
The extent to which the capacity gaps within the GoE are acknowledged, is unclear. We heard several negative comments about lack of coherence in the Ministry of Education, and about the apparent shortage of technical and implementation capacities in some departments, especially the new ones such as the Higher Education Strategy Institute and the Quality and Relevance Assurance Agency.

Appendix: List of individuals interviewed


DfID Ethiopia
Dr Hazel Bines Education Adviser
Ato Alemayehu Minas Education Adviser
Ato Kalayu Gebre-Selassie Capacity Building Adviser

Government of Ethiopia
Ato Haile Melekot Giorgis State Minister of Capacity Building
Wro Yeworqabeba Bahru Expert, HRD, MCB
Tesfaye Bekelle Team Leader, Planning, MCB
Ato Mesfin Tafesse Commissioner, Civil Service Commission
Ato Atakilti Head, Training Department, CSC
Ato Sitotaw Yimam Head of Planning and ESDP II, MoE
Ato Alemayehu Worku Project Preparation and Monitoring Expert MoE
Ato Tizazu Asare Institute of Curriculum Development and Research MoE
Ato Demissew Bekelle Director, Education Media Agency
Wro Kassech Demissie Senior Expert, Women's Affairs Department, MoE

World Bank
Ato Getahun Gebru Education Adviser

USAID
Ato Aberra Mekonnen Education Adviser

Sida
Kenthe Wickmann Education, Research and Culture Adviser
Tony Redmond Team Leader, MoE/EC/British Council Education Accounting and Procurement Project

Amhara regional state government
Ato Yoseph Anteneh Head of Regional Capacity Building Bureau
Ato Yinager Dessie Deputy Head
Ato Tesfaye Mehoye Education Desk Officer
Ato Fanta Moges Deputy Bureau Head, Regional Education Bureau (REB)
Ato Mekbib Alemu Zemenfes Deputy Head, BoFED
Ato Theodoros Shewareget Head of Human Resources, REB
Ato Setu Ainalem Head of Planning and Information REB
Ato Takele Kidan Head of Programming REB
Ato Habtamu Bizuneh Head, Planning REB

Woreda Fogera:
Kibret Mohammed Head, Woreda Capacity Building Office
Mulu Mengesha Head, Woreda Education Office

Woreda Mota:
Melese Belchu Head, Woreda Education Office
Tibebeu Yigzae Education Expert, WEO
Mengistu Anley Head, Woreda Capacity Building Office

Woreda Merawi:
Minas Hiruy Head, Woreda Capacity Building Office
Bayalew Menesha Head, Woreda Education Office

Woreda Tarma Ber-Debresina
Yilma Tilahun Head, Woreda Capacity Building Office

References

Amhara Regional Education Bureau (REB) (undated) Alternative Basic Education for Children Out of School (ABECS): Scaling up.

Checkole, K. (undated) Networking for Development Experiences of the Basic Education Network, Presentation by the BEN Coordinator.

Constitution of the Federal Democratic Republic of Ethiopia, Proclamation No.1, 1995.

Decentralisation Support Activity Project, Progress Reports May 2003, March and May 2004.

Development in Practice (DiP). 2003. Review of Capacity Building Approaches for Local Government in Ethiopia (draft), for DfID.

DfID. 2002. Ethiopia: Country Assistance Plan. London: DfID.

Federal Negarit Gazeta , August 1995 and October 2001.

Fesseha, T. (2002) A Comparative Analysis of, and Consolidated Report on, the Organisational Capacity Assessment and Re-assessment of 24 Ethiopian Local NGOs. Addis Ababa: Pact.

Girishankar, N., Alemayehu, A. and Ahmad, Y. 2001. Handling Hierarchy in Decentralised Settings: Governance Underpinnings of School Performance (Oromia Region). Working Paper 21.World Bank Africa region.

GoE. 1999. Education Sector Development Programme: Action Plan. MoE.

GoE. 2002a. Education Sector Development Programme II: Programme Action Plan. MoE.

GoE. 2002b. Guidelines for Educational Management, Community Participation and Educational Finance. MoE.

GoE. 2003a. National Capacity Building Plan 2003-2008, Vol. 1, Consolidated Strategic Public Sector Capacity Building Programme (PSCBP).

GoE. 2003b. Ethiopia Sustainable Development and Poverty Reduction Programme (SDPRP). Annual Progress Report (MoFED).

GoE. 2003c. EDSP II Joint Review Mission Report (October-November 2003)

GoE. 2003d. Education Statistics Annual Abstract EMIS, MoE.

GoE. 2004a. Programme Implementation Plan for the Public Sector Capacity Building Programme.

GoE. 2004b. Programme Implementation Manual for Education Sector Development Programme.

GoE. 2004c. Situation of Girls Education in Ethiopia. MoE, Women's Affairs Dept.

GoE. 2004d. Civil Society Organisations' Capacity Building Programme: Programme Design MCB
(CSO capacity building programme).

HESO. 2004. Higher Education System Overhaul, Report of the Committee of Inquiry into governance, leadership and management of Ethiopia's higher education system.

Morgan, P. What is Capacity? Going beyond the Conventional Wisdom. Maastricht: ECDPM.

Morgan, P. 2003. Draft Background Paper on Methodology: DAC Study on Capacity, Change and Performance.
Maastricht: ECDPM.

Polhemus, J.H. and Yohannes, L. 2002. Situational Analysis of the Governance Sector in SNNPR and Tigray Regions, vols 1 and 2, for the Embassy of Ireland.

Watson, D. 1999. A Framework for a Strategy in Governance (1999-2002), for the British Council Ethiopia.

Watson, D. and Cross, S. 2003. A comparative analysis of approaches to capacity building in local government in Africa, in S. Cross and D. Watson, Programme Mapping for Decentralisation, for DfID Malawi.

Watson, D. and Khan, A.Q. 2005. Capacity Building for Decentralised Education Service Delivery in Pakistan. Discussion Paper 57G. Maastricht: ECDPM.

Watson, D., Yohannes, L. and Khan, A.Q. 2005. Capacity Building for Decentralised Education Service Delivery in Ethiopia and Pakistan: A comparative analysis. Discussion Paper 57I. Maastricht: ECDPM.

World Bank. 2001. Ethiopia: Woreda Studies, Vol. I. Main Phase. World Bank Country Office Ethiopia.

World Bank. 2003. Higher Education Development for Ethiopia: Pursuing the Vision, World Bank Sector Study.

World Bank. 2004a. Public Sector Capacity Building Programme Support Project. Project Appraisal Document.

World Bank. 2004b. Education in Ethiopia: Strengthening the Foundation for Sustainable Progress (draft).
Human Development Department, AFTH3. Washington, DC: World Bank.


Notes

1 The Ethiopia case study and a comparative analysis of the two cases have been published as ECDPM Discussion Papers 57G and 57I, respectively.

2 Kebeles were created by the Derg as a means to gain control, but were retained and empowered with locally elected assemblies by the EPRDF.

3 In 1998 Ethiopia's GDP per capita stood at $110, less than one-quarter the average of other countries in the region.

4 This report served as a baseline for assessing the growth of capacities at regional and woreda levels. Five years later, the situation at both levels is very different to that described in the report.

5 This was preceded in 1994 by a major government-led assessment by a task force, of management arrangements and practices at all levels of government.

6 Summarised in the PSCAP appraisal (2004-8).

7 See Watson and Cross (2003) for a summary of lessons from local government capacity building in Africa.

8 Devolution Support Activity (DSA) Quarterly Progress Reports 24 (May 2003) and 28 (April 2004).

9 The deputy head of the Bureau of Finance and Economic Development in Amhara provided evidence of dramatic improvements in financial reporting following the introduction of new procedures derived from the DSA project.

10 The author of the DiP report was involved in capacity building in Zimbabwe in the 1980s and 1990s using an approach encouraged by facilitators within target local governments.

11 The 2003 mission included no less than 27 members from government and donors.

12 The international benchmark is for recurrent budgets to be split 70% for salaries and 30% for other expenditures.

13 An initiative launched in 2000, coordinated by UNESCO, in which African governments committed themselves to achieving 'education for all' by 2015.

14 One assessment concluded that 'there is a high degree of budgetary discipline ... the risk of corruption and diversion of funds is low' (cited in GoE, 2003b, p.33), yet Transparency International ranked Ethiopia 92nd (out of 133) in terms of the apparent prevalence of corruption (the same as Pakistan).

15 This process was less than satisfactory for the RCBBs, since they had unrealistic expectations, and had to cut back regional programmes. The experiences of the (few) ongoing programmes such as the DSA or BESO were also not assessed.

16 The DiP study was submitted to the GoE and donors but was not discussed in the PSCAP preparation process. This is unfortunate, since it comments on successful approaches not so far used in Ethiopia.

17 Whether the ESDP is a full SWAp is disputed. It contains several bilaterally funded projects whose coordination is extremely time-consuming for the MoE, and which perpetuate the role of the MoE as an implementer.

18 For example, DfID committed £20 million to PSCAP, but only £6 million to the pool, retaining £14 million for GoE/DfID bilaterally agreed TA to support PSCAP.

19 In particular to limit the roles of Ministries to those defined in the constitution: policy formulation, implementation and monitoring, standard setting, and technical advice and capacity building for implementing agencies at lower levels of government.

20 The federal Civil Service Commission will consult regional CSCs later this year in an attempt to redefine its role, and to provide a new impetus for the CSRP (which is coordinated by the CSC).

21 Staff turnover varies by sector: turnover among professional health staff is reportedly higher than among education professionals and teachers. Turnover is high in remote regions where living conditions are difficult, and no extra allowances are paid for remote postings.

22 A recent review of the CSRP found an encouraging variety of regional initiatives. We saw evidence of one in Amhara. After workshops on job definition and public service transparency, RCBB and REB officers (at least) were encouraged to put their name, job title and photographs on their office doors. Regular reviews of job performance were reportedly being conducted.

23 It appears that such awards are not included in the bureau's budget, but individuals receive annual increments in salary. Since routine annual increments (i.e. a sort of seniority recognition) have now been stopped in Amhara, these performance awards are valued.

24 There is something of a paradox here: most of the capacity building efforts of the E&TBs, PTAs and communities has taken place under the auspices of the RCBB, yet the MoE benefits from the (apparently positive) impact of BCBB capacity building efforts: more community contributions.

25 The head of capacity building is a member of the cabinet, but the head of education is not.

26 An unusual feature of the Ethiopian civil service - at least compared to Commonwealth versions of the Whitehall model - is the absence of a 'principal' or 'permanent' secretary (chief civil servant) and related support staff below the minister(s) of a department.

27 The MCB plans to examine this issue in the near future.