The practice of fiscal decentralization at local level in Ethiopia: Evidence from Oromia and Afar Regional States

Abstract In the past half century, fiscal decentralization has emerged as a critical policy issue among public finance scholars and policymakers around the world. Since 2001, Ethiopia has practiced fiscal decentralization at the district level by devolving certain functions and financial resources for the District Administration. The primary goal of this paper was to assess fiscal decentralization practices at the district level, as well as the challenges that local level authorities faced while implementing it. The paper focused on the intergovernmental fiscal relationship between the Regional and Local Governments (LG) of Oromia (representing the relatively “developed regions”) and Afar (representing the “emerging regions”) Regional States. We selected these regional states, not randomly, but purposively using various criteria such as fiscal capacity, population size and geographical area/distance, and infrastructure and human development. We used both primary (in-depth interviews and field observations) and secondary (published and unpublished sources) data. The key informants for in-depth interview and discussions were selected purposively from regional and Woreda institutions on the basis of their levels of knowledge on the subject under investigation while, the secondary data were analyzed and interpreted through a descriptive technique and presented in form of tables and figures in light of the study’s objective. The major findings show that the regional governments’ devolution of expenditure responsibility to the LGs is not accompanied by adequate financial resources, resulting in high vertical and horizontal fiscal imbalances in Ethiopia. As a result, LGs rely heavily on regional blocked grants to cover the costs of the functions devolved to them by regional governments. This, in turn, limits local governments’ fiscal autonomy and leads to less efficiency in public service delivery and accountability at the local level. This necessitates increasing LGs’ fiscal autonomy by devolving taxation authority in order to reap the potential benefits of effective fiscal decentralization system at the sub-national and national levels.


PUBLIC INTEREST STATEMENT
Despite the topic of fiscal decentralization has been receiving high attention among policymakers, government agencies, researchers, and Civil Society Organizations in Ethiopia recently, little study has been done about its achievements and challenges, especially at the local level.Accordingly, this study attempts to critically assess the major achievements and challenges related to the implementation of fiscal decentralization at local government units in Ethiopia by selecting two regional States, namely, Oromia and Afar.It informs factors that promotes or challenges the implementation of fiscal decentralization at local governments to policy-makers of other regions and use as input in designing and implementing an effective decentralization in their jurisdictions.It provides a comprehensive information and philosophies about the "Woredas Decentralization Program" for various decentralization stakeholders in Ethiopia.Finally, it also fills the existing literature gap on the subject matter and serve as a reference for researchers in the future.sources) data.The key informants for in-depth interview and discussions were selected purposively from regional and Woreda institutions on the basis of their levels of knowledge on the subject under investigation while, the secondary data were analyzed and interpreted through a descriptive technique and presented in form of tables and figures in light of the study's objective.The major findings show that the regional governments' devolution of expenditure responsibility to the LGs is not accompanied by adequate financial resources, resulting in high vertical and horizontal fiscal imbalances in Ethiopia.As a result, LGs rely heavily on regional blocked grants to cover the costs of the functions devolved to them by regional governments.This, in turn, limits local governments' fiscal autonomy and leads to less efficiency in public service delivery and accountability at the local level.This necessitates increasing LGs' fiscal autonomy by devolving taxation authority in order to reap the potential benefits of effective fiscal decentralization system at the sub-national and national levels.

Introduction
Fiscal decentralization, the transfer of power and fiscal responsibilities from the central government to the sub-national governments, has been receiving attention recently among academicians and policy-maker across countries around the world (Thiessen, 2003;Litvack et al., 1998).The main reasons behind the increase interest toward a decentralized fiscal system are: failure of a centralized system in providing efficient public services and promote economic growth; transition to market economy, and growing need to recognize social, economic, and political diversity across regions within a nation (Tanzi, 2002).The basic economic rationale for a decentralized fiscal system is that it has potential to improve the efficiency of economic resources allocation (leads to a maximum society welfare) by matching the supply of public services with the preferences and needs of local people as compared to the centralized system, which provides a uniform public services across all local jurisdictions (R. M. Musgrave, 1959;W. Oates, 1972).Moreover, competition (both vertical and horizontal) among different levels of government could improves efficiency of public service delivery at sub-national levels (Brennan & Buchanan, 1980) Additionally, it also improves the local governance quality through enhancing the responsiveness of local authorities and accountability to local people by improving the political participation of local people ( (Smoke, 2004).
However, fiscal decentralization has also costs if it is implemented in a wrong way that means prior establishment of an effective political and administrative institutions at local levels, including: reducing the efficient of resource allocation; increasing social inequity and unbalance development across regions; and promoting bad governance and corruption at local level (Prud'homme, 1995;Tanzi, 2002).Thus, each country should develop its own optimal fiscal decentralization structure by taking into account its specific conditions so as to realize the potential benefits of fiscal decentralization.
After practicing a highly centralized administrative setup, political and fiscal systems for several decades, Ethiopia has introduced a decentralized and federal form of government de-facto in 1991 and de-jure in 1995, with the aim of ensuring self-administration at sub-national government levels (Solomon, 2006).The government was implemented a decentralization system (policy) in two phases.The first phase was implemented in 1991 by devolving power and responsibility from the central to regional governments and aimed at creating and empowering regional national governments (Tegegne, 1998).The second phase of decentralization was implemented by adopting a District Level Decentralization Program (DLDP) in 2002, with the aim of devolving power and responsibilities to the Woreda level (district) of government (Kassahun & Tegegne, 2004).The DLDP was initially launched in 430 Woredas covering the four big Regional States (like Tigray, Oromia, Amhara, and SNNP) during [2002][2003][2004], but was subsequently implemented in the Wordas of the remaining Regional States and the two city administrations (Addis Ababa and Dire Dawa) during 2005-2008.Following the implementation of DLDP, regional governments in Ethiopia have been devolution more power and fiscal responsibilities to Woredas (districts) since 2002, with the aim of improving the efficient public services as well as enhance the participation of local communities in political and development affairs in order promote local good governance and local economic growth (Kumera, 2006).Although Ethiopia has exercised fiscal decentralization at local levels of government over the last three decades, its effectiveness has not been adequately assess recently.Therefore, this paper attempts to assess the implementation of fiscal decentralization at Woreda (district) level of government in Ethiopia since 2002 by selecting two regional states (Oromia and Afar regional states) and by focusing on parameters like power and function devolution, financial capacity, fiscal autonomy and service delivery.The study has not selected these Regional States randomly, instead, using certain criteria like socio-economic development and infrastructure development.
To achieve the research objectives, the paper has attempted to address the following research questions related to the implementation of fiscal decentralization at local level in Ethiopia This paper is organized as follows: Introduction; Section 2 focus on conceptual framework of the study; Section 3 focus on research design and methodology of the study; Section 4 presents the results and discussion of the study; finally Section 5 presents the conclusion and recommendations.

Definition and rationales of fiscal decentralization
Decentralization commonly defined as "the transferring of authority and responsibility for public functions from a higher level of government to lower levels of governments which ultimately contributes to empowerment of people through their effective engagement and participation in the affairs of local government".It is also defined as the transfer of responsibility and authority for planning, management and the raising and allocation of resources from the central government to lower levels of government.Fiscal decentralization, widely treats as core aspect of a decentralized of government, can be defined as 'the transferring of power and responsibilities for public functions from the control of central authorities to the sub-national governments (regional, province and municipality) with regard to spending and revenue collection.
The basic economic rationale with regard that a decentralized fiscal system is enhances the allocation efficiency of economic resources (as the result maximum social welfare) by supplying public services as per the needs and preferences of local residents.The rationale behind such argument is that a decentralized government (closer to a local people) have better information about the demand or preferences of local citizens than of a remote central government (W.Oates, 1972).Moreover, it also fosters the responsiveness of local authorities and greater accountability to local people by bringing the voice of local communities into the center of policy-making processes.Additionally, it helps to create "laboratories" for innovation and experimentation and innovation of various public service delivery techniques, which serve as models for the latter implementation by other local jurisdictions.Finally, it enhances fiscal discipline and public service delivery efficiency through competition among regional and local governments, which results in delivery at sub-national levels (Brennan & Buchanan, 1980).
However, if fiscal decentralization measures are designed and implemented in a wrong way, that is, without considering the specific circumstances of a country, it may lead to inefficient of resource allocation; increase economic development inequality across regions within a nation; limit the central government to implement redistributive and macroeconomic stabilization policy effectively; and promote bad governance and corruption at local level, due to capturing of local government by local elites (Prud'homme, 1995;Tanzi, 2002;World Bank, 2008).Thus, each country should be developed its own optimal fiscal decentralization structure by taking into account its specific conditions so as to realize the potential benefits of fiscal decentralization.However, there is also a general principle which states finance should follow the public function.Accordingly, subnational governments should endow sufficient taxing power, at margin, that can enables them to finance the cost of functions they are responsible for, so as to realize an effective fiscal decentralization.
Although fiscal decentralization has been become an important topic among policy-makers, academicians, civil society organizations in Ethiopia for the past two decades, a few research work has been conducted with regarding to the subject in Ethiopia, especially at local level of government.Worku (2005) and Meheret (2007), for instance, highlighted the major challenges which constrained to implement an effective decentralized system at local levels of government in Ethiopia As they pointed out, weak administrative and institutional capacity at grass root levels due to lack of well-trained human resource, lack of clarity of power and the assignment of fiscal responsibilities between regional and local governments, top-down decision, authority structures afflicting the regional system and lack of sufficient financial resources, lack of availability of experienced human resources, and tax autonomy as well as lack of decision-making autonomy on the allocation of capital expenditure at local levels of government are considered as major challenges.The implication is that there is problem with regard to the assignment of expenditure and taxing responsibilities between regional and local governments in Ethiopia.Thus, it is essential to assess the implementation of fiscal decentralization at local levels, particularly at Woreda (district) government, in Ethiopia for several reasons.First, the study will have a great significant in providing a detailed and comprehensive information related to the process of fiscal decentralization in Ethiopia for all decentralization stakeholders.Finally, the study will fill the existing literature gap on the area and serve as a reference for researchers in the future.Smoke (2004) identifies the basic preconditions that leads to an effective fiscal decentralization system.Those includes: an adequate enabling environment; assignment of an appropriate set of functions local governments; assignment of own revenue sources to local governments; design and implement of intergovernmental fiscal transfer system and adequate local access to financial markets.Here, it important to note that all those pillars must be operated together in order to realize the potential benefits of an effective fiscal decentralization.

An adequate enabling environment
It can begin with constitutional or legal mandates for some minimum level of power and fiscal responsibilities to the local governments.This provides a foundation in which to build decentralization, but it does not by any means guarantee for effective fiscal decentralization.

Expenditure responsibility
Basically refers to allocation of the authority and responsibility for the public-sector decisions among different levels of government.It is the first fundamental step for designing fiscal decentralization, because assignment of expenditure responsibilities should come first before financing these expenditure responsibilities.Thus, it is important to clearly define the expenditure responsibilities of various levels of government before allocating revenue sources to them, so as to enhance accountability, avoid unproductive overlap, duplication of authority and legal challenges (World Bank, 2003, p. 117).
According to R. M. Musgrave (1959), government in an economy has three major functions, namely, macro-economy stabilization, redistributive income and wealth; resource allocation.The theory (FGT) argues that the macro-economic stabilization function should be carried out only by central government.The rationale is that if decentralized governments undertakes the monetary and fiscal policies individually, it may leads to ineffective outcomes, due to spillover effects arising from the open nature of sub-national economies (Musgrave, 1989).The theory also applied the same argument in undertaking of the redistributive function by central government only.The logic is that decentralized governments are ineffective to ensure income equity at national level using taxes as instrument, due to mobility of production factors like capital and labor for tax reason (W.Oates, 1972, p. 5).
On the other hand, FGT argues that decentralized governments should have a predominate role in resource allocation such as provision of efficient regional and local public goods and services for citizens.The rationale is that decentralized units of government have better knowledge about the preferences and conditions of local people than of the central government, as a result they can improve the efficient of economic resource allocation by matching the supply of public services with the preferences of local residents (W.E. Oates, 2005).While, the responsibility for provision of public services that have spillover effects beyond two or more regional jurisdictions (like Defense, foreign relation) should be provided to the central government.The reason is that, centralized these public functions would improve the efficiency of resource allocation and leads to maximize welfare society as whole, due to realize the benefit of economic scale and internalize the interregional externalities.Thus, the resource allocation function should be carried out by units of government at various levels.

Revenue sources assignment
This concerns basically with which levels of government should has taxing power on what revenue sources (tax base) and which level of government should administer the tax bases (Ambrosania & Massimo, 2006).There is a great consensus among scholars with regard the devolution adequate financial resources to local governments that enable them to finance the costs of functions they are responsible for by constitution or law, so as to realize the benefits of fiscal decentralization.The traditional approach (FGT) argues that tax bases with predominant stabilization effects like corporate and progressive personal income taxes; international trades (customs duties); non-benefit mobile tax bases; unevenly distributed tax bases (like natural resource); and taxes that involve considerable administrative economies of scale (like VAT), should be assigned to central government (W.Oates, 1972, p. 137 andMusgrave, 1999).It also suggest that decentralized governments should impose benefit taxes such as service charges or fees on residents as payment for public services which they provided.Additionally, they should also levy taxes on relatively immobile tax bases like property tax and benefit taxes on mobile factors of production when they receive benefits from the public services provided by them (McLure, 2001, p. 14).

Designing an appropriate intergovernmental transfer system
When the expenditure and taxing responsibility is allocated among various levels of government on the basis of traditional theories of fiscal federalism (FGT), it often creates fiscal problems (both vertical and horizontal fiscal imbalances).Vertical fiscal imbalance refers to the mismatch between the cost of functions that are assigned for SNGs and the available financial resources at local levels.This occur when the taxation power on major tax bases assigned to the central governments; while the responsibility for provision of several public service that have higher costs for subnational governments for efficiency reason.This, in turn, limits SNGs to cover the costs of services they are responsible for by their own revenues (Dafflon, 2006, p. 281;Shah, 2007, p. 29).Horizontal fiscal imbalance, on the other hand, refers to the gap between revenue-raising capacity and expenditure needs of government units presented at the same level.As the result, some SNGs with poor fiscal capacity would impose higher tax rates to provide similar quality and quantity of public services for their residents.This require design of fiscal equalization schemes to close horizontal fiscal imbalance and enable all SNGs to provide comparable public services at comparable tax efforts to all citizens of a nation regardless of their choice of living location.
Generally, many developed and developing countries around the world have designed and implemented various intergovernmental transfer system to achieve different national objectives (UNDP, 2003).However, the process, extent, the nature of transfers to address the fiscal imbalances vary from one country to another, since countries vary in terms of specific conditions.Some countries design the transfer systems to be very decentralizing in nature while others are centralized.According to Boadway (2007: 57-62), there are several economic rationales or reasons for intergovernmental transfer or grants, includes: addressing both vertical and horizontal fiscal imbalances; eliminating inter-jurisdictional externalities and provision of a minimum standard public services.Intergovernmental transfer tools generally can be grouped into two major categories, namely, unconditional (general purpose) and conditional (specific purpose) grants.

Local access to capital markets
Refers generally to the capability of local governments to borrow money from capital markets to finance the functions they are constitutionally responsible for.Local borrowing is appropriate for financing only capital-intensive projects with yield rates above the interest rate and the ability to pay back total investment cost within due dates (Bailey, 1999).Theory of fiscal federalism argues that local access to capital markets is important for three reasons: the resource allocation efficiency; intergenerational equity, and distribution and macroeconomic considerations (Ebel & Yilmaz, 2002).The argument behind the intergenerational equity is that the burden and benefit of capital project should be spread out across generations through debts.Thus, local authorities need not be allowed to finance the current year budget deficits through long-term debt, so as to avoid transferring of debt burden to future generations without any additional physical assets.Similarly, financing the capital projects through imposing higher tax rates on current generation is also undermines intergenerational equity, because future generation will receive benefits from the capital projects without paying for it.
In practice, however, local governments in developing countries are restricted by central government to borrow funds extensively for financing the investment on capital projects in order to ensure macroeconomic stability in long run (Prud'homme, 1995).But, many countries around the world are introduced various debt control mechanisms (such as market discipline, fiscal rule.Cooperative and Administrative approaches) to preserve macroeconomic stability and ensure fiscal discipline at sub-national levels.Of these debt control approaches, the market-based approach is not an effective mechanism to discipline the local borrowings in developing countries such as Ethiopia, due to weak development of capital market.

Research method
There are several types of research methods in social science to choose from.For the purpose of this study, a case study research approach is employed.This is because; the purpose of the study is to provide and gather sufficient information and obtain a deeper understanding with regard the implementation of fiscal decentralization at woreda (district) level in Oromia and Afar regions.The main advantage of this research method is that it effectively deals with and a full variety evidences from various data collection techniques such as interview, Focus Group Discussions, field observation and documents.Generally, the study has employed both qualitative and quantitative approaches to achieve the research objective.

Sampling design and selection of the study area
The study has taken Woreda Administrations as units of analysis to assess the implementation of fiscal decentralization at local level of governments in Ethiopia The reason is that Woredas (districts) are considered as a unit of government for meaningful social and economic development at local level in Ethiopia.However, study has taken into account Woreda governments, which are located in Oromia and Afar regions, as representative of other Woreda governments located in the remaining regions of the country.The study selected these regional states purposively on the basis of various criteria such as fiscal capacity, population size and geographical area/distance, infrastructure and human development, and ethnic diversities.The Oromia Regional State is selected to represent the relatively more developed Regional States (Tigray; Amhara, and SNNPR), whereas Afar Regional State is selected to represent the relatively less developed Regional States (like Benshangul-Gumuz, Gambella, Harari, and Somali).
Particularly, Oromia Region is the most populous region (cover about 35% of total national population) and has the largest geographical area coverage with a total land area of 359,620 square kilometers.The Oromia region' economy is based on agriculture (rain-fed, subsistence), contributing about 45% of the region's GDP and supports 83% livelihood of the population.The region is contribute a larger share for Ethiopia' agricultural export: coffee, hides and skins, pulses and oil seeds.While, Afar Region is the less populous and low population density of 14.9 person per square kilometers region in Ethiopia with a total land area of 92,000 square kilometers.It is one of the regions with a very (Afar Regional State Council, 2010).
For the purpose of primary data collection, the study purposely selected two Woreda Administrations, namely, Gewane and Gelan from the Afar and Oromia regions, respectively, on the basis of convenience and accessibility.The researcher believed that the selected two Woreda Governments are core and represents the selected Regional States better in terms of their political, economic and social activities.

Data source and collection techniques
The study collected and analyzed both primary and secondary data sources.The primary data were collected through conducted an in-depth interviews with key informants (like politicians; plan, program and budget experts, public officials and staffs, and sector office heads).The key informants were selected purposively from regional and Woreda institutions on the basis of their levels of knowledge on the subject under investigation.That means, the key-informant are selected on the basis of level of knowledge related to fiscal decentralization issues and its practices at Woreda level due to their political position and professional capacity.The data from the in-depth interviews were used to identify the factors that promote or challenge facing the implementation of fiscal decentralization at woreda (district), as well as to substantiate data gathered from secondary sources.The study also used focus group discussions (FGDs) to gather primary data related to the problems and challenges facing while implementing fiscal decentralization at local level.Accordingly, two FGDs were held in Gowane Woreda (Afar region) and Gelan Woreda (Oromia region).In each FGD, there were 5-8 individuals from communities with different background, status and age and sex categories.Besides, the study also used checklist-guided field (office) observations as a means of gathering primary data.Accordingly, the researcher, while stayed for over one month in a field work, had opportunity to watch, lesson and informally communicate with workers of various institutions at regional and Woreda levels of government in Oromia and Afar regions The researcher also had opportunity to observe corresponding official letter between regional and Woreda (district) levels of government and the challenge Woredas' public officials faced while exercising the power and fiscal responsibilities which devolved for them by law.
The study also collected secondary data from published and unpublished materials, which available in form of books and journal articles, and Federal and regional' constitutions, proclamations, regulations, policy statements, census reports, statistical bulletins, and annual plans and actual annual reports of the institutions at regional and local levels; as well as research studies which conducted and presented for policy discussion and conferences, etc. Accordingly, panel data (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015) related to fiscal decentralization in Oromia and Afar regions, such as on budget allocation and regional block grants as well as actual expenditure and revenue were gathered from the regional institutions annual reports and thoroughly analyzed and interpreted in the study.

Method of data analysis
The study mainly employed descriptive data analysis method to analyze data and information gathered from both primary and secondary sources.Accordingly, the primary data obtained through interviews, focus groups, and field observations were analyzed using qualitative data analysis method and supplemented with the secondary data.Similarly, data gathered from secondary sources (such as budget allocation, fiscal transfers, actual revenues, and expenditures) were analyzed using descriptive statistical analysis techniques such as average, percentage, and ratio values.After the raw data were converted into ratios such as expenditure and revenue ratios, they were presented in the form of tables and then analyzed and interpreted using the descriptive technique in light of the study's objective.To achieve the maximum validity of the data, the researcher employed the triangulation method.

Ethical Issue
In this research, the researcher has made necessary efforts to address the ethical considerations of confidentiality and privacy.To maintain the confidentiality of information, guarantee was given to the participants that their names should not be associated with their responses during the coding and recording processes.Besides, the participants are involved in the interview and discussions based on their expressed willingness and informed consent.Accordingly, the researcher was given verbal and written description of the study for the participants to obtain their consent.The researcher also provided refreshment for the participants after obtaining information and appreciated them for providing information and their time for participating in the study.

Results and Major Findings
In this section of the study, the paper presents the major finding in line with objective of the study.

The powers and functions Assigned to local (district) governments
The powers and functions of regional and local governments are no defined in FDRE Constitution, since the federal constitution has explicitly specify the powers and functions of federal government and regional states only under articles of 51 and 52, respectively.Article 50 (4) of FDRE constitution, however, requires the regional governments to devolve significant power and fiscal responsibilities to their respective local governments that enable them to function as institutions of self-governance.The implication is that the devolution of powers and fiscal responsibilities to local governments are made discretionarily by regional governments in Ethiopia.
In the case of Oromia and Afar regions, the process of powers and functions assignment between regional and local governments has been done through a top-down approach (i.e.without involving local communities and local authorities).As the result, there is lock of clarity with regarding the assignment of power and functional responsibilities between regional and local governments.This, in turn, create overlapping responsibilities, particularly on shared and would conflict between the two levels governments.Thereby, it reduces the efficiency of public service delivery and difficulty to ensure accountability at regional and district levels of government.
The practice shows that regional governments in the regions under study are devolve power and functional responsibility for administration and provision of public services that have local benefit effects such as primary education and health-care services, water and sanitation, agriculture and rural development, inter-district roads and natural resources conservation) to their respective local governments (particularly districts).In addition to that, the regional governments are also delegated certain functions that may produce inter-jurisdictional externalities and have distribution implications such as provision of secondary schools, TVET, and hospital level health services to the respective woreda (district) governments.This indicates, woreda governments in the regions have responsibility for provision public services that promote the local socio-economy development and improve the living-standard of local people.The practice, however, shows that the woreda governments are ineffective in carrying out some functions that are delegated them by regional governments, due to lack of availability of experienced and competent human resources, as well as financial resources.In Afar, for instance, key-informants claimed that while woredas (districts) in the region have responsibility for providing water supply and major roads, these services are actually provided by regional sector bureaus rather than the districts.
Following the implementation of the DLDP in Ethiopia, the regional governments in Oromia and Afar regions have been devolved some powers to their respective Woreda Councils with the goal of ensuring self-administration at the local level.As the result, Woreda governments empowered to make decisions on matters that affect their internal affairs directly such as political, economic, social, and administrative decisions, without the influences of zonal administrations.According to interview made with key informants at regional level, Woredas in the regions currently make budget decisions on Woreda blocked grants independently and are not required to submit their annual budgets and expenditure plans to zone administrations for approval.The practice, however, shows that the elected Woreda Council members of some Woredas in the Afar region are not yet exercised the powers devolved to them by regional government.They often meet regularly only to approve the annual plans and budgets, which have been formulated by the higher levels of government.Moreover, some of the power and functions which devolved for local governments are still carrying out by regional sector bureaus.This is due to the fact that most of local governments in the study regions are suffer from capacity problems, which includes: financial, human and material resources; as well as administrative and technical capacity.However, the Woredas in Oromia region are in better position in term of capacity mentioned early when compare with Woredas in Afar region.As a result, most of Woredas in the regions under study, especially in Afar region, are highly vulnerable to the interference and manipulation of regional organs and authorities.Thus, it can be argue that there is a gap between the constitutional provisions and the reality on ground.

The Degree of Expenditure Decentralization
This paper used the ratio of Woreda governments' expenditure to the total regional state expenditure to measure the extent of expenditure (de) centralization in Oromia and Afar regional states.As can be seen from Table 1, the expenditure share of Woreda to total Regional States expenditure is low in both regions under study, which account on average about 40% and 33% in Afar and Oromia regions, respectively, during 2007/2008-2015/2016 fiscal years.This indicates that expenditure decisions are highly dominated by regional government in both regions over the study periods.Moreover, the expenditure share of woredas has shown an increment trend over time (from 8% in 2007 to 38% in 2015/2016) in Oromia region, but it shown a decline trend (from 51% to 32%) in Afar region.The implication is that, the role and performances of the woredas in the provision of public service delivery in Oromia region has been shown an improvement over recent periods.
According to the discussion with FGD, the Woredas in Oromia regions have been shown an improvement in delivery basic public services (like education, health, agriculture, water and rural road) in terms of quality and quantity after the implementation of fiscal decentralization at local government units.Thus, it can be argued that such finding match with the theoretical argument that suggests that fiscal decentralization can improve the efficiency of public services at local level by enabling the local authorities to supply public services as per the needs and preferences of local people.However, in reality the real contribution of local governments for the improvement in basic public services delivery is very limited.The reason is that most of basic public services facilities constructed (like primary school, health posts and clinics, shallow wells) in the Woredas of Oromia region are built by regional and zonal sector bureaus as well as international and local NGOs supported by foreign donors (like World Bank, UNDP, UNICEF, etc.).Thus, when we take into account actors outside the woreda institutions, the role and performances of local governments in delivery of basic public services delivery is very minimum.Factors contributed for Woredas' poor performance includes: weak managerial and technical capacity, due to lack of availability of competent human power; lack of availability of sufficient financial and material resources.
Numerous factors can be identified as contributing to the existence of a high degree of expenditure centralization in both regions.First, poor political and leadership commitment on the part of regional authorities to decentralize the expenditure responsibilities.The regional sector bureaucracies have been hesitant to decentralize the public services devolved to the Woreda by the regional constitutions.Regional authorities explain that they are concerned about losing their prestige and resources to Woreda authorities.Regional authorities frequently regard local authorities as competitors rather than collaborators in the administrative hierarchy.Such a perception stifles regional decentralization processes.Second, the regional sector offices believe that the Woreda governments lack administrative and institutional capacity to handle all of the expenditure responsibilities that are legally assigned to them.As a result, they use the Woredas' limited capacity to justify keeping public functions centralized at the regional bureau levels (World Bank, 2004).

Fiscal autonomy of local level governments in Oromia and Afar Regions
Local spending autonomy refers to the extent of local governments to have the right and ability to determine the nature and size of overall public expenditure and break it down into various public goods and services based on the needs and preferences of local people, without being influenced by higher levels of government (Dafflon and Madiès, 2009).It is better measured by using nonfinancial indicators such as policy decision-making power, input autonomy, monitoring and evaluation autonomy, and political decentralization to ensure the accountability of local authorities.Thus, the qualitative indicators were also used in this paper to assess the extent of Fiscal autonomy of local governments in Afar and Oromia regional states.
Despite Woreda governments in the regions under study are established to acts as independent and autonomous local administrative units, in reality they do not have the decision-making autonomy in developing own policies and strategies as well as preparing own budgets.The key informants responded that decision with respect to the allocation of capital budgets into various sectors and the implementation of most capital projects (like construction of schools, digging wells and health units) are carried out by the regional sector bureaus in the two regions.The key informants also responded that Woredas in the regions have no decision-making autonomy to hire, promote, transfer, and impose disciplinary measures on their civil servants.Most Woredas in Afar region, for instance, the decisions to hire and promote civil servants in remote areas are carrying out by regional sector bureaus.In addition to that, Woredas in both regions lack decisionmaking autonomy when it comes in setting standards for local public services they responsible to deliver, mainly due to weak technical and managerial capacity.Finally, the Woredas do not have the autonomy to monitor and evaluating the efficiency and effectiveness of capital budget spending.This is because the monitoring and evaluation of most capital projects in the regional states has been carried out through regional sector bureaus.Regional sector bureau officials explain that local administrative levels lack an effective information system for gathering data and reporting the level of capital budget spending.
Generally, Woreda governments in the regions under study have limited decision-making autonomy with regard to political, fiscal and administrative matters that affects directly their affairs.This suggests that there is a gap between the constitutional provisions and the reality on the ground in terms of local government spending autonomy in both Afar and Oromia regional states.There are several factors contributes to the Woredas' low decision-making autonomy in the regions under study.First, there is poor political and leadership commitment on the part of regional authorities to decentralize power and the expenditure responsibilities to Woreda governments.The regional sector bureaucracies have been hesitant to decentralize political power and fiscal responsibilities, which devolved to the Woredas by the regional revised constitutions.Regional authorities explain that they are concerned about losing their prestige and resources to Woreda authorities.Regional authorities frequently regard local authorities as competitors rather than collaborators in the administrative hierarchy.Such a perception stifles regional decentralization processes.Moreover, the regional sector offices believe that the Woreda governments lack administrative and technical capacity as well as adequate financial and material resources to handle all power and expenditure responsibilities that are legally devolved to them.As a result, they use the Woredas' limited capacity to justify keeping political power and fiscal responsibilities centralized at the regional bureau levels (World Bank, 2004).
Second, most of Woredas in the regions under study are facing serious challenges with regard to weak administrative and technical capacity to mobilize funds and utilize the resources in projects that could improve the living standards of their local people (discussion made with key-informants and FGD).This is mainly due to lack availability of trained and experienced human resources as well as necessary material resources infrastructure facilities at local levels.Additionally, almost all Woredas in the regions have no capacity to generate sufficient own revenues that can enable them to meet the increasing socio-economic development demand of their local people.This is mainly due to several factors includes: delegation of few revenue sources that have very low yielding rates; lack of tax autonomy; and poor tax administrative capacity as well as unwillingness of taxpayers to pay taxes and corruption.These capacity constraints, in turn, limited the autonomy of local governments to develop own policies and allocate their budgets into recurrent and capital expenditures as well as various sectors according to the needs and preferences of their local communities.Finally, the community participation in the policy formulation, planning and implementation of capital project in the Woredas of regions under study is very limited (field observation and discussion with FGD).

Degree of Revenue Decentralization
From Table 2, shows that Woredas' share of own revenue to total regional revenue is very low, which account on average about 20.1 percent and 6.3 percent in Oromia and Afar, respectively, during the 2007/2008-2012/2013 fiscal years.This indicates, the revenue collection is highly dominated by regional governments, and makes Woredas heavily dependency on regional blocked grants to cover larger portion of the cost of functions they are responsible.Woredas' revenue raising capacity, however, shown an improvement trend (mainly from domestic indirect taxes and non-tax revenues) and reduces the level of their dependency on regional blocked grants (from 93 percent to 87 percent in Oromia region) during the study periods.The practice shows that, despite the Woredas in regions have been made more efforts to generate more own revenues, still yet their level of dependency on regional transfer remain unchanged during the selected periods (refer table 3.2).Consequently, it further limits the autonomy of Woreda authorities to allocate their budget resources in expenditure areas that demanded highly by local residents.This, in turn, could undermine the potential benefit of fiscal decentralization such less efficiency of local public services delivery and less accountability at local levels.
The key informants explained the factors that contributing to high revenue collection centralized in the two regions are: delegation of few revenue sources with very low yield rates; poor tax administrative capacity at local level, due to lack of trained and experienced manpower; lack of availability an appropriate mechanisms and procedures to assess and collect local taxes; unwillingness of tax payers, due to lack of awareness; and lack of tax autonomy.

Own revenue composition of local governments in Oromia and Afar Regions
Table 2 presents the composition and share of the major own revenue sources (taxes and nontaxes) and regional transfers of Woredas to total budget revenues in Oromia and Afar regions for 2007/08-2012/13 fiscal years.Tax revenue is the major own revenue source and it plays an important role at Woreda level, which accounted on average about 88.6 percent and 86 percent of the Woredas' total own revenue in Afar and Oromia regions, respectively.However, the share of tax revenue to total own revenue has shown a declining trend (from 90 percent to 85 percent) in Afar region and (from 93 percent to 83 percent) in Oromia region during the selected periods.
In terms of tax structure, direct tax revenue is the dominant tax revenue source of the Woreda governments in the two regions, which accounted about 79.4 percent and 90.5 percent of the Woredas' total tax revenue in Afar and Oromia regions, respectively.This shows that Woreda governments are generated more tax revenue from direct tax bases than of indirect tax bases during the same periods.Moreover, the contribution of direct tax to total tax revenue of the Woredas has shown an increment trend (from 73 percent in 2007/2008 to 96 percent in 2012/ 2013) in Afar region, while it remain unchanged in Oromia.Out of the total direct tax, the Woredas generated more tax revenue from employment income tax followed by enterprise business profit tax and "Chat tax" during the selected periods.
Despite Afar' economy is based on agriculture (mainly on livestock production), Woredas in the region are generated little taxes (not more than 3 percent of their total direct tax) from livestock income tax and rural land use fees during the selected periods.Key-informants from the Woreda revenue offices in Afar region, explained the factors contributed for the low revenue collection performance from these taxes includes: late implementation of the region land use policy by the Woreda' revenue authorities; and lack of appropriate revenue collection mechanism that suits with the life style of the pastoral community at local level.
In contrast, the share of indirect taxes to total tax revenue of the Woredas is low, which accounted about on average 20.6 percent and 9.7 percent in Afar and Oromia regions, respectively, during the study periods.The reason is that most of indirect tax bases are levied and collected by regional governments in the two regions.The Woredas have been collected indirect tax revenue mainly from value added tax (VAT), turn over tax (ToT) and excise tax on locally produced goods and services (refers table 3.2).Finally, the share of non-tax revenue to total own revenue of Woredas in the regions is very low, which accounted on average only 11.4 percent and 14.2 percent in Afar and Oromia regions, respectively.The Woredas in the regions have been collected non-tax revenues mainly from charges and fees and sales of public goods and services.

Vertical fiscal imbalance
Vertical fiscal imbalance refers to a mismatch between the costs of functions that LGs responsible for and the revenues available locally to them.This often occurs when taxation on productive revenue sources are allocated to higher levels of government, while responsibility for provision of public services which requires high costs are allocated to LGs, for efficiency and equity reasons.The study applied the share of Woredas' expenditure needs are financed through their own revenue to measure the extent of vertical fiscal imbalance in Oromia and Afar regions.
Form Table 3, on average, the share Woredas' own revenue to their total expenditures was about 10. 57% and 20.36% in Oromia and Afar regions, respectively, during 2007/082015/16 fiscal years.This indicates that, devolution of power and functions for Woredas by regional governments in both regions is not accompanied with devolves of sufficient financial resources.As the result, it create a high vertical fiscal imbalance, which accounted on average about 0.89 and 0.79 during the selected periods.This indicates, Woredas in the regions have no fiscal capacity that can enable them even to finance their recurrent expenditures, and thereby they heavily depend on regional transfers to finance some portion of their recurrent expenditure and full of capital expenditures.Such high mismatch between the revenue sources and expenditure responsibilities could reduce the fiscal autonomy of local authorities and thereby undermines the efficiency of public services delivery and accountability at grass root levels, due to lack of strong linkage between revenue raising and spending decisions.Moreover, although the fiscal capacity of Woredas has shown an improvement (from 1.8 percent to 19 percent in Afar region and from 5.6 percent to 28 percent in Oromia region), the level of dependency of Woredas on regional transfers are remain unchanged over the selected periods.
The factors contributed for the existence of a high vertical fiscal imbalance in the regions under study are: devolution of revenue sources that have very low yielding rates and lack of tax autonomy; poor tax administrative capacity; lack of appropriate enforcement mechanisms and unwillingness of taxpayers to pay taxes.In addition to that, the key informants at regional level also claimed that high expenditure requirements of Woredas due to low-level infrastructure development (especially Woredas in Afar region), increase the salary and wages of civil servants and high corruption practices at local levels also contributed to the high vertical fiscal imbalance in regions under study.

Horizontal fiscal imbalance
Horizontal fiscal imbalance refers to the fiscal capacity gap between of government units presented at the same level.HFI commonly measures by comparing the potential revenue raising capacity with the total expenditure needs of local governments.However, measuring the fiscal capacity gaps between LGs within a region in developing countries includes Ethiopia, due to weak data collection and management capacity at sub-national levels.Thus, this study has used the actual data related to revenue raising capacity and expenditure needs of each woredas in Oromia and Afar regions to measure the extent of fiscal capacity gap in the regions.
The existing data from regional BoFDE in the two regions shows that the existence of a wide horizontal fiscal capacity gaps between the Woredas located within each regions, due to several factors.Some of these are: (i) significant variation between Woredas in terms of revenue raising potential, due to differences in the level of economic and infrastructural developments as well as availability of natural and mineral resources in the regions.In case of Afar region, for instance, the Woredas located in Easter Zone (like Afdera, Elidaar, Amibera, Aysaita and Dubti) have higher revenue raising potential when we compares with Woredas located in Western Zone (like Kuri, Bidu, Adaar, Afambo and Argoba) during 2011/2012 and 2012/ 2013 FYs (refer BoFED annual report, 2014).The reason is that the Woredas in Easter part of Afar region are in better position with respect of economic activity and infrastructure development; urbanization rate, and natural resources (Salt, potassium and Zinc).As the result, these Woredas are collected tax revenues more than half of the region' total PIT and BPT during 2012/2013 fiscal year (Ibid).Second, a significant variation in per capital expenditure needs between Woreda Administration within the two regions, mainly due to difference in terms of sociodemographic characteristics (such as population dispersion, urbanization rate, and social composition and age structure), level of infrastructure developments and physical and climate conditions.For instance, Woredas located in mountain topography of Oromia region are required to spend larger amount of their budgets to provide certain public services like local road as compares with Woredas located in lowland areas.Moreover, Woredas with less density of population are required higher budgets to provide comparable quantity and quality of basic public services, due to lack of from economic of scales.
From the discussion above, the fiscal relationship between the regional and local governments in Oromia and Afar regional states is characterized by high vertical fiscal imbalance and a wide horizontal fiscal capacity gap among the local level governments.This, in turn, requires designing and putting in place an effective fiscal transfer mechanisms that fit with the specific situations of the regions in order to address the fiscal problems.Additionally, in designing a fiscal equalization transfer system, it needs to take into account international norms and principles related to intergovernmental fiscal transfer, such as clarity in grant objectives, autonomy, revenue adequacy, predictability, equity, transparency, and simplicity.

Assessment of Intergovernmental fiscal transfer systems in Oromia and Afar Regions
Intergovernmental fiscal transfer is the major revenue sources of sub-national governments in Ethiopia, accounted about 70% of their total revenue.As the result, they could not able to carry out their functions properly without federal transfers.In Ethiopia, however, the federal governments is not providing transfer funds for local governments directly, but through its regional governments.Hence, Woreda governments in Oromia and Afar regions have been receiving funds in form of Woreda Blocked Grant from regional level of government.In principle, there are two major questions that needs to address in designing a mechanism for transfer funds in form of unconditional grants: how the aggregate grant pool be determined?And how the total grant pool to be distributed among lower levels of government?.
With regarding to the first question, the regional BoFEDs of the two regions are determined the total Woreda block grant pool as part of annual regional budget and approved it by the regional council.The practice shows that the total size of grant pool with respect to the share of grant pool to annual regional budget has inconsistent trends over the past 10 years.The size of Woreda grant pool, for instance, has shown an increment trend over the fiscal years in nominal terms, but inconsistent with respect to the share of Woreda grant to total regional budget.This, in turn, makes the current transfer systems of the regions less stable, unpredictable and non-transparent, which further limits the capability of local governments in preparing an effective budget and planning long-term objectives to provide local public services.This calls for the involvement of all stakeholders in determining the size of Woreda grant, in order to improve the predictability and transparency as well as acceptability of the transfer system of the regional states.
When we concern with the second question, regional governments of the Oromia and Afar regions have been allocating the Woreda blocked grant pool among their respective Woredas and urban administrations on a basis of grant formula that adopted by the Regional BoFED since 2002 and 2005, respectively.The regional governments revised the grant formula several times since that, mainly due to frequently changing in the socio-economic conditions of regional states and lack of reliable data (Meheret, 2007).Pre-2009/2010, the regional governments adopted a grant formulas that considered some socio-economic variables and assigned weights for each variable subjectively.During this period, for instance, however, the pre-2009/2010 grant formulas had severe shortcomings including: discourages Woredas to generate more revenues for their jurisdictions; and undermines the efficiency of local public services.Thus, regional governments of Oromia and Afar regions are replaced the socio-economic variable based formula with a new grant formula (commonly known as Expenditure Need Based Grant Formula) in 2010/2011 and 2009/2010, respectively.The main aim of replacing old grant formula by new grant formula is to match with the real specific circumstances of the Woreda governments located in the regions on ground and improve the fairness distribution of regional block grant.However, key informants claimed that the regional governments did not fully involve all stakeholders such as local executive, local council, and local communities in the process of designing the new grant formulas.Consequently, it reduced the acceptability of the new grant formulas by the authorities at local level.
Overall, the new grant formulas adopted by both regions gives more emphases for expenditure needs for allocating the Woreda block grant among Woreda governments (located in both rural and urban) within the regional boundaries.This indicates that the new grant formulas excluded revenue raising effort which considered in the previous grant formulas.As a result, it discouraged the Woredas to generate more revenues from the sources delegated to them and highly depends on regional transfers.Besides, it is further increase the fiscal capacity disparity between the Woredas within the regions, because the grant formulas failed to consider the variation of Woredas in terms of revenue raising potential.The new grant formulas are used in various expenditure indicators to estimate the capital and recurrent budget needs of each Woredas separately.The formula uses similar indicates to estimate the Capital budget need of each Woreda located in rural and urban areas, with some exceptions.The formulas, for instance, are used in the level of industry and trade development to estimate the capital budget need of each Woredas located in urban areas; while extension service for agriculture, and rural road density to estimate the capital budget of Woredas located in rural areas (refer BoFED annual report of Oromia and Afar, 2014:40).Accordingly, the formulas generally allocated higher capital budget for Woredas that are in worst in terms of level of socio-economic and infrastructure developments than of other Woredas.
The new grant formulas, on the other hand, uses unit cost of each unit of output for each basic services (like education per student, health service per patient, cost of agricultural extension service per household, etc.) to estimate the recurrent budget need of each Woreda (Ibid).Moreover, they use similar expenditure indicators to assess the recurrent budget needs of Woredas, which is located in rural and urban areas, with some exceptions.The grant formulas, for instance, do not consider the expenditure needs for municipality services such as water supply, waste disposal etc, to assess the recurrent budget need of the respective urban administrations.The rationale is that municipality services are exclusively financed from local revenue sources such as service charges and fees.
Finally, the regional blocked grant share of each Woredas is determined by summing the estimated recurrent and capital budgets of each Woredas of all expenditure indicators mentioned in grant formulas.The practice, however, shows that a larger portion of total Woreda block grant allocated in form of recurrent budget for the Woredas in both regions.Of the total Woreda blocked grant in Oromia region, for instance, about 88% and 12% are provided in form of recurrent and capital budgets, respectively, for the Woredas in 2013/2014 FY (BoEFD annual report of Oromia, 2014:41).This indicates that, the new grant formulas of both regions are not provided sufficient financial resources for the Woredas that can enables them to finance larger portions of their capital expenditures.

Access to Financial Market of Woredas (Districts) Governments
In case of Oromia and Afar Regional States, regional Governments endowed right for respective Unban Administrations to borrow funds from domestic financial institutions owned either by federal government or regional governments to finance only their capital expenditures.The practice, however, shows that Urban Administrations in the regions are not exercised their rights to borrow funds from domestic credit sources until now.The reason is that local access to financial markets are subjected to the direct control by the regional administrative council in the regions under study.This indicates, therefore, the existence of a gap between the constitutional provision and the actual practice on the ground with regarding local government's access to domestic credit sources in Ethiopia.Moreover, although fiscal indiscipline of local authorities such as bailout problem is the main issue of regional governments in many federal countries around the world, it is not a serious problem currently in the regions under study.This is because, local governments located in the regions have very limited capacity, especially in Afar region, in utilizing their budget resources including regional transfers effectively, due to lack of managerial and technical capacity.

Discussion
The main findings are summarized on the base the implementation of fiscal decentralization at local level of government (Woredas) by considering the performance of two regional states; namely: Oromia and Afar.With this perspective, the analysis is made on the major pillars of fiscal decentralization, which includes the extent of enabling environment, assignment of appropriate expenditure responsibility and revenue sources with reference regional-local levels of government, vertical and horizontal fiscal imbalances, design of intergovernmental transfer, and adequate local access to capital markets.
With regard to enabling environment, regional governments in Oromia and Afar regions have been devolved more power and fiscal responsibilities to the respective Woredas (districts) following the implementation of Woreda decentralization policy since 2002.However, there is lack of clarity and formality with regarding the assignment of power and fiscal responsibilities between regional and local governments, particularly on shared responsibilities.The reason is that the process of assignment of expenditure responsibilities to Woredas is done solely by regional governments without involved local authorities and communities.Moreover, the revised regional constitutions are not also provided guidance for resolving potential conflicts that could occur between regional and local governments with regarding to the shared responsibilities.As the result, it create overlapping responsibility problem and may leads to several conflicts between regional and local governments.Thereby, it further reduces the efficiency of local public services delivery and accountability at local levels.
The study also found that although the Woreda governments in the regions have some powers to make decisions with regard to political, socio-economic, administrative and fiscal matters that affects their affairs directly, in practice some of these powers are still centralized at higher government levels (regional and zonal level).The reason is that the authorities at regional and zonal levels in the regions often make interferences over the power/authority of Woredas.Factors contribute for the limited fiscal autonomy of local governments in the regions includes: poor political commitment, weak technical and managerial capacity, and poor revenue raising capacity.
When we concerns with the trend of expenditure assignment at Woredas (districts) level, the result shows that expenditure share of Woredas to total regional expenditure is relatively higher in the Oromia region (45%) than of Afar region (33%) during 2007/082015/2016 fiscal years.Moreover, the Woredas were spent approximately 14% and 15% of total regional capital expenditure in Oromia and Afar regions, respectively, during the study period.This shows that expenditure decisions, particularly capital expenditure decisions, are made at regional levels in both regions.This could be due to lack of clearly define the power and functions of local governments; insufficient financial resources and weak managerial and institutional capacity at local level.
With regarding to the revenue assignment, Regional governments are delegated responsibility to collect revenue on certain revenue sources to their respective local governments.The practice, however, shows that Woreda governments in both region have no revenue autonomy on revenue sources delegated them by regional governments.As the result, it reduces the efficiency of public services delivery and accountability at local level.Moreover, when Woredas are collected more than the targeted revenue amount which set by regional' BoFED, an equal amount of money would reduce from their shares of block grant entitlements.Thus, such tax system may discourages the local authorities to generate more own revenue from their jurisdictions and rationalize the expenditures at local level.
Overall, finding of the study shows that devolution of power and functional responsibilities to local governments by regional governments is not accompanied with devolution of sufficient financial resources and in both regional.Woredas in the regions under study, for instance, have capacity to generate own revenue that can only finance on average about 10.6% and 20.4% of their total expenditure in Oromia and Afar regions, respectively, during the study periods.This shows that the Woredas are depends on average about 90% and 80% of their total budgets in Afar and Oromia regions, respectively, on regional transfers to cover the costs of functions they are responsible for.This indicates, the Woredas in Afar region have better financial capacity as compares to the Woredas located in Oromia region.The implication is that, the Woreda level decentralization program is relatively implemented in effective manner in the Oromia regional state as compared to the Afar regional state.As a result, the Woreda Administrations in Afar region are highly dependency on regional grants and makes them more vulnerable to the manipulation and interference of regional governments.This, in turn, limited the Woredas' fiscal autonomy and led to less efficiency of public service delivery and accountability at local levels.The factors contributing to very low Woredas' financial capacity in region includes: delegation of a few revenue sources that have unproductive nature; lack of tax autonomy; poor tax administrative and enforcement local capacities; lack of awareness of taxpayers to pay taxes honestly and corruption.Additionally, there is a high horizontal fiscal capacity gap fiscal between local governments within the regions understudy, mainly due to significant differences in revenue-raising potential and expenditure needs.
In addition to own revenue, local governments are receiving funds in the form of block grant from regional governments, with aim of addressing both vertical and horizontal fiscal imbalances.The regional governments have been distributing the block grants among Woredas on basis of a grant formula that considered only Woredas' expenditure needs.Finding of the study shows that the grant formulas of both regions are less effective in addressing the fiscal capacity gaps between Woredas within the regions.The reason is that the grant formulas do not take into account Woredas' revenue-raising capacity while allocating the Woreda block grant pool among Woredas in regions.Finally, although Woreda governments located in urban areas endowed the right to borrow funds from domestic financial institutions, in practice they are not exercised their borrowing rights until now, mainly due to high direct control of regional executive councils.
Finally, the study found that the major problems and challenges that constraint the Oromia and Afar regional states to implement an effective fiscal decentralization at local level of government are: inadequate devaluation of decision-making power and autonomy; weak administrative and institutional capacity at local level, lack of availability of trained and experienced human resources and necessary infrastructure; Poor political willingness and commitment of authorities at regional level to decentralize fiscal power in real sense; lack of sufficient own revenues; poor design of regional-local transfer system; lack of participation of local community in identifying, implementing and monitoring of capital projects; and lack of local governments access to domestic financial markets.

Limitation of the Study
The study was constrained by several factors, which includes: lack of availability of sufficient, reliable and consistent secondary data, especially at local level of government; unwillingness of public officials and staffs to provide primary (first-hand) information; and lack of availability of globally recognize standard on the measurement of expenditure and revenue decentralization.In addition to these, the study is not attempted to analyze the impact of fiscal decentralization on various macro-economy variables such as economic growth, inflation, inequality and basic public service delivery (like Education, health-care).

Conclusion
Ethiopia has been implementing fiscal decentralization at local governments through devolution more power and fiscal responsibilities to Woredas (districts) since 2002, with the aim of enhancing the participation of local communities in political, economic and development affairs in order to improve the efficient of local public services delivery and local good governance as well as promoting local economic development.The practice, however, shows that there is lack of clarity with regarding to assignment of expenditure and taxing responsibility between regional and local governments, particularly on shared responsibilities.As a result, it creates conflicts between both levels of government and reduces efficiency of public services delivery and accountability at subnational levels.Besides, the devaluation of power and expenditure responsibilities to local governments by respective regional governments is not accompanied by sufficient taxation power.This, in turn, creates a high vertical fiscal imbalance and makes the local governments to heavily dependent on regional transfer to finance the costs of functions that they are responsible for.As the result, it further limits the fiscal autonomy of local governments to allocate their budgets as per the needs and preferences of their respective residents.This, in turn, limited them to realize the benefits fiscal decentralization such as improve the efficiency of public services delivery and accountability as well as promote economic development at grass root levels.

Recommendations
This study provides some practical recommendations that help improve the efficiency, effectiveness and equity of the fiscal decentralization system of Ethiopia.
(I) Regional governments needs to clearly specify the power and functional responsibility of regional and local governments, particularly on shared functions, so as to avoid the potential conflicts between the government units at different levels and improve further the efficiency of public services and ensure accountability at grass-root levels.Thus, the regional governments needs to take a comprehensive survey regarding the actual power and expenditure responsibilities of the local governments by unbundling different aspects of a certain public functions in terms of policy making, regulation, service delivery and financing.
(II) Regional governments need to make more political and leadership commitment in enhancing the managerial and technical capacity of Woredas as well as transferring necessary financial and material resources for Woredas institutions.This, in turn, would improve the capability of local authorities in planning and implementing projects and programs that could improve the living standards of their local people.In addition, the regional governments need to establish Capacity Building Offices at Woreda levels in collaboration with MCB that responsible for improving the managerial and technical capacities of local governments.
(III) Regional governments need to abolish the zonal administration from the administrative structure of the regional states to avoid the potential interference of regional sector bureaus in internal affairs of local governments through the zonal administration.The existing functions of zonal administration could be replaced through introducing other mechanisms such as establishing a special district or Woreda governments association to coordinate their functions horizontally.This mechanism might enable the local governments to control their activities without interference of zonal administration.
(IV) Regional governments need to revise the existing assignment of power and fiscal responsibilities for local governments by taking the specific conditions of each Woreda government into account such as population size and density, administrative and institutional capacity, so as to realize the benefits of fiscal decentralization such as improving the efficiency public service delivery and promoting economic growth at local levels.This because; certain local governments (especially in remote areas) are encountered several challenges to carry out the power and fiscal responsibilities which devolved by the constitution, mainly due to weak administrative and institutional capacity as well as lack of economic scale.
(V) Regional governments require to enhance the revenue generating capacity of local governments by taking various measures, which includes: delegating more tax bases with high yield rates; allowing them at least to set tax rates on own revenue sources; and improving the tax administrative capacity of Woredas.Local governments need also to introduce an appropriate tax enforcement mechanisms and create awareness of their tax-payers to pay taxes honestly, so as to improve the efficiency and effectiveness of their revenue collection This, in turn, would improves the autonomy local authorities in allocating their budgets by considering the needs and preferences of their local people.Thereby, it could enable to realize the benefits of fiscal decentralization at Woreda levels such as improve the efficiency of public services delivery and accountability of local authorities.
(VI) Regional governments need to take into account the revenue collection efforts of Woreda governments while distributing the Regional Block Grants horizontally among Woredas.
(VII) Regional governments need to establish a regional development bank in order access local governments to domestic credit sources.This, in turn, enables local authorities to expand the production and provision of public services that can improve the living standards of local people through investing in capital projects.Additionally, it enables local authorities to ensure intergenerational equity through financing some portion of the cost of long-term infrastructure projects by debts.
(i)  To what extent the regional governments in Ethiopia are devolves powers, functions and resources to local governments in practice?(ii)To what extent are the local governments in Ethiopia have financial and administrative capacity to carry out their own functions which assigned to them by law?(iii)What are the major challenges hindering the practice of an effective fiscal decentralization at local government levels in Ethiopia?

Table 1 . The expenditure ratios of woredas in Oromia and Afar Regional States during the 2007/2008-2015/2016 fiscal years (in Million Birr)
Note: WE = Woreda Expenditure, RE = Regional Expenditure and RWT = Ratio of Woreda Expenditure to Total Expenditure of Regional State.
Source: Data from BoFED; IBEX database of Oromia and Afar Regional States.

Table 2 . Composition of own revenue and fiscal capacity of Woredas in Oromia and Afar Regional States during 2007/2008-2012/2013 fiscal years
Total Revenue; BG = Blocked Grant; ST to TOR = Share of Tax to Total Own Revenue; SDT to Tax = Share of Direct Tax to Tax; SIT to Tax = Share of Indirect Tax to Tax; SNT to Tax = Share of Non-Tax to Tax; STOR to TR = Share of Total Own Revenue to Total Revenue; SG to TR = Share of Grant to Total Revenue Source: Own computations based on BoFED data of Oromia and Afar Regional States.