Aid, ownership, and coordination in the health sector in Ethiopia

ABSTRACT The Government of Ethiopia is seen as a owner of its national programs and policies and thus also as a strong coordinator of the foreign aid it receives. This is also the case in the health sector in Ethiopia, where the Ministry of Health have shown leadership in the last two decades. National health plans have been clear-cut and had ambitious objectives, to which the international donor community has adhered. The government-led coordination structures and joint health financing arrangements have been instrumental for improved donor coordination and aid effectiveness in the sector. This has led to impressive results, looking at the poor state of health that the government inherited from former regimes. However, the sector has at once been heavily dependent on foreign sources and characterized by high aid fragmentation. In this paper, we describe the health plans and health financing between 1990 and 2015. We also look at health leadership, donor coordination, and the results of investments in health.


Introduction
In the aid 'reformist' debate, the argument most pronounced at present is that foreign aid did not bring desired results because it was an exogenous push in the absence of a strong partnership and sense of ownership by recipient partners for their development agenda (OECD 2008(OECD , 2011aRiddell 2009;Goldberg and Bryant 2012;Gore 2013). Since the introduction of the sectorwide approach (SWAp) in the 1990s, donors and recipient governments have promoted the need for countryowned and -led development cooperation. However, the debate on aid effectiveness is inconclusive, and there is limited empirical evidence showing aid effectiveness as a result of country-led development cooperation in low-income countries (Bigsten and Tengstam 2015;Haque, Hill, and Gauld 2017). The Ethiopian health sector is no exception.
Ethiopia is an interesting case to study aid effectiveness at the sectoral level, because it is one of the African countries that has shown remarkable progress in the health sector in the context of a consistent health policy environment (IHP+ Results 2010. The sector has been one of the top aid investment destinations for several bilateral and multilateral donors. According to OECD-DAC data (2018a), from 2013 to 2015, there were, on average, 26 donors and more than USD 1 billion flowed annually to the health and reproductive health sectors in Ethiopia. The average annual share of bilateral ODA to the health sector in the years 2015 and 2016 was one of the highest, at 22% of total net ODA disbursement, second only to humanitarian aid at 25%.

Aid effectiveness in the health sector
A series of multiple attempts to reform the landscape of the traditional aid approach finally took formal shape in the Paris Declaration on Aid Effectiveness in 2005. Based on the new model of aid effectiveness, the International Health Partnership and other related initiatives (referred to together as IHP+) (the Universal Health Care 2030 [UHC2030] since 2016), which is one of the international aid and development joint coordination platforms in the health sector, developed seven behavioural principles (UHC2030 2018): (1) Support a single national health strategy; (2) Record all funds for health in the national budget; (3) Harmonize and align with national financial management systems; (4) Harmonize and align with national procurement and supply systems; (5) Use one information and accountability platform; (6) Support south-to-south and triangular cooperation; (7) Provide well-coordinated technical assistance. The Busan High Level Forum recognized these principles as 'effective development cooperation practices,' adding 'private sector engagement' to the list (OECD 2011a). This clearly shows that six of the interconnected principles are anchored in 'country ownership. ' IHP+ describes country ownership specifically in the health sector as the existence of a single national health plan with a longstanding vision and clear health priorities, as well as medium-term expenditure and result frameworks that are jointly assessed and endorsed in a participatory approach under the guidance of strong leadership by the recipient partner in a 'One Plan, One Budget, One Report' approach. (IHP+ Results 2010. A well-developed and country-owned health development plan ensures positive health results through reinforcing alignment and harmonization as well as guiding financing of health priorities and facilitating result-based coordination and mutual accountability for improved aid effectiveness in the health sector (Guillaumont and Chauvet 2001;Ulikpan, Mirzoev et al. 2014).
Overall, some studies concluded that aid contributes to the development of a recipient country when development partners (DPs) support country ownership by building internal capabilities under a true partnership framework (Booth 2012; Goldberg and Bryant 2012;Sweeney and Mortimer 2016) and when aid is invested in national development priorities (Burnside and Dollar 2000;World Bank 2005;Hasselskog and Schierenbeck 2017), with a considerable, stable flow of aid to propoor sectors like health (Mosley and Suleiman 2007) and with strong backing for fighting corruption (OECD 2008). Most of all, a visionary and politically-committed government leadership by a recipient partner that is capable of defining its priorities and creating proper health development coordination platforms makes a difference in ensuring an improved and sustainable health system (Balabanova et al. 2013;Ulikpan, Narula et al. 2014;IHP+ Results 2015;Reich et al. 2016).
In principle, country ownership includes government and non-state actors with respect to owning the policies and coordination endeavors (World Bank 2005;OECD 2008;Carothers 2015), but practically, it would be difficult to consider country ownership in this broader sense in countries like Ethiopia, which are led by a 'developmental state' ideology, in which the role of civil society and the private sector is limited. Therefore, in this paper, country ownership narrowly refers to government-led health development cooperation.

Organization of the Ministry of Health in Ethiopia
The Ministry of Health at the federal level is led by a minister and directors of the directorates under the supervision of two state ministers, along with the heads of the federal hospitals and five agencies, including the Pharmaceuticals Fund and Supply Agency. At regional level, the health system is organized in a hierarchy of regional health bureaus (RHBs), zones/sub-cities (for some regions with a zone or sub-city administration), woreda (a districtlevel administrative unit in Ethiopia) health offices, health centers, and health posts.
In 1997, the government of Ethiopia developed a 20year health sector development plan (HSDP) to implement the 1993 health policy. Several platforms have been organized to coordinate the implementation of the HSDP. The coordination of the health sector involves two levels. The first is steering committees to coordinate with the regions and woredas. Accordingly, the Joint Steering Committee (JSC) coordinates the Ministry of Health and the regional health bureaus (RHBs) at the central level; the regional joint steering committee (RJSC) at the regional level; the woreda joint steering committee (WJSC) at the woreda level; and the Health and HIV/AIDS Committee at the kebele (village-level administration) level. At the second level are the joint coordination structures between the Ministry of Health and the development partners (DPs), which is the focus of this study.

Materials and methods
This is a qualitative study based on thematically-organised semi-structured interviews conducted with 42 respondents drawn from donors residing in Ethiopia and from the Ministry of Health, along with an analysis of other relevant government documents and health data. Of 26 providers of health aid to Ethiopia in 2015, 17 of them participated in the interviews. Apart from the heads of agencies directly responsible for coordinating development cooperation, 12 of the participants were health specialists. The participants were randomly selected to include the large, medium, and small-sized multilateral and bilateral donors. These interviews were administered in two rounds: the first in February and March of 2016 and the second in March and April of 2017. About half of the participants were interviewed in both the first and second rounds of interview sessions to check the data for consistency and to track new developments.
A systematically organized interview guide was employed to assess donors' health programs and portfolios in Ethiopia and the level of their alignment to government health programs and priorities as well as their views towards the government health policies and plans, government leadership in the health sector coordination, the functionality of coordination platforms and practices in the health sector, quality of health policy dialogue, and role of big and small donors in the health coordination platforms. The guide and topic lists for the interviews were derived from and informed by the international literature on aid coordination like the principles of the Paris Declaration and its evaluations and related scientific articles.
Almost all the interviews were audio-recorded, and they were transcribed in verbatim and systematically coded and colored based on themes from the interview guide. We then applied framework analysis matrix using Excel sheets in order to organize, summarize, and analyze the interviews coded into themes and subthemes. Framework analysis is one of the commonly used methods in qualitative research, particularly in the analysis of interview data collected based on structured themes (Smith and Firth 2011;Gale et al. 2013).
Data on development assistance for health (DAH) and trends in health financing and expenditure on the major health status indicators specific to Ethiopia, as well as to Sub-Saharan Africa (SSA) countries, were extracted from reliable and timely published database sources including the Credit Reporting System (CRS) of OECD-DAC, Development Assistance for Health Database by Institute for Health Metrics and Evaluation (IHME), Global Health Observatory (GHO) and Global Health Expenditure databases by the WHO, the World Development Indicators of the World Bank open database, and UNdata. The interviews, the aid data, and the HSDP together with other relevant documents were triangulated and thematically analyzed.

Health plans: hinges for country-led coordination
Analysis of the consecutive health sector plans and relevant documents as well as our interviews show that the HSDPs had some distinct features that contributed to the strengthening of donor coordination and improved health results in the country. First, the HSDP was initially developed based on Sector Wide Approach (SWAp) principle, which was introduced into health sector in Ethiopia in 1997(Ministry of Health 1998, 2002. This encouraged the practices of country ownership, in a broad partnership with the health DPs, from the start. Second, unlike in many other African countries, the HSDP was 'home-made' and the Ministry of Health played a stewardship role in the design and implementation of the program, as well as in defining national health priorities, as perceived by most of the interviewees. Third, the HSDPs were clear in portraying the national health goals and priorities, which have evolved from the rehabilitation and expansion of basic health services, emphasized during the first two phases of the HSDP, to health service quality and equity in the Health Sector Transformation Plan (HSTP), which started in 2015/16 and will run until 2019/20 (see Table 1). This has helped the Ministry to firmly ensure that, on the one hand, support from the DPs fits the health priorities, and on the other hand, that all donors equally value the health priorities. As stated by one of the respondents, 'Ethiopia has one state health sector strategy, we are required to align with it. The priorities of the government are really respected, even by USAID [United States Agency for International Development], which uses parallel systems' (Participant MMH_1a, personal interview, 4 April 2017).
On the other hand, the consultative engagement of health partners in the development of the HSDP, particularly during the last two phases of the HSDP, resulted in increased mutual trust and a sense of shared ownership of the HSDP, as verified by most of the interviewees. They said that this has made the alignment of their programs with the government's health priorities easier and increased their level of confidence in investing more in the health sector, with sustained engagement in joint health development planning and coordination.
Last year we were active together to shape the Health Sector Transformation Plan. The health part was really pleasant, really good. With 90% of the plan, we are very pleased, and we think it is going to be the right direction. (Participant BMH_1a, personal interview, 2 February 2016) Principally, the practice of woreda-based health sector planning with a 'One Plan, One Budget, and One Report' approach has empowered the Ministry to reinforce strong country ownership and country-led health development coordination, as stated by the interviewees. The Ministry introduced woreda-based plans during the HSDP-IV, and they have been operational since 2013/14. Through this process, woredas prepare a woreda-based health sector core plan based on an indicative plan and the respective national priorities, as defined by the federal Ministry of Health. Then, the woreda plans are consolidated to form a national annual health plan. The core plan has a results framework and cost plan, and it is the only operational plan in the sector to which the government at all levels and all other health partners adhere. This helps local-level implementing partners, including non-governmental organizations (NGOs), to align and jointly coordinate their programs with the health priorities of this single plan.
Implementation of the HSDPs has also benefited from global initiatives including the Millennium Development  Goals (MDGs) and the Paris agendas on aid effectiveness, as well as from the national development policy context. HSDP-III, for instance, was employed under the ambitious national plan, known as the Plan for Accelerated and Sustainable Development to End Poverty (PASDEP). Global dynamism from the Paris agenda on Aid Effectiveness inspired the implementation of HSDP-III with a wider scope of country-led health development cooperation. During this period, the Ministry endorsed the HSDP Harmonization Manual and the IHP+ Country Compact along with the Joint Financing Agreement (JFA) for the Millennium Development Goals Performance Fund (MDG PF). The Healthcare Financing Strategy, which recognized community-based health insurance, social health insurance, and other core strategies in the health sector, was introduced and implemented in HSDP-III. The context that drove the momentum gained by HSDP-III continued during the implementation of HSDP-IV. The HSDP finished in 2014/15 and has been replaced by another 20-year health sector strategy called 'Envisioning Ethiopia's Path to Universal Healthcare through Strengthening of Primary Healthcare' (Ministry of Health 2015b). The first phase of this strategy is the HSTP.

Leadership of the Ministry of Health: government in the 'driver's seat'
The health sector in Ethiopia has had successive leadership at the top that demonstrated professional capability and political commitment to lead in the 'driver's seat'. This started with former minister Dr Tedros Adhanom, who introduced, in the eyes of our respondents, significant changes in the health sector. According to the WHO (2017a), 'the transformation he led as Ethiopia's Minister of Health improved access to healthcare for millions of people. Under his leadership, Ethiopia invested in critical health infrastructure, expanded its health workforce, and developed innovative health financing mechanisms.' Dr Tedros, who has a PhD in community health, is a malaria expert and former head of the Tigray Regional Health Bureau, led the Ministry from 2005 to 2012 and was elected as Director General of WHO in 2017.
The successor of Tedros, as Minister of Health, was Dr Kesetebirhan Admasu (2012-2016), a medical doctor and a public health specialist. He was a state minister from 2010 to 2012 under the supervision of Dr Tedros. He sustained the momentum of the health sector reform agenda and showed remarkable leadership in family planning and maternal and child health, and in directing the Health Development Army (HDA), especially the women's HDA, according to our respondents. Prof. Yifru Berhan Mitke, (minister during the interview period) is a medical doctor by training and a university professor. He led the medical schools of some of the distinguished universities in Ethiopia. The state ministers and most of the directors are also health professionals and have, in the eyes of our respondents, a clear vision on the health sector.
Further analysis of the interviews shows that the Ministry of Health has been one of the strongest ministries in Ethiopia with a 'practical' country ownership, which has played a leading role in the joint health development coordination process. Some of the strengths identified by the interviewees were: First, the Ministry has enough political determination to say 'no' when support from a DP is against the national health priorities and the principles of ownership in development coordination.
They are quite capable and strong enough to defend their strategic goals and strategies. This is good, especially for the DPs using the country system, budget support, and the pooled fund. If you did not have such a strong ministry, it would be a crisis. (Participant MLH_1, personal inteview, 28 March 2017) Second, the top leadership has demonstrated technical and professional ability to coordinate resources from the health partners towards the goals of the sector: 'They have the experience, they work with different donors, and they have the expertize. Most of the Directorate heads are medical doctors and capable people that are bright, risk takers and committed' (Participant BLG_1a, personal inteview, 1 March 2016). Also, the leadership has shown that it has the practical experience to deliver consistently on the aid effectiveness principles: 'When we go to the other countries, they are aware of the IHP+ framework of all these donor coordination preambles and statements to be made, but Ethiopia has made quite good use of that framework' (participant MLH_2a, personal interview, 2 February 2017).
Third, the level of corruption in the health sector was perceived as low, and the leadership has shown the political commitment to fight and reduce corruption in the health sector. One of interviewees from a large multilateral organization said: 'The perception of the public was that corruption in the health sector has been one of the lowest in all the ministries' (participant MLH_2a, personal interview, 2 February 2017). Generally, most of the health partners interviewed acknowledged that stability of the health plan and strong country ownership of the health leadership in Ethiopia has inspired DPs to provide increased and stable DAH for strategic results in the sector. This will be elaborated in the following section.

Health sector financing and aid fragmentation in Ethiopia
Donors have increased their support to the health sector in Ethiopia, with a rise in health and reproductive health aid to over a billion USD in 2014, which makes the country one of the top-five DAH recipients in SSA next to Nigeria, Tanzania, and Kenya (Appendix A). DAH steadily increased following the health sector reforms and execution of the HSDP in the late 1990s. The average growth rate of DAH to Ethiopia before HSDP (1990HSDP ( -1996 was only 4%; however, it grew by 68%, 65%, and 129% during implementation of HSDP-I (1997HSDP-I ( -2001, HSDP-II (2002-2004 and HSDP-III (2005/06-2009/10), respectively (Graph 1).
A sharp rise in DAH to Ethiopia was observed during HSDP-III. This can partly be attributed to an increase in vertical funding by the Global Alliance for Vaccine and Immunization (GAVI) and Global Fund (Ministry of Health 2014a), as well as an increase in global commitments and support following the Paris Agenda on Aid Effectiveness in 2005, the relatively better policy environment, and the presence of strong leadership at the top during that period.
However, a closer look at the health financing in Ethiopia shows that the sector has faced two fundamental challenges, which indicates the need for effective coordination of DAH in the country. The first one is that the sector has been heavily dependent on external sources; and the second one is that aid fragmentation in the sector has been high, even higher than in the other top DAH recipient SSA countries and the level of fragmentation increased after the Paris Declaration. On the other hand, the government has shown a strong commitment to increasing public health expenditure and has coordinated the use of program-based approaches like pooled funding to minimize the effect of aid fragmentation in the sector. The following sections show details of this analysis.
The financing structure of the HSDP gradually changed from domestic sources to largely external financing (see Appendix E). Particularly, HSDP-I and HSDP-II were highly dependent on domestic financing. At the beginning of HSDP-II, for example, donors contributed only 10.3% of total health financing, while the share of domestic financing was 89.7%. Because of the Ethio-Eritrea war, most of the donors declined to support Ethiopia during that period. However, from the beginning of HSDP-III, the level of external financing increased more than three-fold. From only 15.2% at the beginning of the HSDP in 1997, it reached half (50.1%) of total health funding in 2010 and then declined to 41.7% at the completion of the HSDP. This carries substantial risk in terms of the sustainability of health service provision in the country, in the event that, for whatever reason, major donors were to walk out on financing the sector. One of the interviewees from among the lead donors in the sector expressed the concern as follows: It is a sector where we have a commitment from the government side, but it is a good example of where donors have contributed to creating distortions, as about 50% of the funding of the sector comes from donors and lenders, which is an external source. (Participant MLG_1, personal communication, 23 February 2016) An analysis of average annual DAH flow to Ethiopia over the five years from 2011 to 2015 also shows that the top three of the 26 donors contributed two-thirds of the total health aid in those years (Appendix B). This resulted in high aid fragmentation in the sector, as analyzed in Table 2. Of the different methods to measure incountry aid fragmentation, we use the OECD-DAC approach, which utilizes country programable aid (CPA) disbursement to evaluate the level of significance of donors' relation in a country or in a sector. According to the broader definition of this method, donor's relation in a sector is 'non-significant' if the donor's support is neither 'important', which means that the donor is in the bottom list of small−sized donors that cumulatively represent less than 10% of total CPA to the sector, nor 'concentrated', which implies that the CPA share of the donor to the sector is below its average CPA share countrywide (OECD 2009(OECD , 2011b. In a narrow sense, aid fragmentation in a sector can be measured by analyzing the number of donors whose cumulative CPA contribution to the sector is below 10%. Our analysis takes the broader approach. From 2004 to 2014, the total number of donors in the health and reproductive health sub-sectors in Ethiopia increased by 50% and 40%, respectively. Also, the total CPA contribution of both sectors in 2014 was nearly four times the amount in 2004. Subsequently, the aid fragmentation ratio for the health and reproductive health sub-sectors during the period increased by 10.5% and by 67.5%, respectively. A study by Alemu (2009) also substantiated that the aid landscape in the health sector of Ethiopia in the mid-2000s was characterized by high aid fragmentation and showed no substantial improvement despite the global aid effectiveness declarations. The fragmentation in the reproductive health sub-sector was high in both periods. A closer analysis of the sources of aid fragmentation shows that support from the bilateral donors (DAC countries) in the sub-sectors was more fragmented than that of the multilaterals in both periods. (see Appendix I).
Compared to the top-ten DAH recipient SSA countries from 1990 to 2014 (Table 3), the 2013 and 2014 average fragmentation ratio for Ethiopia was the highest (59%) followed by South Africa (55%), Kenya (54%), Zambia (54%), Mozambique (53%) and Uganda (51%) that over 50% of the CPA they received was fragmented. The high level of fragmentation in Ethiopia was due to high fragmentation in the population and reproductive health sub-sector, in particular in 2013. The average total CPA Ethiopia received during this period was the second largest after Nigeria. However, Nigeria was able to mobilize larger CPA from a relatively low number of donors with more significant relations. As a result, the fragmentation ratio for Nigeria was relatively low, which implies that Ethiopia needs to increase the number of significant relations or to optimize the resources of all CPA providers without decreasing their number, by using program-based approaches like pooled funding. How then does Ethiopia manage these program-based approaches?
Given this level of aid fragmentation, Ethiopia institutionalized joint health financing arrangements and coordination practices, which appeared to reduce the adverse effects of aid fragmentation on health aid effectiveness. The Sustainable Development Goals Performance Fund (SDG PF), which was established in 2009 as the Millennium Development Goals Performance Fund (MDG PF), has been a SWAp implementing tool and one of the pooled funding and coordination mechanisms through which the DPs have committed to supporting 'One Plan, One Budget, One Report' principles in Ethiopia. The fund was initiated by two donors with USD 10.6 million; this grew to 11 contributors with total funds of around USD 235 million in 2014. DFID has been the largest (providing 61% of the total fund  in 2014) and most consistent contributor to the fund, and the World Bank is second, followed by The Netherlands and Irish Aid (based on 2015 figures).
The share of the fund in terms of CPA disbursements grew sharply from 2% in 2009-28% in 2014 (Table 4). But, support from other principal donors like USAID and the Global Fund has been missing, despite the fact that the fund has been one of the preferred financing modalities of the Ministry of Health (Ministry of Health 2008). The fund is administered by the Ministry of Health itself and, hence, the Ministry has the flexibility to channel the resources through its own systems, which has also been good in terms of strengthening the health system, according to our respondents.
The Fund has helped to mobilize resources, including from smaller contributors (Australia, Ireland, Italy, Spain, UNFPA, UNICEF, WHO), and to fill funding gaps in priority areas in the HSDP with less harm from aid fragmentation and donor proliferation. The SDG PF has also been used to finance some health priority areas left underfunded by donors because donors have been more involved in reproductive health like HIV/AIDS programs (Appendix J). For example, in 2014/15, the fund was used for public health commodity procurement (68.4%) and health system strengthening (22.2%), (MoH 2015). The pooled fund is also used to strengthen the health systems of regional states by distributing resources in kind and technical support based on their woredabased annual plans, according to an interviewee from the Ministry.
The SDG PF is one of the effective funds with 'real' country ownership by the Ministry of Health built on mutual trust with its DPs, as described by our interviewees. One of the interviewees from the large bilateral donors and an important contributor to the fund described it by saying: 'I have seen many pooled funds in many sectors, but this is one of the better ones and more effective than the pooled funds with which we worked over the years' (participant BLH_1, personal interview, 23 March 2017).
Analysis of the health sector financing trends in Ethiopia also shows that health expenditure of the government and the financing role of donors during the implementation of the HSDPs has gradually increased. As indicated in Graph 2, the government of Ethiopia persisted in increasing its health expenditure as a percentage of total health expenditure throughout the HSDP period. Prior to the introduction of the HSDP, domestic health financing was dominated by private health expenditure. Private health expenditure includes resources from out-of-pocket payments by patients, non-profit institutions, and private insurance and other related finance. However, out-of-pocket as a percentage of total private health expenditure in Ethiopia was nearly 80% during the HSDP implementation period. In that sense, the burden on individual citizens' health expenditure, dropped from 46.4% in 1997-32.3% at the close of the HDSP in 2014. To further decrease the pain of out-ofpocket payments and also the level of dependence on external sources, the government recently introduced Graph 2. Health expenditure by expending agents in Ethiopia (1995Ethiopia ( -2014. Source: Authors' calculation based on data from the Global Health Expenditure Database by the WHO (2017b). social and community-based health insurance schemes. According to Ali (2014), these insurance schemes 'increased health care utilization, access to medicines, and quality of services'. The health expenditure by nonprofit institutions also steadily increased, despite the fact that its share of total health expenditure was still quite small.
Evidence further shows that the proportion of heath expenditure as part of total government expenditure and as a percentage of the national GDP has increased since the implementation of the HSDP. For example, general government health expenditure (GGHE), as a percentage of total general government expenditure increased from 7.3% in 1997-15.75% in 2014, putting Ethiopia at third place in SSA compared with Malawi (16.8%) and Swaziland (16.6%), (Appendix F [column 2] and Appendix K).
Total health expenditure of Ethiopia as a percentage of its GDP increased from 3.2% at the beginning of the HSDP in 1997-4.9% at its completion in 2014 (Graph 3). This was mainly due to an increase in the GGHE, which constituted the major part (59% in 2014) of it, even if the share of out-of-pocket expenditure by patients remained high during this period. The increase in the GGHE as a percentage of GDP was partly linked to an increase in external resources. The share of external resources grew more than three times during HSDP-III and nearly seven times in HSDP-IV compared to the level in HSDP-I. Ethiopia's GGHE as a percentage of its GDP (2.9%) in 2014 was among the top 10 spending countries in Africa (Appendix K), but it was far below the minimum 5% target recommended by the WHO (2001).
As indicated in Graph 4, total per capita health expenditure for Ethiopia at the completion of HSDP in 2014 (USD 26.7) was six times higher than the per capita expenditure in 1997 (USD 4.4). Even if it was low compared to the WHO's recommended minimum per capita target of USD 34 (WHO 2001), the growth in per capita expenditure during the HSDP was Graph 3. Health expenditure by financing sources as a percentage of GDP in Ethiopia (1995Ethiopia ( -2014. Source: Authors' calculation based on data from Global Health Expenditure Database by the WHO (2017b).

Platforms
The coordination landscape in the health sector in Ethiopia comprises the Joint Consultation Forum (JCF), Joint Core Coordination Committee (JCCC), Joint Review Mission (JRM), Annual Review Meeting (ARM), and SDG PF (formerly the MDG PF). The JCF is the highest coordination and dialogue forum, in which the Ministry organizes policy discourse and oversight of coordination jointly with health partners. The JCF comprises the Minister, state ministers, all directors of the Ministry of Health, heads of agencies, DPs the Health, Population, Nutrition (HPN) Donor Group, two NGO consortiums, and members of the private sector as well as representatives from health professional associations. The JCF is chaired by the Minister of Health and co-chaired by one of the HPN Donor Group co-chairs. The JCCC is the technical arm of the JCF and operates using technical task forces and ad hoc sub-committees. It involves technical people from the Ministry of Health and the HPN Donor Group, and is chaired by the Director of the Policy and Planning Directorate of the Ministry of Health (Figure 1).
The JRM and ARM are the common monitoring and evaluation platforms in which the government and its health partners track the annual progress of HSDP implementation. The JRM is a joint mission whereby groups of representatives from the health partners, government, and other stakeholders visit selected samples of regional states, districts, and health facilities to assess HSDP implementation issues and to verify on the ground health-related data coming through the Health Information Management System. The consolidated report from the discussion of different joint missions serves as input for the ARM and JCF. The ARM is an annual gathering with representatives of the health partners, regional states, districts, hospitals, and health facilities, as well as health extension workers. Based on input from the JRM, issues like the performance of the sector, next plan priorities and other policy issues are thematically discussed. It reviews and endorses the woreda core annual plan.
The HPN Donor Group is a donors-only sector working group for the health sector under the Development Assistance Group (DAG) and has more than 26 members, including the 11 contributors to the SDG PF, which some of our respondents called the 'home of health'. It provides technical support to the JCCC through its technical sub-groups and facilitates policy dialogue through its role in the JCF.

Assessment of donor coordination in health sector
Most of our interviewees believed that these platforms have facilitated the participation of donors in the joint planning, implementation, and evaluation of the HSDP and subsequent medium-term strategic plans, although our interviewees indicate that this is still weak at the regional and local levels. According to our interviewees, large and small donors, bilateral and multilateral, have played a role in the coordination process, with different expertize, approaches, and levels of engagement, which was generally seen as productive to the health sector. The World Bank, for example, has been one of the active players in the health sector, especially in the financing area where most of its expertize lies. It has been among the major contributors to the SDG PF through a performance-based financing approach and has been active in the JCF meetings and the HPN Donor Group. The interviewees noted that the European Union Delegation has also been actively involved in the HPN Donor Group, the SDG PF, and in the JCF and JCCC.
The role of the UN agencies has varied across the agencies. But, their role in the coordination platforms and their contribution in general, and in the SDG PF financing, in particular, has not been satisfactory, according to the majority of interviewees. The WHO, for example, was not an active actor in the coordination fora and policy dialogue, as it should have been. UNICEF, however, was relatively active in the coordination platforms and was strong in coordinating technical assistance in program areas like the Expanded Programme on Immunization, maternal health, epidemics, and follow up support, in the eyes of our interviewees. On the other hand, the presence of emerging vertical donors including the Global Fund and GAVI in the coordination platforms was perceived as weak by the interviewees. GAVI coordinated its projects from Geneva and was irregularly present at the meetings, while the Global Fund has become more active in the coordination platforms in recent years.
Among the bilateral donors, DFID was one of the largest and strongest partners in the health sector with the biggest, most consistent contribution to the SDG PF, as perceived by respondents. The Netherlands is a medium-sized donor who was active in the sector, contributing a significant amount to the SDG PF with consistent technical support over a long period. USAID is the biggest donor in the health sector, with its special implementing structures and a funding system parallel to the system of the Ministry. Although USAID uses parallel systems, it has been efficient and active in the health sector coordination platforms, according to its donor counterparts.

Major results in the health sector in Ethiopia
In the 1980s and early 1990s, the health status in Ethiopia was one of the poorest, even among Sub-Saharan African (SSA) countries. However, Ethiopia is, these days, seen commonly as one of the 'success stories', because of its achievements in the health sector (for example, IHP+ Results 2012Balabanova et al. 2013;Spicer et al. 2014;Reich et al. 2016). Its overall health status is still lower than the global average, but most of its health indicators are now rapidly improving, due to the effective implementation of the HSDPs under strong country leadership. A health specialist interviewee from one of the large bilateral donors noted the progress as follows: They have done much better than any other country in this part of the world with resources less than what they have spent. However, the counter-argument is that because they were low at the start, they still have a long way to go. (Participant BLH_1, personal interview, 23 March 2017) Data from the Global Health Observatory of the WHO (2017c) shows that during implementation of the HSDP, life expectancy in Ethiopia increased, the fertility rate decreased, and the country achieved most of the health and health-related MDGs. Life expectancy at birth started to rise above the SSA average at the beginning of the HSDP in 1997/98. It increased at a rate of 28.4% to reach an average of 64.5 years in 2015 from 50.3 years in 1997. The fastest growth rate (9.1%) in life expectancy and a decrease in the fertlity rate compared to the SSA average was observed during HSDP-III (2004/ 05-2009/10). (see Appendix G). According to Naghavi et al. (2017), this improvement was 'beyond expectation,' given the level of economic development in the country.
The success that Ethiopia achieved in attaining the health and health-related goals of the MDGs could explain Ethiopia's remarkable achievements in its health sector. According to the 2014 UNDP MDGs report (2015), 'Ethiopia has successfully achieved six of the eight MDGs'. The majority of the MDGs achieved were health and health-related goals as shown in Appendix L, and the major ones are discussed as follows: Ethiopia reduced the under-five child and infant (less than 1 year) mortality rate (<5MR and IMR) per 1,000 live births by 71% and 65%, respectively, from the year 1990-2014, and met the 67% (two-thirds child mortality) target of MDG 4 early in 2012/13. (See Appendix L and Graph 5). During the HSDP implementation period (1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014), <5MR was reduced by 62% and IMR by 57%, while the highest reduction for <5MR was recorded during HSDP-III (26%) and HSDP-IV (25%). Subsequent to the introduction of the HSDP in 1997, Ethiopia had gradually reduced child mortality and had the lowest average <5MR (68) and IMR (46) per 1,000 livebirths in SSA (which had averages of 95 and 62, respectively) and for low-income countries (which had averages of 90 and 60, respectively) at the completion of the MDG period. The achievements in the reduction of the child mortality rates were attributed to family planning interventions in Ethiopia (Yigzaw et al. 2015) and the intensification of primary health service coverage, which covered 94% of the population in 2013/14 (UNDP 2015).
Maternal mortality rates per 100,000 live births for Ethiopia dropped by 71% from the year 1990-2015, although the country did not meet the respective MDGs target of three-quarters reduction (UNDP 2015). However, the MMR, which was higher (1,250) than the MMR average for SSA (987) and for low-income countries (1,010) in 1990, dropped to 353 and fell far below the SSA average (547) and the average for low-income countries (496) at the completion of the MDG period in 2014/15. Throughout the HSDP period, the MMR for Ethiopia decreased by 65% on average, with the highest rate of reduction recorded (52%) during HSDP-III (Graph 6).
Moreover, Ethiopia reduced the prevalence of HIV among the population aged 15-49 years by more than three-quarters from its peak prevalence rate (4.5) in 2000-1.1 in 2015, which exceeds the MDG target of halving and reversing the trend. Some of the other achievements in the area include: malaria incidence per 1,000 people at risk reduced by 91%, deaths associated with malaria reduced by 87%, TB incidence rate per 100,000 population per year reduced more than half, and percentage of TB cases successfully treated with TB diagnosis and treatment approach (DOTS) 92%. (See also Appendix L) The overall achievements in improving the health status of the country mainly have been related to the HEP, which was developed in 2003/04 as an innovative community-based universal health service development approach (Spicer et al. 2014;Ministry of Health 2014a, 2015bWang et al. 2016). At the completion period of the HSDP in 2014/15, Ethiopia had mobilized more than 38,000 female health extension workers (HEWs) and 442,000 health development army (HAD) groups to cover more than 12 million model household families (Ministry of Health 2015b) at a lower cost than before (Balabanova et al. 2013;Wang et al. 2016). Health Graph 5. Child mortality rate per 1,000 live births in Ethiopia vs. SSA and low-income countries . Source: Authors' calculation based on data by UN/DESA (2017).
Graph 6. Maternal mortality rate per 100,000 live births in Ethiopia vs. SSA and low-income countries . Source: Authors' calculation based on data by UN/DESA (2017) and World Bank (2018). infrastructure development, especially the primary healthcare units, also showed a massive improvement. There were only 96 hospitals, 282 health centers, and 802 health posts at the beginning of the HSDP in 1997, but at the end of the HSDP period the number of health facilities had significantly increased to 311 hospitals, 3,547 health centers, and 16,440 health posts (Ministry of Health 1997/98-2014/15, 2015bReich et al. 2016). This increased access to primary health services at the grassroots level in a flexible and equitable health delivery approach (Balabanova et al. 2013;McIntyre and Meheus 2013) contributed to healthcare access and health service quality 'beyond what was expected'.

Conclusions
Ethiopia, through its visionary HSDPs, has achieved fast progress and impressive results in the health sector. This is largely due to the joint and consistent efforts of the Ministry of Health and its health partners as part of country-owned and government-led health development coordination. This study provides evidence that the accomplishments in the health sector can be credited to the synergetic concurrence of health plans, the leadership of the Ministry of Health, health development assistance from the DPs, and the joint efforts of both parties through coordination platforms.
First, the HSDP set the foundation for country ownership. Principally, the woreda-based planning exercises in Ethiopia, as part of 'One Plan, One Budget, One Report' principles, further boosted country ownership, which extends from the Ministry at the federal level to the local and community level (through the HEP). In addition, the health sector goals and its clear priorities were the basis for alignment of the health DPs with the government's priorities and essential to the country-led health development coordination. At the same time, these health priorities and targets set result frameworks for the partnership between the DPs and the government, so that the DPs committed to bring about changes in the health sector.
Second, given the conducive health policy and program, the Ministry was strong enough to take the 'driver's seat' in the designing and implementation of its own health development agenda. The political commitment and professional capability of the top leadership and its firm commitment to exercise country ownership of health development played a significant role in bringing about results. The government also showed ownership by increasing public health expenditure, as well as a strong commitment to the priorities of the health sector. The total health expenditure for Ethiopia increased throughout the implementation of the HSDP; the share contributed by the government in terms of per capita expenditure and as a percentage of total health expenditure, as well as a percentage of GDP, have risen constantly. The low level of corruption in the Ministry, which has been exceptional in Africa, also contributed to the strengthening of trust in country ownership in the sector.
Third, looking at the principles of the Paris Declaration (ownership, alignment, harmonization), the commitment of the DPs to support the sector with a substantial amount of health aid and their high level of engagement in the health development coordination platforms were indispensable and imperative to the remarkable results achieved by Ethiopia in the health sector. Given its level, aid was also of a 'high quality', because it was consistent and well aligned with the goals and priorities of the health sector programs. The DPs shared the ownership of the health programs and showed a strong commitment to align and harmonize their support with local demands through the preferred financing channels, like the MDG PF, which gave more leverage for the Ministry to develop solid ownership over its health development agenda.
Fourth, the coordination platforms have also been essential units in multiplying the joint efforts of the Ministry and the DPs in the country-led health development coordination process. These platforms were essential to sector-level policy dialogue and joint decisions in relation to financing and coordinating all the efforts in the health sector. The good results in the health sector attracted both big and small donors to the coordination platforms, making the health sector a 'donor darling' sector in a 'donor darling' country.
However, there are also challenges in Ethiopia's health sector that call for improved coordination. Primarily, there is the high level of dependence by the health sector on external financing, which is distorting the structure of health financing in Ethiopia, even though it is low compared to some SSA countries like Malawi. The other major challenge is that aid fragmentation, which increased after the Paris Declaration, is high in the sector. This is mainly due to small-sized bilateral donors involved in the population policies and reproductive health sub-sector. On top of that, almost half of the donors' contributions, like that of the Global Fund, goes through different competing channels, which creates parallel structures in the Ministry of Health, and, in turn, creates parallel accounting and power centers. However, the coordination platforms and the joint financing arrangements contributed to reducing the effects of aid fragmentation and increasing aid effectiveness in the sector. Ethiopia can thus be seen as an example of the fact that aid fragmentation should not be a major problem when most of the aid is aligned with local ownership and coordinated in joint financing arrangements in 'One Plan, One Budget, One Report' practices. Private sector and civil society participation in high-level health development coordination platforms, however, is also almost absent, and the coordination capacity declines as it goes down to the regional and local levels. Also here improvements could be found.
The HEP has been exemplary in promoting a community-based and innovative primary health service delivery approach, which brought about significant change in the health sector and has changed the health system structure (Burki 2016;Mullan 2016) and helped the country to realize national and international health development goals and targets (McIntyre and Meheus 2013;Admasu, Balcha, and Getahun 2016;Reich et al. 2016). The role of the leadership in achieving these results, with less investment than anticipated, has attracted more backing for the health sector from the DPs and increased the level of country ownership in the health sector. Generally, a lesson can be drawn that aid works if anchored in country and community ownership and when donors increase their commitment in strong partnership and managed according to the aid effectiveness principles. This combination of (1) strong and professional leadership (also in combating corruption), (2) visionary plans which are also based on the lowest level inputs, (3) alignment of the donors with plans and financing instruments, and (4) an emphasis on Primary Health Care with capacity building at the lowest levels should or could be an example for other countries in Sub-Saharan Africa.