Report No. 1 8898-ET Ethiopia Regionalization Study June 15, 2000 Macroeconomics 2 Country Department 6: Ethiopia, Eritrea, Somalia, Sudan Africa Region Document of the World Bank GOVERNMENT FISCAL YEAR July 7 - July 6 CURRENCY EQUIVALENTS Currency Unit: Ethiopian Birr Official Rate: US$ 1 = Birr 7.07 (end of period 1997-1998) ABBREVIATIONS AND ACRONYMS ACSI Amhara Credit and Savings Institution ADLI Agricultural Development Led Industrialization BWUD Bureau of Works and Urban Development CBE Commercial Bank of Ethiopia CSA Central Statistical Authority DBE Development Bank of Ethiopia ESRDF Ethiopia Social Rehabilitation and Development Project, FDRE Federal Democratic Republic of Ethiopia GDP Gross Domestic Product GOK Government of Kenya GTZ German Technical Cooperation LAS Indian Administrative Service KLGRP Kenya Local Government Reform Program MEDAC Ministry of Economic Development and Cooperation MLG Mimstry of Local Govenmment MWUD Ministry of Works and Urban Development NEP National Environmental Policy NGOs Non-Govemmental Organizations NRM Natural Resources Management and Conservation NUPI National Urban Planning Institute PARUL Poverty Alleviation through Rural-Urban Linkages PHRD Population and Human Resource Development PRAs Participatory Rural Appraisals RTPC Rural Trade and Production Center SNNPS Southem Nations, Nationalities and People's State STDP Small Towns Development Program UDSS Urban Development Support Services UNDP United Nations Development Program UNCHS United Nations Center for Human Settlements USAID United States Agency for International Development WABs Women's Affairs Bureaus WAO Women's Affairs Office WDC Woreda Development Committee WID Women in Development Vice President: Callisto Madavo Director: Oey Astra Meesook Sector Manager: Peter Miovic Task Manager: Frederick Kilby PREFACE This draft report has been prepared as a collaborative exercise between the Bank and the government of Ethiopia. The Bank team comprised Messrs/Mmes Fred Kilby (AFTM2 and Task Team Leader), David Bevan (Consultant), Eyerusalem Fasika (World Bank RM), Sudarshan Gooptu (World Bank RM), Jessica Seddon (World Bank DECVP), Prof. Tapas Sen (Consultant), Paul Smoke (Consultant), Vinaya Swaroop (World Bank DEC/PREM), Yitbarek Tessema (World Bank RM). The main mission visited the regions of Amhara, Garnbella, Oromiya, SNNP, and the new Administrative Region 14, Addis Ababa in Ethiopia from June 27 to July 14, 1998. Other core members of the team not present in this mission were Mines. Gita Gopal (World Bank AFTH4), and Anna Maria Inverso (Consultant, AFTH4), whose contributions were based on earlier work undertaken in collaboration with the Women's Affairs Office, Federal Republic of Ethiopia, for the joint report "Implementing the Ethiopian National Policy for Women". The counterpart team of the Government was led by H. E. Ato Sufian Ahmed, Minister of Finance, H. E. Dr. Mulatu Teshome, MEDAC, and consisted of members representing sector ministries and regional bureaus. The mission would like to express its appreciation and gratitude to the many officials of the Federal Government of Ethiopia and the Regional Governments and municipalities of Addis Ababa, Amhara, Gambella, Oromiya and SNNP with whom we met. The report was prepared under the general guidance of Mme. Oey Astra Meesook (Country Director), and Peter Miovic (Technical Manager, AFTM2). Mr. Vinaya Swaroop was the Peer Advisor. Peer Reviewers were Marianne Fay (World Bank, .AFTU1) and Anwar Shah (World Bank, OEDCR). The preparation of this draft of the Regionalization Study would have not been possible without the able and efficient document processing from Lilian Canamaso, (AFTM2), Isabella Mwaura (AFMIKE) and research analysis support from Maria Gabriela Gonzalez (Consultant, AFTM2). ETHIOPIA Regionalization Study Contents SUMMARY AND POLICY RECOMMENDATIONS .................................................. i The Constitutional Structure and Scope of Regional Discretion ........................................ ii Expenditure Assignment .................................................. iv Tax Assignment .................................................. v Fiscal Imbalances and Regional Transfers ...................................... vii Capacity Constraints and Governance Issues ............................................,.,, x The Role of Municipalities Under Decentralized Development ..................................... xiii Regionalization and Gender Issues ........ . ................................... xiv :1. REGIONALIZATION AND THE DESIGN OF ECONOMIC CONSTITUTIONS IN ETH IO PIA 1.............................................. A. A MULTIGOVERNMENT STRUCTURE .............................................. 2 Federal Government .............................................. 2 Regional Government ............................................... 2 Zonal and Woreda Administrations .............................................. 3 B. EXPENDITURE ASSIGNMENT IssuEs ............................................... , ., 4 Devolution of Expenditure Responsibilities .............................................. 4 Subnational Budget Planning ................................................. . . 4 Expenditure Assignment: An Evaluation .................... .......................... 8 C. IssuEs iN TAx ASSIGNMENT ...................................................... ...... 14 Tax Assignment in Ethiopia ..16 Centralization of Revenues .. 17 Sharing of Joint Revenues ..17 Decentralization and Tax Reform ..18 2. FISCAL IMBALANCE, INTERGOVERNMENTAL TRANSFERS AND SUBNATIONAL BORROWING 23 A. FISCAL IMBALANCE: VERTICAL AND REGIONAL .24 Vertical Imbalances in Ethiopia ..24 Regional Fiscal Imbalances in Ethiopia ..26 B. INTERGOVERNMENTAL TRANSFERS ..27 Intergovernmental Transfers: A Conceptual Perspective . .27 Intergovernmental Transfers in Ethiopia ..28 Intergovernmental Grants Mechanism: An Evaluation of the Ethiopian System ... .30 C. SUBNATIONAL BORROWING. 33 Strong Performing Regional Governments and Domestic Borrowing . 35 3. ROLE OF MUNICIPALITIES UNDER DECENTRALIZED DEVELOPMENT IN ETHIOPIA.......................................................................................................................... 37 A. IMPORTANCE OF MUNICIPALITiES AND RECENT EFFORTS TO STRENGTHEN THEM ....... 37 The Urgent Need for Strengthening M unicipalities .............................................. 37 Positive Steps Towards M unicipal Reform in Ethiopia .............. .......................... 40 B. THE PRESENT MUNICIPAL SYSTEM AND PERFORMANCE CONSTRAINTS ...................... 41 Urbanization and M unicipal Status ..................................................................... 41 Legal Basis of M unicipalities ...................................................................... 42 Institutional Framework in Which Municipalities Function ............ ...................... 43 Municipal Functions ...................................................................... 44 M unicipal Revenue Sources ............................................................. ........ . 45 M unicipal Budgeting ....................... ............................................... 46 Governance ...................................................................... 47 Summary of Key Concerns Regarding Ethiopian Municipalities .......................... 48 C. MOVING FORWARD: A PROCESS FOR STRENGTHENING MUNICIPALITIES .................... 48 The Need for Deeper M unicipal Reform . ............................. .............................. 48 Some Examples of Municipal Policy and Procedural Reform in Developing Countries ...................................................................... 49 A Strategic Approach ......................... ............................................. 55 4. REGIONALIZATION AND GENDER ISSUES . ................................................................ 59 A. INTRODUCTION AND BACKGROUND ......................................................................... 59 B. REGIONALIZATION: POTENTIAL AND CONSTRAINTS FOR GENDER-SENSITIVE DEVELOPMENT ......................................................................... 62 Sustainable and Responsive Development ........................................................... 62 C. IMPROVED AND EFFICiENT DELIVERY OF SERVICES ................... ................................ 68 Agricultural and Related Services ...................................................................... 71 Equitable Development ......................................................................... 72 D. CONCLUSIONS ......................................................................... 79 Tables: Table 1.1: Federal and Regional Shares in Public Expenditure in Ethiopia: 1993-94 to 1997-98 Table 1.2: Share of Federal and Regional Governments in Revenues in Ethiopia: 1993-94 to 1997-98 Table 2.1: Vertical Imbalance in Selected Countries Table 2.2: Regional Imbalance in Ethiopia Table 2.3: Budget Shares and Budget Subsidies for Regional Governments in Ethiopia: 1997-98 Table 4.1: The Gender Dimension of Poverty in Ethiopia Table 4.2 Women's Representation in Decision-Making Bodies Table 4.3 Selected Socio-Economic Regional Indicators Boxes: Box 1. 1: The Evolution of Subregional Expenditure Allocation: An Example from the Amhara Region Box 1.2: The Indian Administrative Service Box 1.3: Incentives for Self-Improvement: An Example of a Civil Service Reform in a State in Brazil Box 1.4 Tax Assignment in Federal Countries Box 1.5 Split Tax Bases Box 2.1 Evolution of the Formula for Federal Grants (Subsidies) to Regions Box 2.2 Contingent Liabilities of the Regional Governments due to Agricultural Loan Schemes Box 2.3 Problems with Subnational Borrowing: Lessons from Experience Box 4.1 Key Elements of the Women's Policy Box 4.2 Women's Health and Education Status in Ethiopia Box 4.3 Gender and Employment Chart: Chart 4.1: Institutional Arrangements for Implementation of the National Policy on Women SUMMARY AND POLICY RECOMMENDATIONS 1. Ethiopia's development strategy is to pursue balanced regional progress, and is based on the principle of participatory development from the highest to the lowest levels of government. The Ethiopian Constitution is a federal system of ethnically based regions, where 48 of the country's 63 Ethnic groups have established national self - government structures. Minority ethnic rights are further protected through institutional arrangements ,designed to ensure representation for all, including decision-making bodies with special representatives for smaller ethnic groups. Although the Constitution was approved only relatively recently (in December, 1994) and the practical details of implementing this vision are still evolving, the pace of decentralization of the powers and functions of the state to lower levels of government is impressive. The present report provides an overview of developments and indicates some policy areas which may require attention as the system develops. 2. The first part of the report deals mainly with federal - regional fiscal relationships, and related regional planning and budgeting processes. Ethiopia has embarked on a bold and thoughtful process of decentralization, which has been supported by a widely shared consensus over both development strategy and objectives, and very large transfers of untied resources from the federal government to the regions. At this point the system is unquestionably working well. But as regions assume greater responsibility for their development, it is likely that differences may emerge between regional and national policies. Such differences are for the most part likely to be healthy reflections of specific regional circumstances, and indeed the very promise of decentralization is that policies and programs may be adapted to suit regional needs. The report argues that given the very substantial redistribution of funds that is taking place from the federal to the regional level-and the power of the regions to act in areas that may have national significance - the priority for policy design would be to look at the regional transfer system to give more incentive for regions to raise their own resources to contribute to their development, and to consider whether the present block grant system needs to be modified to create stronger economic incentives for regions to support national programs (rather than relying so heavily on political consensus and moral suasion). 3. Regionalization brings government closer to the people, and by bringing decision- making down to lower levels can foster decisions that are more in line with local needs. However, the impact of this strategy is by no means guaranteed, as illustrated in the second part of the report, which covers municipal development and gender issues. Municipal governments fall under the jurisdiction of the regions. But the legislation that governs them dates back to the 1940s, and as a result they are treated very unevenly by regional governments. Their accountability to local electorates is often murky, and they ii Summarv and Policv Recommendations lack the direct authority to set the tax rates and user fees that fund the services they provide. As a result, their ability to serve their citizens effectively, or contribute more broadly to rural-urban growth is impaired. The issues regarding women are more subtle. Here again decentralization holds the promise of greater empowerment. But generally women are under-represented in both the legislature and administration, and may lack an effective voice in their local communities, especially where social norms favor men. A wide variety of strategies are needed to strengthen women's influence and to promote gender sensitive development at the regional and community level. Otherwise, the National Women's Policy, which champions the role of women in Ethiopian society, may not fulfill its promise. The Constitutional Structure and Scope of Regional Discretion 4. Ethiopia has a four-tier system of government. Below the Federal Government, are Regional Governments. At the regional level, the country is divided into nine states and two "special city administrations" representing the two largest cities - Addis Ababa and Dire Dawa with a status equivalent to regions. These regional administrations are divided into zones (66), which in turn are divided into woredas (550) and special woredas (6). The woredas, which on average cover 100,000 people, are considered to be the key local units of government. As local capacity grows, they are increasingly administering and prioritizing the provision of public services. Below this level, communities are further subdivided into smaller electoral units, known as Kebeles, to further enhance community grassroots participation in the formulation and administration of policy decisions. In addition, most urban areas in the regions are governed by a parallel system of municipal government. Apart from Addis Ababa and Dire Dawa, the smaller municipalities derive their institutional structure, rights, and responsibilities from legislation of the 1940s. In contrast to the Woredas, the municipalities do not receive any fiscal transfers from higher levels of government and are, in general, relatively underfunded in relation to their responsibilities. 5. A series of proclamations made by the Federal Government, culminating in the 1994 Constitution, lays the basis for devolution of decision-making powers and responsibilities to the nine regional governments and two city administrations. The constitution assigns extensive powers to the regions. Each region can establish an administration that advances self-government and democratic rule in accordance with the Federal Constitution, enacts the state constitution and other laws, formulates and executes economic, social, and development policies, strategies and plans, administers land and other natural resources found in its territories, levies and collects certain taxes not reserved for the Federal Government, designs standards for state civil service conditions and pay, and maintains state level security forces. The functions assigned to the Federal Government, are national in scope - monetary policy, foreign relations, defense, etc. In addition to implementing policies with nation-wide benefits, central ministries are expected to support their regional counterparts by undertaking research, collecting data, and providing technical assistance and advice. SummaMr and Policy Recommendations iii 6. The principles of fiscal federalism suggest that assigning expenditure responsibilities and decision-making powers to lower levels of government can substantially improve a state's ability to efficiently identify and address its citizen's needs. The extent of regional discretion has three aspects. * the first concerns the allocation of responsibilities/discretion over expenditure. This has been laid out in some detail in a series of proclamations, which in keeping with the spirit of decentralization have given the regions the responsibility for the management of the bulk of local infrastructure and social service provision. * the second aspect concerns the allocation of revenue-raising powers, which has some unusual features, where the nature of ownership determines whether some taxes are paid to the federation or the regions. * and the third concerns the design of the transfer system (denoted by "subsidy" in Ethiopia) which bridges the gap (usually termed the "vertical imbalance") between the expenditure levels and revenue raising powers of each level of government. 7. The current arrangement in Ethiopia, is highly unusual. There is a large vertical imbalance between the revenue raising powers of the Federation and the expenditure responsibilities of the regions, which is bridged by unconditional block grants, dependent only on "objective" factors, such as population, development levels and local revenue generation, rather than the effective fulfillment of specific tasks. Because constitutional powers have been so strongly decentralized, upward accountability is relatively limited, reflecting only constitutional and legal requirements, such as properly accounting for the use of government funds. 8. The use of unrestricted block grants - which allow the regions to choose what they spend on -- encourages public spending which reflects local priorities. Local people have an interest and voice in service delivery, and to the extent that they participate in the trade- offs between different services, the outcome should reflect both the opportunity costs of these spending choices, and their value to the community. But regional choices are no longer sharpened by the perception that they are financing these choices through local taxes, which might lead to higher public spending in the aggregate. And taxpayers from other regions who provide these resources have relatively little say in the use of these funds. In terms of accountability and governance, this arrangement implies that the long terrn economic (and perhaps political) integrity of the system depends on the efficiency with which these resources transfers are used, and local corruption is controlled. Moreover the underlying support for such a large scale of transfers is also likely to be strongly influenced by the public's perception of their effectiveness in meeting widely held social goals, and the transparency of the results. From a practical point of view, it also raises thorny issues regarding the administration and funding of national development programs, which may run into conflict with the goal of local empowerment. iv Summarv and Policy Recommendations Expenditure Assignment 9. Excluding debt service, regional budgets accounted for 50 percent of consolidated government spending in the 1997/98 budget. At the sector level the regions account for the bulk of spending on economic development (agriculture and natural resource management), social development (health and education), and they account for more than one-third of spending on roads. Regional budget planning in Ethiopia is a complicated bottom-up process which is difficult to put in a nutshell since it involves multiple channels of coordination. Typically, information gathering in the yearly budget exercise begins at the woreda level and flows vertically within sectors as well as horizontally across as many as three levels of sub-regional government. The budgets are consolidated by the regional sector bureaus which create budgets in keeping with longer-term sector development plans. The Regional Planning and the Finance Bureaus use the infornation, as well as the budgets passed from the zonal planning and economic development departments, to create regional capital and current budgets. These budgets are then submitted to the Regional Council for approval, which in turn become law, and are issued in the Regional Official Gazettes. The Regional States thus emerge as the key players in Ethiopia's public service delivery system. It is the Regions that aggregate communities' preferences, create specific development plans within the more general national strategies, and, in most cases allocate the budgets between sectors, as well as between capital and recurrent expenditure. 10. At present, the form and extent of devolution within a region has been left very much for the regional governments to decide. In consequence, practice has been quite varied. While the details differ, the systems adopted among the regions are all very decentralized The extent to which woredas hold budgets, which they can allocate across sectors differs, and defies generalization. In some regions, there are fairly complicated arrangements for holding back funds within sectors, for example, budgets to fund inter- woreda roads or zonal hospitals may be held at the Zonal level. In other woredas, the budget envelopes are notional, and sector bureaus (which are accountable to either the zonal or regional authorities) actually retain control over their budgets. Some regions have adopted a form which very closely mimics that between the federal and regional governments, albeit in miniature - for example, the structure of the transfer formula from the SNNP Region to the zonal level has been copied exactly - while others have been somewhat more innovative. In general, there appears to be a move towards allocating the zonal and woreda budgets according to some criterion, and then allowing these (within reason) to be allocated between sectors according to preferences articulated within those units. Here the intraregional budgetary transfer arrangements mean that the woredas can align spending priorities with local preferences. While this is consistent with the bottom- up philosophy, which is central to the thrust of the Ethiopian decentralization, it could pose difficulties in managing nation-wide sector programs, for example, in education and health, to which the Ethiopian government is also committed. Summar and Policy Recommendations v Tax Assignment 11. In Ethiopia the federal government currently collects most taxes and revenues. In 1997-98, the federal government collected 82 per cent of the aggregate tax revenue and 83 per cent of the aggregate non-tax revenue. This is mainly due to the concentration of larger tax bases within the federal domain, including foreign trade taxes, corporation tax, and federal stamp duties. The federal government also has constitutional rights to some other smaller tax bases, including taxes on income from monopolies and winnings on national lotteries and other games of chance. An unusual feature of the Ethiopian system is that many taxes and fees are shared between the federal and the regional governments, where the split is determined on the basis of taxpayer categories rather than by types of tax. For example, in the case of public enterprises, ownership determines jurisdiction; income of employees, sales, profits, dividends and production of federal enterprises are taxed by the federal government, while those from regionally-owned public enterprises are taxed by the regional government. In the case of private companies, profits and dividends are taxed jointly by the federal and regional governments. But for non-incorporated companies, profit tax is levied by regional governments when an entity is a sole proprietorship or a farmers' co-operative, and by the federal government in the case of partnerships. Purely regional tax bases are small, and include fees from usufructuary rights to land, and royalties for the use of forest resources 12. The Constitution provides that any residual taxation power should be assigned by the House of Federation and House of Peoples' Representatives in a joint session. Since neither the federal nor the regional governments have residuary powers of taxation; assignment of any new tax is completely flexible. The Constitution is not exhaustive and does not mention some commonly used taxes. In particular, value added tax, is not covered. This is especially significant, because the introduction of value added tax is under consideration at present. Depending on its assignment, this could potentially lead to a significant shift in the distribution of tax receipts from the federal to the regional level. Other potential taxes, such as motor vehicle tax and entertainment tax, are not specifically assigned in the Constitution. These arrangements provide substantial scope for policy discretion regarding the future distribution of taxes between the federal and regional governments, without requiring changes to the Constitution. 13. Assignment of taxes to different levels of government involves several considerations which can produce conflicting recommendations. For example, conventional wisdom is that the incentive for efficient use of public resources is best served where the benefits of public expenditure are internalized among the taxpayers who benefit from these services. In the Ethiopian system, where the bulk of expenditure responsibilities are assigned to the regions, this would argue for a corresponding revenue decentralization. But there are also strong economic arguments for centralizing tax policy and administration. If regions are assigned the powers to levy taxes which are highly progressive, or redistributive, in nature this can create artificial and economically harmful incentives for-and barriers to-production and trade. For this reason, it is commonly argued that taxes should be levied on producers and products whose mobility is restricted to the geographic jurisdiction of the taxing authority. For example, economic activity may vi Summary and Policy Recommendations shift within a country simply to minimize tax burdens, and interregional trade may be discouraged if regions attempt to tax products or services produced by other regions. Preserving and strengthening nationally integrated markets clearly benefits everyone. Economies of scale in tax administration are an important consideration arguing for centralization in countries where administrative capacity is short. Finally, the desire of the federal authorities to preside over a strongly redistributive fiscal system also argues for some centralization of revenues. 14. Centralization of Revenues. The present distribution of tax revenue is heavily tilted towards the federal government in Ethiopia, while expenditure responsibilities have been so extensively decentralized that there are clearly arguments to be made for encouraging the development of stronger regional tax bases. From the point of view of regional tax policy, if there is scope for expanding local taxes that do not raise economic or administrative efficiency concerns this could be a win-win situation. There are some developments which are likely to shift the distribution of taxes from the federal to the regional level. The privatization of federal public enterprises, will shift the personal income tax obligations of workers from the federal level to a joint tax base shared by federal government and the regions. Rising agricultural incomes (which are taxable by the regions), as well as improvements in the administrative capacity of regional governments could further boost regional tax yields. But given the dominant role of the federal government over the major tax bases these developments are not likely to lead to a major decentralization of tax receipts. 15. The most promising scope for expanding the tax base of the regional governments is likely to be from greater revenue sharing, or the introduction of VAT. Joint revenues are not significant at present, and their distribution (after being collected by the federal apparatus) is not yet a settled procedure. Currently, joint direct taxes are shared on a 50:50 basis, while for indirect taxes the ratio is 30:70 in favor of the federal government. The process of distribution among regions is under discussion, and so far not every regional government has received a share of this revenue. Shifting the ratio in favor of the regions would be a relatively simple step. The reliance of regional governments on federal transfers could also be reduced, and their own taxing powers enhanced, by creating suitably chosen new tax instruments. For example, a regionally administered comprehensive value-added tax (VAT) with necessary safeguards for harmonization purposes could be studied. Such a move needs to be considered carefully, since VAT is an administratively complex tax. In view of the limited administrative capacity in the regions, regionally administered VAT may not be the best use of scarce skills. But even if VAT is centrally administered, the revenue base could be assigned to the regions as a means to strengthen their fiscal autonomy. 16. Subnational Borrowing. In Ethiopia, regional governments are technically allowed to borrow domestically with the concurrence of the federal government, but in practice very few regions have exercised this power, except for very short-term cash management purposes. Since regional expenditures levels are in effect determined under the transfer formula (see para. 21), and any planned deficits are financed through grants (subsidies), in the current system, the incentives for regional borrowing are non existent. Summary and Polica Recommendations vii Municipalities within regions have some scope for borrowing, but at present their weak financial position and limited revenue base prevents them from doing so. Consequently, the decentralization of spending authority has not led to a run up in sub-national borrowing and the associated macroeconomic instability or crowding out of private sector that has arisen in some other decentralized systems. Several regions have, however, used their borrowing privilege in an indirect way to guarantee - with the concurrence of the Federal Ministry of Finance - commercial bank lending to peasant farmers under the government's agricultural modernization program. This program, which is proving popular with farmers, has the potential to expand rapidly and could therefore lead to a rapid build up in the contingent liabilities of regional governments. Because Ethiopia is prone to drought this situation poses considerable risk for regional governments and local communities. The mission was told of one woreda where school building programs had been cut in order to repay the loans of farmers who had defaulted on their loans. Fiscal Imbalances and Regional Transfers 17. Extent of vertical imbalances in Ethiopia. Vertical imbalance refers to the mismatch between expenditure responsibilities and assigned revenues at the sub-national level. A vertically balanced system would imply that each level of government financed its own respective expenditures, and would therefore be in accordance with the goal of trying to improve decisionmaking by marrying tax burdens with expenditure benefits. In practice, assignment of regional expenditure and tax responsibilities is typically considered independently, as different principles govern the two sides of the budget. In Ethiopia, as in many developing countries, the reliance of the government on trade taxes and a desire for a relatively progressive and redistributive tax structure have pushed the balance of taxation toward the center. But on the expenditure side, the concern to accommodate Ethiopia's diversity and to bring government closer to the people have favored radical decentralization. Great efforts have been made to achieve, to the extent practicable, a bottom-up structure that reflects local information and preferences while being coherent with regional and national policies. The outcome of these two opposing goals is a degree of vertical imbalance which is striking by international standards. In the sample of countries we looked at, Ethiopia's vertical imbalance coefficient (0.52) was considerably higher than the federal systems of Mexico(O.36) or Malaysia (0.37), and was exceeded only by Bolivia(0.66) and South Africa (0.88). However, the share of regional government spending in total government spending is far higher in Ethiopia (41.2 percent) than in Bolivia (19.8 percent) or South Africa (30.9 percent). 18. The conventional wisdom is that such a high degree of vertical imbalance may distort policy decisions regarding the level, timing and distribution of public spending. Because the costs of public expenditure are not fully internalized and spread over the entire country, there is a reduced incentive to provide public services at the least cost. However, since the public expenditure envelope in Ethiopia is predetermined, the usual problem of burgeoning public expenditures may not be so serious. Still, in theory, the tax price of public services as a whole would be lower than otherwise relative to privately supplied services, which would distort the private-public mix of the bundle of services that viii Summarv and Policy Recommendations the citizens would demand. This would particularly apply to those services which go to determine the level of the regional expenditure envelope and therefore grants (subsidies). For example, the system provides a theoretical incentive to wait for a publicly financed school even when it is possible to start a private school because (a) the additional tax cost of a new government school would be much less to the community than that of a private school and (b) any new school would reduce the subsidy to the region to some extent. 19. It could also be argued, that there is always some loss of sub-national autonomy when expenditures are dependent upon transferred resources, even when there are no formal conditionalities attached to the transfer; this is especially so when the rules of the game are framed by the grantor (the federal government). Too large a deviation from the grantor's preferences might cause hesitation in adopting such policies on the part of the grantee (regional governments) even when warranted, because of apprehensions regarding loss of grants that may follow. The greater the share of grants in their total expenditures, the greater would be the degree of such hesitation. And in the event that differences emerge between the federal government and regions, there are likely to be pressures towards introducing conditional grants to encourage spending on centrally favored programs. 20. Intergovernmental Transfers in Practice. In the last few years, regional governments were able to finance, on average, 30 percent of their expenditures from their own revenue sources; the remaining 70 percent was financed by the federal government in the form of a transfer. Clearly the transfer formula is one of the most important policy variables at work in the Ethiopian public finance system. The present Ethiopian system of intergovernmental transfers is based on block grants and the regions are free to spend according to their own choices and preferences. The block grant system has considerable merit, because it supports local decisionmaking based on community preferences. 21. But the block grant is determined quite differently from most other federations, which usually allocate a fixed grant either through one or more chosen formula(s) or some other method including an unilateral judgment regarding the quantum of grants that the grantor government can afford or the grantee government may need. The method used in Ethiopia involves determination of the total expenditure envelope first (the federal government decides the total expenditures that the federal and regional governments taken together can be allowed to incur), and then giving gap-filling grants to cover expenditures not financed by (estimates of) the grantee's own revenues or donor flows. This has the considerable practical merit of enforcing macroeconomic discipline by setting a cap on overall public expenditure. On the other hand, since total expenditures are determined first and grants are subject to both regional tax netting and budget offsets, the system does not provide regional governments with much scope to decide the size of their budgets. This could lead to potential disagreement over the appropriate scale of regional spending. 22. In addition to the restrictions that such a system places on regional autonomy, there are three additional concerns which policy makers will need to consider as the system evolves. Summaa and Policv Recommendations ix First, the budget formula actually penalizes regional tax efforts in determining the transfer to individual regions. Although regional pride, and a widely shared desire to contribute to national development, seem to currently outweigh what would otherwise be an incentive for regions to avoid taxing themselves, this may lead to free rider problems in the future, and may result in the system being perceived as being unfair to regions that make a stronger tax effort. * The second concem is that the "budget offset" system whereby budget subsidies to the regions are reduced by an amount equivalent to the projected disbursement of direct donor funding reduces the incentive for regions to tap donor resources. The clear advantage of this system is that it enables the federal government to retain control over the both the scale and geographical scope of public spending. Some regions might otherwise become favorites for donor support and hence undermine the system of budget transfers. If donor resources are fully fungible between regional projects and federal programs there may not be much to worry about, but if not, the absence of any direct financial incentive for regions to work with donors could lead to a diminution of overall flows. * Thirdly, the authorities may wish to look at the development indicators used for assessing regional need. At present information on regional incomes and poverty outcomes are somewhat limited. The authorities have therefore tumed to some simple and useful indicators of regional development levels, based on economic and social infrastructure levels. But as Ethiopia begins to expand and refine its socioeconomic survey and analysis to capture the regional dimensions of poverty more accurately this information could be factored into the regional transfer mechanism more directly. Greater public awareness of the extent of regional poverty, and information on the performance of regional governments in improving the lives of the poor, could also be helpful in maintaining regional accountability for the very substantial amount of federal funds which flow to the regions, as well as national popular support for the system of regional transfers. 23.. In countries like Ethiopia which are starting afresh with a low level of development, priorities for govermment spending-roads, health, education, water supply and agricultural development are not so difficult to determine. There seems to be a strong consensus at the regional and federal level on these interventions, which in tum underpins the widespread support for the block grant system. As long as regional preferences more or less correspond to national ones, the case for other types of grants is quite weak. Yet the exclusive reliance on block grants' may also merit a fresh assessment to examine the 'The subventions from the Road Fund is an obvious exception, since the funds so transferred can be spent in the road sector only. x Summary and Policy Recommendations scope for introducing incentive based conditional grants in the future. As the country moves along the path of economic and social development, this uniformity of preferences may weaken over time, and in that situation, other types of grants could be legitimate tools for the federal government to influence the spending pattern of regional governments. Yet paradoxically it may be better to undertake such a review now while there is consensus over the key programs than to wait for differences to emerge. There is a wide literature available on the experience of federal countries - developed, developing and transition - in instituting conditional grants, that might provide potential models for Ethiopia to adapt to its own special circumstances. 24. This may be an appropriate opportunity to consider a more formal institutional arrangement for determining the level, form and distribution of grants. At present, MEDAC and the Federal Ministry of Finance formulate and execute the entire system of federal grants to regions. The present system of fiscal federalism in Ethiopia can be characterized as "cooperative" federalism - few disputes and widespread cooperation between the federal government and regional governments, as well as between regions. International experience shows that over time cooperative federalism often changes into "competitive" federalism with an increasing polarization over the allocation of funds. To avoid such an eventuality, the present system may need to become more institutionalized to protect the current consensus. One possible way to lirnit the risks of competitive federalism is to set up an independent Grants Commission to ensure that the inter- governmental transfer mechanism is seen fair by most, if not all, stakeholders, and is not subject to a frequent tinkering exercise involving minor changes here and there. Among the industrialized countries, Australia has such a system, where a permanent independent agency (Commonwealth Grants Commission), consisting of professionals, determines the intergovernmental grants on the basis of formula(s) worked out through continuous research in this area. But there are many alternative models to choose from, each of which have their own different strengths and weaknesses. Ethiopia may wish to consider this model or an alternative institutional arrangement for lending transparency and predictability to the determination of budget transfers, as well as to nurture the consensus on regional transfer principles Capacity Constraints and Governance Issues 25. The regionalization drive is an important vehicle for improving the institutional environment for efficient and equitable public expenditure. However, an overall shortage of administrative capacity and marked regional variations in the availability of skilled administrators and technical staff are a major constraint. It is sometimes argued that the regionalization program is unwise given these capacity limitations. In one sense this is mistaken, since governing a nation as large, dispersed and varied as Ethiopia is not made intrinsically more difficult by having a decentralized as opposed to a centralized structure. However, decentralization inevitably involves some duplication of function, and thereby risks drawing scarce administrative talent into less productive functions. Guarding against this risk requires local governments to make tough choices about what services they wish to provide and the activities they regulate. In some countries decentralization has been Summary and Policy Recommendations xi associated with a major increase in civil service employment, particularly where local governments attempted to expand their mandates into areas that might have been better left to the private sector. To date this does not seem to have occurred in Ethiopia. In view of the budgetary implications, and the desirability of raising civil service pay to r ecruit and r etain high quality staff, it is very important that this position be maintained. 26. The widespread shortage of capacity in the civil service is universally acknowledged, and rectifying this has become a major objective of government. The ongoing civil service reform program is extremely ambitious and therefore will take many years to bring to completion. In the context of service delivery in the regions, what is most urgent is probably to organize a widespread improvement in low-level skills. From this perspective, the focus of the civil service college on providing long duration, high level courses is not especially helpful. A related point is that the current informal system of federal support for administrative capacity building in the four least developed regions .-Afar, Somali, Gambella, and BeneshangulVGamuz - seems quite modest in relation to the level of financial support that is being channeled to these regions. There may be a case for a more aggressive federally supported program to build administrative skills. There may also be opportunities to take fuller advantage of the better-off regions' experience. The current arrangement whereby the Road Authority of Tigray is responsible for implementing some road construction projects bordering Tigray and Amhara, which economizes on technical expertise, provides an interesting example of the potential for interregional collaboration. 27. The capacity issue is not restricted to the types of technical skills which can in principle be taught, important though these are. Capacity also lies in the capability to, and habit of, applying experience and common sense to resolve practical problems. First it is necessary to perceive or register the problem, second to articulate solutions. During the Derg era the civil service became increasingly unresponsive to citizens' needs. Accordingly, it lost the habit of even recognizing peoples' problems, let alone formulating an adequate response. Unless it is possible to draft into the civil service a whole new cadre who are innately "client oriented" and who can act as seeds to re-energize their colleagues, it will be necessary to tackle this problem by instituting some source of outside pressure. Regionalization has created new channels for citizen input into policy-making and increased leaders' accountability to beneficiaries and incentives to perform well. Nevertheless, there is a risk that the lack of capacity within regional government may lead to poor quality investment decisions. This suggests that it would be highly desirable for the federal government to develop a set of evaluation criteria for general use across the regions, so that cost/ benefit and beneficiary assessment techniques can better inform local choices, for example in donor-supported Sector Investment Programs (SIPs). Given the current regionalization philosophy, the federal government would presumably not choose to make these mandatory, but if they can be shown to bring practical benefits, they would very likely be adopted voluntarily in current circumstances. 28. Governance and corruption. The Federal Government clearly wishes to devolve very substantial discretionary powers to the regional governments. It is also concerned that these changes should lead to improvements in governance, and not be associated with xii Summarv and Policy Recommendations increased corruption. Moving from the old situation where the lower echelons were used to operating top-down regulation, but had little control over resources, to one where they have considerable discretion, and now have control over substantial resources, will be very risky for the evolution of good governance. However, Ethiopia has not developed a culture of acquiescence in corruption, and when cases come to light they appear to be followed by dismissal and in some cases prosecution. Ethiopia has already developed its own mechanisms to encourage discipline and transparency, notably the "gum gum" evaluation system which subjects officials to regular and public review. This is claimed to encourage openness, improve motivation and performance, and reduce corruption. However, the political roots of the system may mean that it is not a long term answer to the problem, which would perhaps involve a wider and more pluralistic engagement of ordinary citizens and a variety of institutions of civil society. 29. To prevent a rise in corruption, action on a variety of fronts is required. The first is to attempt to institutionalize what appears to be a high commitment to a public service ethic amongst many of the officers of the regional governments, both civil servants and politicians. It will also be important, second, to raise civil service pay to appropriate levels to provide adequate incentives, and to provide clear statements of their duties, rights and obligations as well as the nature of disciplinary procedures. The third requirement is for transparency and accountability to the general public. 30. Governments have two quite distinct types of role in econornic activity. One is as an economic actor, obtaining resources from taxation, user charges and borrowing, and using them to purchase goods and services and make transfers. The other is as the instigator and organizer of the "rules of the game" for other economic agents (and itself). Much of the discussion of the merits and demerits of decentralization has been restricted to its impact on the first role. Thus, as noted above, one of the classic arguments for decentralization is that it permits the supply of local public goods to be better suited to local preferences. In one sense this possibility -or even promise - of diversity is at the heart of the whole regionalization program. 31. But it is important to take stock of just how far the authorities wish to take this exercise in pluralism. It is for consideration, therefore, whether it might not be desirable for the federal government to impose a relatively high degree of uniformity across regions in respect of these legal and regulatory features, even though it continued to cede to them considerable autonomy in respect of expenditures, and possibly increased autonomy in respect of revenue-raising powers. While the regions may adopt varied "rules of the game" within their territories, there is a high premium on these different rules being operated in a transparent, reliable and uniform fashion. Second, many countries are coming together to secure the benefits of a "level playing field" through the international harmonization of many legal instruments, tax policies and the like. If the federal government would wish to limit regional differentiation in this area, then it would be better to act now, before it has become any sort of problem, and while the existing consensus holds. If, instead, it is attempted to introduce restrictions when the issue has become live, it will be at least more contentious and possibly infeasible. 'Summary and Policy Recommendations xiii The Role of Municipalities under Decentralized Development 32. Ethiopia is among the least urbanized countries in Africa. According to the 1994 (Census, there are 855 urban centers in the country, accounting for only 13.7 percent (8.2 rnillion of 54.9 million) of Ethiopia's population. Ethiopia's 1994 Constitution did not e,xplicitly deal with municipalities, a form of local government that was established in rnajor urban areas and small towns in the 1940s. Municipalities in Ethiopia suffer from a number of serious problems. First and foremost, they operate without clear legal status in the context of the broader decentralized public sector structure being developed by the Government. Formally, regional councils have administrative responsibility for all government entities within their jurisdiction, including the municipalities. In practice, municipalities may be required to report to lower levels within the regional administration, i.e., the zone or the woreda, rather than to the regional council. This results in somewhat arbitrary and inconsistent treatment by higher levels of government, which have to date had few clear standards to govern their dealings with municipalities. '33. Municipalities in Ethiopia have principal responsibility for some important public services that could support private sector development and improve the welfare of urban i-esidents, including the provision of roads and bridges, markets and transport terminals, public hygiene, water supply sewerage and solid waste management and fire and ambulance services and street lighting. They do not receive any direct budget support fiom regional (or federal) government and rely on local property taxes, business licenses and user fees to finance their activities. These revenue sources have generally been poorly exploited, often due to regionally imposed restrictions on the base and rates charged on local revenues. Due to inadequate revenues, municipal service delivery is frequently weak, and of poor quality. This creates a vicious cycle, as poor services undermine the incentive for residents to pay municipal taxes and charges. In addition to their weak financial position, municipalities frequently lack the skills to provide basic urban services. Local organizational structures, job descriptions, and managerial processes and procedures are poorly defined and inadequate, and the degree of accountability to local constituents is generally weak. And some urban centers are not even designated as municipalities, depriving them of both municipal functions and sources of revenue. 34. The focus of the Ethiopian government on ADLI (Agricultural Development Led Industrialization) has to some extent led to neglect of municipal governments. However, ,evidence from other countries suggests that rural development strategies can suffer if there is inadequate urban development. Even small urban centers often function as communications and transport centers for their region, providing critical linkages to rural areas and serving as concentration points for marketing activities. Nurturing rural-urban linkages is thus a key component of a successful local economic development strategy. Municipalit:ies are often uniquely well placed to understand this and to design supportive interventions. Effective management and development of smaller municipalities can also have an impact on national population policies and urbanization patterns by helping to curtail the excessive growth of population and economic pressure in primary urban centers. Moreover, better municipal government has the potential to stimulate private xiv Summarv and Policy Recommendations sector development more generally, the creation of jobs for a rapidly growing urban population, and to enhance the well being of urban residents, especially the poor. 35. There is clearly an increasing Government awareness of the need to pay more attention to municipalities, and a number of positive steps towards reform have already been taken. First, the Federal Ministry of Works and Urban Development (MWUD) recently prepared draft legislation to define criteria for municipal status. Second, the MWUD has been supporting the development of projects through its Urban Development Support Services (UDSS) group to enhance the financial and managerial capacity of municipalities through system development and on-the-job training. However, these efforts have been sporadic, limited in geographic focus, and not very closely coordinated. In order to raise the performance and to increase the development impact of municipal development a more comprehensive and integrated set of reforms will need to be pursued. Key suggested reforms to consider would include: * Developing and applying objective criteria for the creation and classification of municipalities as a basis for more clearly defining their detailed roles and responsibilities and relations with higher levels of regional administration. * Creating guidelines for the organizational structure, job responsibilities, revenue bases, and managerial process and procedures to be used by various types of municipalities. * Developing mechanisms to revise municipal tariffs and salary scales on a regular basis. D Enhancing accountability to municipal residents through reforms to standardize the electoral process and opening channels for greater citizen information on and participation in municipal decision making. Rcgionalization and Gender Issues 36. The National Policy for Women, 1993 (Women's Policy), formulated concurrently with the regionalization initiatives, acknowledges that neglecting the concerns of women has resulted in social and economic costs, and underscores the need to move towards a more institutionalized approach through enhanced policy measures. The Policy articulates a comprehensive strategy to institutionalize the socioeconomic and political rights of women and promote gender-sensitive policies and interventions. The potential and constraints for gender-sensitive development may be divided into two main components: sustainable and responsive development, and equitable development. 37. Sustainable and Responsive Development. The recent changes in institutional arrangements have succeeded in bringing gender issues to the table at the regional level. Each region has established Women's Affairs Bureaus WABs, and most regions have appointed coordinators for women's activities in two to three of the important sector bureaus. In some regions, women's focal points have also been established in the lower government echelons, i.e., zones and woredas. Many of these offices are dynamic, and are striving for change. But for a variety of reasons, the new arrangements are not Summary and Policy Recommendations xv fi'nctioning effectively. First, WABs are not linked in any formal maimer to Planning and Finance, except in Amhara and the SNNP regions. Moreover, most WABs function in isolation, unrelated to those involved in mainstream development planning, since there is no formal relationship between women's departments in sectoral bureaus and WABs. There is also a distinct lack of women in leadership or managerial roles. In Oromiya, for example, there are no women representatives in the regional council or executive committee. The Constitution permits the use of affirmative action as a tool to address gender inequality. Yet, in almost all the regions heads of sector bureaus are - except for WNABs - all male, even though these are political appointments, which provides considerable scope for appointing women. 38. With the exception of the SNNP, the regional budget arrangements do not provide the WABs with resources to independently fund projects or programs specifically targeted toward women. Providing WAB's with the resources to allocate grants to sector bureaus to cover the costs associated with mainstrearning gender sensitive activities into their programs could help tilt the balance. Modifying the current block grant system to create more direct incentives to ensure that regions allocate adequate budget for the implementation of the National Women's Policy would provide an even stronger push. :39. Regionalization has the potential of enhancing equity in the delivery of services resulting from wider local involvement in decisionmaking. However, the high illiteracy rate, low levels of educational attainment, lack of skills, the absence of leisure, given their household and other responsibilities - combined with lack of information on developmental opportunities - mean that in practice women's participation in the planning and implementation of development interventions is currently limited and weak. Thus it becomes critically important to have strong women's groups or representatives to speak on their behalf. One way to strengthen women's influence may be to decentralize responsibilities for the design of service delivery down to the facility level, so that local women may have a greater say in the way these services operate. This is something already envisaged in the National Health Policy. Another way may be to build gender specific interventions into the design of programs. For example, in Ethiopia, there are significant inter-regional variations in the degree and extent of women's participation in the agriculture sector and related agro-processing activities. Regionalization permits extension services to tailor their assistance to reflect differences in agricultural systems. 'Yet despite their contribution, women are regarded as marginal workers and do not receive equitable access to critical inputs of production such as land, fertilizers, credit facilities, and are not targeted as potential recipients of extension and other supportive services. Correcting these design shortcomings may require interventions at the highest policymaking levels. 40. Equitable Development. Regionalization permits increased attention to governance issues, a necessary pre-condition for equitable development. A predictable legal framework is essential, and should provide for equitable allocation of resources between men and women and a satisfactory dispute settlement method. The Civil Code (1960) has invalidated all customary laws and practices - which differ across regions - that govern allocation of resources within a household. Women are often unaware of the progressive provisions of the Civil Code, and Customary laws which are inequitable in the xvi Summarv and Policy Recommendations allocation of resources, and whose control is mostly with men, are applied by the majority of rural people. Moreover, the Constitution further clouds the issue since it contains a clause which permits the applicability of customary laws to personal disputes, should both parties to a dispute agree. Amendment to the legal framework needs to focused on these regional differences, so that women's rights are more uniformly protected across the country. For disputes concerning marriage and betrothal, the civil code rests powers with family arbitrators, which has the important advantage of providing women with much greater access to legal remedies than if their only recourse were to the formal judicial system. In practice there are many shortcomings to the system, and it is presently under review. The responsibility for change lies with the regions, who could potentially develop a system with much stronger safeguards for ensuring that the applicable laws are applied in a more gender sensitive fashion. Regional governments, also have the authority to ensure gender equity in distribution of rural land and are actually obliged to do so, under the Federal law and the Constitution. 41. It is not possible to do full justice to the range of issues covered by this report in a short summary, but the interested reader may wish to consult the specific sections of the report for more details. Chapter 1 provides a description of the assignment of tax and expenditure responsibilities in Ethiopia. Chapter 2 looks at the fiscal imbalances that have emerged betwveen the regions and the federation as a result of this assignment, and the design of the regional budget subsidy arrangements. Chapter 3 describes the institutional arrangements governing municipalities and the potential for improving urban service delivery and fostering rural growth, based on the experience of other countries. Finally, in Chapter 4, the report takes on the issue of regionalization and gender. 1. REGIONALIZATiON AND THE DESIGN GE ECONOMIC CONSTITUTIONS IN ETHIOPIA Regionalization in Ethiopia is intricately linked to the federal government's efforts to pursue equitable development for all regions of the country. The righting of past wrongs ('urban bias, ethnic intolerance, and neglect of infrastructure and social development) motivates many of the federal government's current efforts to increase citizen participation, transfer resources to peripheral areas, and promote balanced social and economic growth. A key component of the strategy is to reallocate resources and devolve decision-making to the lowest possible level of government. 1.1 After the fall of the Derg regime in 1991, socio-political factors led to a federal system of governance in Ethiopia, which is built along ethnically based regions. Adopted in December 1994, the Ethiopian Constitution celebrates its ethnic diversity in outlining the country's governance structure. The Preamble of the Constitution starts with "We, the Nations, Nationalities and Peoples of Ethiopia..." The antecedent to the Constitution -- a 1991 Proclamation that establishes national regional self-government -recognizes 63 different ethnic groups, 48 of which are large enough to be able to establish their own self-government structures at the woreda level, the fourth tier of the government, or above. ' 1.2 Ethiopia's development strategy is to secure balanced regional progress and is based on the principle of participatory development with involvement from the highest (federal) to the lowest (woreda) levels of government. There is strong belief in the principle that a decision making process that includes all nationalities and ethnic groups, is key to sustained national development. Institutional structures are designed to ensure representation for all ethnic groups, and in some cases decision-making bodies such as the regional councils (see below) include special representatives for minorities.2 The pace of decentralization of powers and functions to the lower levels of government is impressive and the process is still unfolding. 1.3 This chapter first describes the structure of multilevel govemment in Ethiopia. The following two sections describe and evaluate the design of the prevailing expenditure and tax assignrments. A series of suggestions for improvement are also provided in these sections. 1 Article 3 of Proclamation 7 (1991) of the Federal Govemnment of Ethiopia. 2 Article 54 (2) says that "Provisions shall be made by law for special representation for minority Nationalities and Peoples" in the House of Peoples' Representatives. 2 Chapter I A. A MULTIGOVERNMENT STRUCTURE 1.4 The ongoing process of regionalization has created a four-tier system of government in Ethiopia: the Federal Government, Regional Governments, Zonal and Woreda Administrations. At the regional level, the country is divided into nine states and two special city administrations.3 Together these are divided into zones (66), which in turn are divided into woredas (550) and special woredas (6). Woredas, the key units of local government, are further subdivided into kebeles, primarily to facilitate implementation and to facilitate two-way flow of information. In addition, urban areas in the regions are governed by a system of municipal government. (For more information on municipal government in Ethiopia, see Chapter 3). Federal Government 1.5 Each of the four levels of government has more or less the same structure: a legislative body, a court system, and a number of sector-specific administrative institutions that constitute the executive branch. At the federal level, the legislative arm consists of two bodies, the House of Peoples' Representatives and the House of Federation. Members of the House of Peoples are elected by the country's population, while the House of Federation is made up of at least one representative from each region. Both houses make general national laws, but the House of Federation focuses more on issues in intergovernmental relations. The Federal Executive Arm consists of a President and a Prime Minister who heads the Council of Ministers. The Council carries out the functions assigned to the Federal Government - monetary policy, foreign relations, defense, etc.- as well as designs national policies which in turn, provide guidance to the lower levels of government. In addition to implementing policies with nation-wide benefits, central ninistries are expected to support their regional counterparts by undertaking research, collecting data, and providing technical assistance and advice. The Federal Supreme Court has the highest judicial power over federal matters and is the court of final appeal for regional cases. Regional Government 1.6 A series of proclamations made by the Federal Government, culminating in the 1994 Constitution, lays the basis for devolution of decision-making powers and responsibilities to the nine regional governments and two city administrations. The constitution assigns extensive powers to the regions. Each region can establish an administration that advances self-government and democratic rule in accordance with the Federal Constitution, enacts the state constitution and other laws, formulates and executes economic, social, and development policies, strategies and plans, administers land and other natural resources found in the territories, levies and collects certain taxes not The nine Regional States are Afar, Amhara, Beneshangul/Gumuz, Gambella, Harari, Oromiya, Somali, Southern Nations Nationalities and Peoples (SNNP), and Tigray. The two special city administrations are Addis Ababa and Dire Dawa, Ethiopia's capital city and second largest city respectively. Regionalization and the Design of Economic Constitutions in Ethiopia 3 reserved for the Federal Government, designs standards for state civil service conditions and pay, and maintains state level security forces. 1.7 The Regional Council, elected by the citizens of the region, makes region-specific legislation and approves the budget and development strategy. Members of the Regional Council appoint an Executive Committee - a subset of the Council itself - which oversees day-to-day administration of the region. The Council also appoints a President to perform the executive functions, and a set of Bureau Heads who act as the regional equivalent of ministers. According to Proclamation 41 (1993) of the Federal Government, the number of and exact division of responsibilities between Bureaus can vary between regions. All regions have Finance and Planning Bureaus, and Sector Bureaus in priority policy areas: Agriculture, Education, Health, and Roads. Functioning as regional mninistries, these bureaus prepare short- and long-term action plans in keeping with central government targets. Regions also have their own civil services, though pay scales and conditions of services are aligned with the Federal Civil Service. Regional supreme courts oversee the interpretation of regional constitutions and the application of regional laws. Zonal and Woreda Administrations L.8 The structure of subregional government varies, as each region is given the right to adjust its institutions to suit its particular development needs. Zones, the level of government just below the region, are mentioned in the Constitution as structures to be tised for "administrative convenience." The Zonal administration consists of a set of counterparts to the Regional Bureaus. Their main task is to aggregate the information coming from the smaller communities, which are represented by Woredas. In most regions, zonal administration is appointed by the regional government. (In the SNNP region, however, the zonal administration is an elected body.) They usually serve as conduits for implementing the budgets that are planned at the regional level. 1I.9 Constitutionally, woredas, the lowest level of elected, full-time government (covering roughly 100,000 people) are the most important unit of subnational government. They consist of an elected council, a president chosen from among the council, and a set of sectoral offices. In most cases, these sectoral offices serve as implementing agencies and contact points with the communities rather than development policy planners. Within wvoredas, communities organize themselves into kebeles, peasant associations, and other community groups. Kebeles usually have an elected leader and a small council, but these officers are not part of the paid, full-time government structure and usually have shorter terms than higher-level officials. 1].10 This standard four-tier institutional structure applies mainly to the rural and semi- rural areas of Ethiopia. The two larger cities, Addis Ababa and Dire Dawa, have "special administrations" with status equivalent to regions. Smaller urban areas are typically outside the governance structure envisioned in the new constitution. These ',municipalities" derive their institutional structure, rights, and responsibilities from legislation of the 1940s. In relation to their responsibilities they are typically 4 _ Chapter 1 overburdened and underfunded relative to the Woredas (for more information on municipalities, see Chapter 3). B. EXPENDITURE ASSIGNMENT ISSUES .11 One of the underlying arguments for fiscal federalism is that assigning expenditure responsibilities and detcision-making powers to lower levels of government can improve a state's ability to effliciently identify and address its citizens' needs. Being closer to the beneficiary population, lower levels of government have an inforrnational advantage over the federal government as they can better see their constituents' needs. They also know the local conditions that are importanit in internalizing benefits and costs of public service provision. Moreover, local governments have a greater incentive to improve service delivery, beneficiaries and citizen groups can easily monitor their actions and evaluate them through local elections, etc. Realization of these benefits, however, is not guaranteed. Careful attention needs to be given to policy design, particularly in quickly evolving, highly decentralized systems such as in Ethiopia. Devolution of Expenditure Responsibilities 1.12 In terms of intergovernmental fiscal relations, Article 94 of the Constitution provides the following three guidelines for public expenditure policy: (a) "The Federal Government and the States shall respectively bear all financial expenditures necessary to carry out all responsibilities and functions assigned to them by law," (b) "The Federal Government may grant the States emergency, rehabilitation, and development assistance and loans, due care being taken that such assistance and loans do not hinder the proportionate development of States," and (c) "The Federal Government shall have the power to audit and inspect the proper utilization of subsidies it grants to the States." 1.13 The first clause gives the states the duty and the right to plan and implement their own development policies, while the second two establish the federal government's role as supporter and monitor. In practice, the central government and its line ministries offer guidance and technical assistance, while the regional governments are encouraged to develop their own development strategies. The latter have taken over most of the day-to- day responsibilities for planning and implementing social policies. Subnational Budget Planning 1.14 The Ethiopian Constitution provides the directives for subnational budget planning. Article 50(4) requires the states to grant "adequate power" to the lowest units of government "to enable the People to participate directly in the administration of such L&egionalization and the Design of Economic Constitutions in Ethiopia 5 units." Article 89 (6) states that "Government shall at all times promote the participation cf the People in the formulation of national development policies and programs; it shall also have the duty to support the initiatives of the People in their development endeavors." 1.15 Creating the budget plans has two dimensions: identification of priorities and goals, and managing funds in order to fulfill these goals. The first of these tasks is participatory and bottom-up, but the second, financial management, is mainly carried out by the regional governments rather than subregional offices. 1.16 Identification of priorities and goals. This process involves multiple channels of information flow, vertically within sectors as well as horizontally across as many as three levels of government. Information gathering in the annual budget exercise begins at the woreda level where the sector development offices (and other agencies) prepare their plans and present them to the Woreda Development Committee (WDC). This committee reviews the plans, consolidates and adjusts them, and submits the proposal to the Woreda Council - an elected body parallel to the Regional Council - for approval. The information gathered by the woreda is passed upwards to the zonal level via two channels: in disaggregated form from sector offices to sector departments, and in aggregated form fi-om woreda council to zonal administration. The zonal sector departments aggregate sectoral information, adjust the plans to ensure that allocation of resources to woredas matches the longer term sector development goals, and pass the information to the zonal planning (dealing with the capital budget) and economic development (dealing with the recurrent budget) departments.4 The zonal planning and economic development departments use this information as well as the aggregate woreda budgets to create zonal capital and current budgets. These budgets together form the zonal development plan wrhich is then passed to the regional planning and finance bureaus. Like the transfer of information from woredas to zones, the zonal sector departments also pass up information on individual sector development plans to the regional sector bureaus. The regional sector bureaus create sector budgets in keeping with longer-term sector development plans, then pass their capital budget proposals to their planning bureaus and recurrent budgets to their finance bureaus. These two bureaus use the information as well as the budgets passed f-om the zonal planning and economic development departments to create regional capital and current budgets. These budgets are then submitted to the Regional Council for approval which in turn become law and are issued in the Regional Official Gazette. 1.17 One of the key features of the budgeting process is the strict hierarchy of spending decisions. First priority goes to funding recurrent expenditures (salaries, maintenance and supplies). After these activities are funded resources are allocated to financing the investment outlays associated with ongoing projects. Finally, any surplus budget is directed toward new investment projects. This facilitates implementation, timely project completion and avoids stretching investment programs too thinly. This separation between capital and current budgets, which also occurs at the federal level, is analyzed in greater detail in the 1998 Ethiopia PER Report Number 1 8369-ET. 6 Chagter 1 1.18 The most important way in which the real planning process differs from the idealized version (outlined above) is the timing of the announcement of budget ceilings. The idealized version of bottom-up planning assumes that budget ceilings are known at each step so that meaningful adjustments of development plans could be made. Unfortunately, this is not the case. The overall budget ceiling for the region is not known until late in the planning process.5 Moreover, the subregional governments rarely know the envelope they are working with as they create their budgets. The information they pass up to the regional government is thus more of a wish list than a calculated, well- debated development plan that is based on realistic resource constraints. The regional planning exercise is the key step in creating a yearly budget, for it is where the constraints are known and where decisions about trade-offs and cuts are made. In most cases, however, there is not enough time to go back to the lower levels of government with this information in order to have a proper iterative process. 1.19 Financial management. This process involves turning unconstrained wishlists into budgets within the resource envelopes and mostly takes place at the regional level. There are only a few regions where zonal administrations are actively involved in the financial management. The SNNP Region, for example, has a particularly strong zonal- level administration and zonal sector bureaus interact directly with regional finance and planning bureaus. Transfers from regional capital budgets are allocated to zones according to a population size, socio-economic status, and revenue-generation capacity. In the Amhara Region, (see Box 1.1) zonal administrations have the power to decide the capital and current mix as well as the sectoral allocation of a formula-determined budget ceiling. hi FY94, and FY95, budget ceilings were announced in late February of the fiscal year, FY96 they were released in May, and in FY97 they were announced in March. (The Government Fiscal Year is from July 7 through July 6.) Regionalization and the Design of Economic Constitutions in Ethiopia 7 Ai:i iDiI n i.-. 3iiar Rg i Al iqn ~ .c d. in i mph the im c ppula participaton i : ; ~ati. d .. og .t. mi~__ti . g``s- have bening thin i oi ?vwyii'pe a'idii tohe ait ?f heir subreinl overmens.. Intrgovernental 3 ,,~~~~~~~~~~~~~.. ........ & . P? ;Z A_.. .:?U 9- P.. 5 O............................... . t16nst1li the A e n p e - nudel o- thisevolving, increangly bottm-i pn . . ,,? ii A' rngi?iaot~ "::rni~sg e... .... of enie m ent .n th A. ;j.i>tL d~mou~rate. a tiyp~calzoul~ for ...thebuplenatln of rgionaliratin. n a.. ongoing. .exer .iment., Si: 3^ 3 ement issoe1ueuty ?iam Rd >Xa@w-@.o n pereeiv*d- F0r t rs, the ~egio ce mid ng burea divided ow- revenues as wl - .i b .,3 ... ...e.pl whk we. .reoil t.he rego burea. .. RP 9: 9 oft. ecs. The ?'a o m be g t a formuIa hi.. ~ ~hsur ha xiling saary liations ould in aet Uong reveq,mew.e, i~e~iner~ W&1Ue.theall.,caion nsea of .' in tht ajsteainies oral osswol s?~~~~~~~~~~~~~~~~~~~~~~ iDe: .: SWl ?i?.,~ ~~V iiid> ? ? .>h? .t ? o tl- f;gial&9 :It ~ ~ recmre~tt ...g.t __o..mm*iteslve Ieo inet an tode. ? ? di.ut:tat . sum tranerto thi: z--i_ eal ura us ion .gi he zaneIns.. A h.there.. .ns'.. ... in i~et recom imen.t budetsr 3?3~ ~ ~ ~ ~~~~~~~~t al,-?X -? 0?iit.S>%.lE]0 1f ??i3jo cvW c?n tSrcu8u cot an naimha w O *i* p.}nn. i thepow~rt6~letenino caital an curren expendtushs ow btee tranfere t'o t ClZrly,Ztixe vm nre z s' fluenc? ino. .!ur '* ?i flat?it e'.?rity t-he , o tensio be?e b?tmu an t?- i plnnn . . : V i 4 i i t~ ~~~~~~~~~~~. . .... e&.-X.. :WI;I4 ~ ~ but p~en~cesl~orexaple reciiry oev zoent~ eedoallcae ',3,ott1S zji M, hd,b'l ,,, sp3rt :~fm Hoee, tt3ouhprsuasion regional fficials wr i3 't? ? eiet? sa ??'':'.? ?l xeoez the iotaneo acntaility Moitrg te zones fromaboe a ?? d$ i? ?? 3B ?? ,f:::ib, ::'.... ..ly ...th .ee n uior genral- Meanwh Mmeile, a .. b l iB: s ... ..ra .... 1.2 Te itAmiisr o A s Aabprovrs are exeted froa wredatcivls aedan soph..i..i.t. b geta panning tpeir.broders Furmsrn als, Region al, the City of AddisAbabahas .e.ated foremaa.tl ititutions hopromoe ehnded-seorkarel cordntin.Te ueas r dvie it tre omites:abuge dmnitato ... . . .... ~~199g.tl coordinaion. Th bureau are di ided into -thre dommifttees a Sugt Iadmnstration 8 Chapter ] committee which includes both finance and planning bureaus, an economic committee which includes the infrastructure, trade and development, and agriculture bureaus, and a social committee which includes health and education. These committees are currently made up of the bureau heads, but city officials are thinking of including citizen representatives in future deliberations. The Gambella Region, a poor region, depends heavily on the regional sector bureaus to identify needs and plan development programs. There is very little direct participation by the woredas and kebeles. Even the newly formed zones do not play a strong role in the decision making and are often unaware of the process. 1.21 Implementing the Budget. Once the regional budget is passed, the expenditures for the year are set. Reallocating funds between different sectors is difficult at the regional level (it requires the Regional Council's approval) and almost unheard of at lower levels of government. Most regions allocate funds to the lower levels of government according to the inputs (workers, construction materials, etc.) that would be needed to meet existing recurrent cost commitments, finish projects, and, at the margin, start new projects that correspond to the overall regional development plan. Within the sectors, expenditure responsibilities are allocated to the lowest level of government capable of internalizing the costs and benefits of the service. In the health sector, for example, the woredas run the health clinics and health centers, while zones run the hospitals and/or referral facilities, and regions run specialized higher-level clinics. In roads, the Federal Government builds and maintains national highways,6 the regional government builds inter-zonal rural roads, the zones construct small roads within their jurisdiction, and woredas build and maintain paths within their areas. In education, the regions set curriculum policies, teacher standards, and coordinate the purchase of textbooks, school furniture, etc., while zonal and woreda administrations oversee the construction and maintenance of (respectively secondary and primary) schools in their area, and recruit teachers. In agriculture, the woredas take most of the responsibility for paying and hiring agricultural extension workers, while the zones coordinate efforts and ensure an equitable distribution of manpower, and the regions focus on work with region-wide significance such as research and policies to improve access to credit. 1.22 The regions thus emerge as the key players in Ethiopia's development strategy. It is the regions that aggregate communities' preferences, create specific development plans within the more general national strategies, and, in most cases allocate the budgets between sectors as well as between capital and recurrent expenditure. Expenditure Assignment: An Evaluation 1.23 Has the new and evolving regionalization strategy (based on participatory development) of Ethiopia led to efficient and equitable provision of public services? In this 6 The Ethiopian Road Authority (ERA) has also been contracted for some longer regional roads in Gambella and Amhara region. This federal body has a much more active role in the regions because the regions simply do not have the equipment, expertise, or access to contractors that the federal body has. Regionalization and the Design of Economic Constitutions in Ethiopia 9 context, have decentralized expenditure policies worked well? This section addresses these and related issues. Suggested options for reforms are outlined. 1.24 There are several dimensions along which public service provision in an economy can be evaluated: Is the role of the state appropriate? Is the level of expenditure sustainable? Are public expenditures being turned into desirable outcomes? Is it being done in an efficient, least cost manner? Is the distribution of public expenditure. across and within regions, equitable? 1.25 The role of the state. The Ethiopian Constitution reveals a strong preference for an activist, welfare-state government that is heavily involved in providing social and economic infrastructure for development. Article 90(1) states that "to the extent the country's resources permit, policies shall aim to provide all Ethiopians access to public health and education, clean water, housing, food, and social security." Resources, however, do not permit full public provision of all of these services. Moreover, having the responsibility for a service does not mean that governments should themselves provide that service. The regions should focus on appropriate government intervention rather than universal service provision by the public sector. Most regional governments accept this argument, but face a variety of problems. 1.26 In many regions, particularly the poor ones, a modern private sector capable of meeting the demand for social and economic infrastructure is at a nascent stage, or non- existent. The main obstacles to private sector development appear to be (a) a lack of access to credit (or other means) to finance the large initial investments, and (b) an inadequate scale of operations. In the roads sector, for example, regional officials seem eager to contract work out, but, in some areas, the private sector simply does not exist for contracting out road projects. In other cases, there are not enough contractors, especially on the maintenance side where equipment is expensive and profits are uncertain. While there are no easy and quick ways to overcome these problems, possible solutions are: (a) construction equipment belonging to the Government could be leased, sold, or rented; (b) the Government could facilitate credit availability by designing an appropriate insurance scheme; and (c) bundling of projects can also help increase the scale of operations and hence profitability of public contracts, but care needs be taken to avoid handicapping bidding for attractive projects by tying them to less profitable ventures. Also, any bundling scheme should be designed to lower the actual transaction costs of carrying out the projects not just simplify the initial bidding process. For most regions, moving toward widespread private sector supply of public services may be a distant goal. But injecting some element of competition into this arena would challenge public agencies to improve their performance, and communities would be empowered by the pressure of alternative choices. 1.27 The urban areas face slightly more complicated issues in private-sector development, for the services which they could contract out are unlikely to be duplicated in other parts of the country. The City Administration of Addis Ababa, for example, attempted to contract out liquid waste disposal, but was unable to find a contractor with sufficient capacity. Most contractors who enter the market do so in traditional fields such 10 Chapter 1 as road and social infrastructure construction, rather than the newer fields where government contracts are fewer and not as definite. 1.28 There is scope for improving the regions' ability to use the private sector. The Federal Government can help in this respect by running seminars or training courses on contracting, bidding, and working with private contractors. Although construction contracts are relatively straightforward and many regions have been using them for years, maintenance contracts are not so clear cut. Regional governments often have trouble drawing up and enforcing good maintenance contracts, mainly due to inadequate scale of operations. Rather than take the risk with an unfamiliar contract, regional officials continue to use the labor intensive "force account" method of maintenance that involves hiring temporary semi-skilled workers. 1.29 Efficacy and efficiency of public programs. Popular participation may improve the match between peoples' preferences and the mix of public services provided, but it does not necessarily improve the efficiency of allocation. While regionalization has improved the institutional environment for efficient public expenditure in Ethiopia, service delivery remains inadequate and in some cases the quality of public service is extremely poor. Simple basic public services, such as drinking water and sanitation, basic education and health services, street repair and lighting and trash collection, are frequently unavailable or in short supply. 1.30 In many regions social equity, rather than obtaining the highest marginal return in terms of the overall development plan, is the guiding principle for each year's expenditure allocation. Several regions have begun to use their own five-year development plans, but progressively each individual year's budget appears to be influenced more by the project wish-list emerging from the bottom-up budgeting process than by realistic assessments of the relative costs and feasibility of expanding coverage to a particular area. Popular participation also does not necessarily determine the most efficient sequencing of investment, particularly when local communities are not aware of the opportunity costs of the investment under consideration. For example, in the Amhara Region investment in health facilities were planned without adequate regard to the costs of expanding the coverage. In its quest for social equity and attention to all communities, some health posts were built before access roads were improved. A more cost-effective approach would have been to concentrate investment in the relatively accessible areas and expand coverage of remote areas as access to these areas improved and the cost of investment decreased. Another example of a lack of attention to sequencing comes from the Gambella Region. In this case, the majority of road funds are being spent on relatively expensive access roads to link remote areas of the regions. These roads are considered important for the region's development and for ensuring all citizens' access to health, education, and other services, but according to the annual budget and road costs data provided by regional officials, these plans would take approximately 15 years to complete. A more cost- effective option could have been to use more of the budget to build relatively inexpensive rural roads radiating from existing towns in order to increase surrounding farmers' access to markets and services in the towns. Regionalization and the Design of Economic Constitutions in Ethiopia 11 1.31 Another constraining factor in the efficacy and efficiency of public programs is the nationwide shortage of human resources. Regions have wide disparities in governing capabilities due to differences in the availability of skilled staff, particularly at the management and technocrat levels. In several regions a lack of professionals such as accountants, economists, engineers, managers and planners, hinders the implementation and efficacy of public programs. 1.32 Both federal and regional governments are addressing the national skill shortage by undertaking a variety of measures. The Regional Affairs Office within the Prime Minister's Office, the Ethiopian Management Institute, the Ethiopian Civil Service College, and the Ethiopian Mapping Authority are involved in training of officials from the regions. The Federal Civil Service Commission has also held management and administration training seminars for high and middle level professionals and managers. Moreover, the Ministry of Economic Development and Cooperation (MEDAC) has attracted and coordinated bilateral and multilateral assistance for training programs and advisory support in planning, implementation, monitoring and evaluation, and management. These institutions have adapted their programs to suit the regions' needs and a majority of the programs are specifically directed toward the less-developed regions. Regions are also developing their own training programs. Many regions have Teacher Training Institutes, vocational schools, and programs to train midwives and primary health care workers. In addition to increasing the supply of civil service workers, both regional and federal governments are working on ways to preserve the stock of human capital that they have. Loss of professionals to the private sector is being addressed in a variety of ways. A program in the Amhara Region, for example, allows doctors to work part-time in private clinics in order to keep them from having to leave the public sector for financial reasons. 1.33 Notwithstanding the current efforts to solve the problem, regional disparities require more attention. There may be opportunities for strengthening current mechanisms of federal support to the four least developed regions - Afar, Somali, Gambella, and Beneshangul/Gamuz - without compromising the objective of encouraging local self- reliance. One possible way of institutionalizing the regional capacity-building process is to consider the option of creating a cadre-based national civil service. Several federal countries (see Box 1.2, for an example of the Indian Civil Service) have such a system, which could be useful in addressing the capacity issue, particularly at the top. (Box 1.3 provides an example of the wonders of an appropriate incentive system in the civil service in a state in Brazil.) There is a risk here that creating a more unified civil service administration might generate pressures for unrealistically high salary structures or excessive staffing patterns, and through collective bargaining, and undercut regional 12 Chapter I autonomy. The civil service college already provides additional quotas for applicants from the least developed region, and there may be scope for developing the federal level, an affirmative action programs by which students from these regions could be trained to become accountants, economists, doctors, teachers, or other civil service professionals.' After training these students could then return to their under-served home areas. Finally, promotion of inter-regional coordination could be helpful in reducing governance and capacity problems. This could be in the form of inter-regional investments, joint ventures along the borders, development of integrated commercial development strategies such as a jointly financed road from a coffee area in one region to a market town in another. 7 Affirmative action for the least-developed regions does not mean that the comparatively well-off urban areas should be ignored. These areas require a more advanced set of skills for management, so comparing skill levels to other regions may not be so enlightening. The needs of regions like Addis Ababa are unique the challenge of monitoring over f00,000 taxpayers, of usig computer netw mo.are difficult to address in the context of large- scale training programs. Regionalization and the Design of Economic Constitutions in Ethiopia 13 1.34 Achieving equitable public service outcomes across regions. One positive outcome of the regionalization process has been a more equitable distribution of public expenditure between and within regions. The grant formula explicitly transfers resources to those regions with poor socioeconomic indicators and low revenue potential. The residents of states such as Gambella, Afar, Somali, and BeneshangultGumuz, receive far more in per-capita transfers from the central government than residents of better-off regions such as Amhara or Oromiya. Within regions, the distribution of expenditure also appears to be highly progressive. The Amhara Region, for example, distributes money to zones according to socio-economic need (among other elements in the grant formula). The SNNP Region uses the progressive federal grant formula to distribute funds to zones in the regions. Several regions have created special programs to redistribute resources to the nomadic people. Among public officials in the region, there is a strong belief in equity over other goals of public expenditure programs.' 1.35 Although regionalization has improved the distribution of financial resources between regions, variation in human resources and implementation capacity prevents these resources from being fully translated into equitable outcomes. The human-resources related inefficiencies outlined in the previous section are not distributed evenly across the country. Availability of trained doctors, nurses, financial managers, etc. vary from region to region. Data on intra-region inequities are not available, but anecdotal evidence suggests that the same kinds of inequities occur within regions. Officials in Gambella, for example, mentioned that, although money could be allocated, it was extremely hard to ensure service delivery staff in school, health clinics built, roads completed-in some remote regions. 8Officials in the SNNP Region were surprised when the Bank mission suggested the unit cost of providing facilities might vary and that there might be a trade-off between expanding service delivery and reaching remote areas; equity is regarded as the relevant goal. 14 Chapter I C:. ISSUES IN TAX ASSIGNMENT 1.36 Assignment of taxes to different levels of government involves several considerations: increasing accountability, preserving common markets, enhancing equity, and encouraging efficient administration These considerations, however, can often produce conflicting recommendations. The fiscal federalism literature does underscore several important considerations that need to be bore in mind while discussing tax assignment (Musgrave, 1983; Oates,n1991; Shah, 1994): . To provide a strong incentive for efficient use of public resources, the costs of public expenditure ought to be internalized among the taxpayers to the extent feasible. This encourages accountability and prevents an expense account spending spree. . Taxes that are strongly progressive and highly redistributive in nature should be levied only by the federal government unless the tax structure is near uniformn across subnational unitsh Otherwise, persons with higher incomes can easily migrate to units which impose a lower marginal tax rate. This is particularly true of tax bases that are not region-specific. Regionalization and the Design of Economic Constitutions in Ethiopia 15 * Immobile bases should be taxed by the subnational government, as socially inefficient factor movements (i.e., which encourage relocation from lower cost locations simply to avoid taxes) are ruled out in such cases. * Taxation of natural resources, due to the uneven regional endowment within a federation, are best taxed by the federal government to contain regional inequity. * Administrative capacity and costs should be considered in deciding which level of government should levy particular taxes. Economies of scale in tax administration are often relevant due to some indivisible costs and minimum scale of operation. Convenience in obtaining relevant information for proper assessments is also an important related factor. * Benefit taxes and user charges can be levied by all levels of government depending upon the benefit span and the level of government providing the service, respectively. 1.37 In general, tax assignment in various federations conform to these guidelines (Box 1.4), although there are individual cases of departures from these norms. ~~~~~~~l a sinet in- Feea 7nre . -- - . ...>':. - j , ..'. . ,, . .. . .. . . . . .. , , , .. ... .? t ? ,? ' i , ?: : i ' ? s i ?? ? i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.. .. .. . . ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... ; ? .... . :s?ii ??">; Ty?e of Ta? . Nuber sf ! ntr.Si.: ,,h . . s. ? .: ...?-i-.j. ;i'::P...'::. -? i? . !?;! ... , - ! i~~~~~~~~~~~~~~~~~~~~~~6.i i iS ti .... . .. .. ..... -........... ' ' . ; g i 0 i,f g ;.s- :! & biti!: i:M .. . -:.:. ustoms .. . 2 ;:fEN- T A l T.. s .. .. .. .. .... L...T 1.ls ..... e. . .. Sa.ta nao huiah?, I?naM...Z Misu' N'.eria, Pakistan de & . une Many~ ~~ ofteaoecuTries hav -ihy taie txas as it oeo h bodla ae 16 Chapter I Tax Assignment in Ethiopia 1.38 Tax assignment in Ethiopia is somewhat unusual in the sense the jurisdictions of the federal and regional governments are, to a considerable extent, structured according to taxpayer categories rather than by types of taxes. The exclusive domain of the federal government (laid out in Article 96 of the Constitution) covers foreign trade taxes, corporation tax, federal stamp duties, taxes on monopolies and tax on income and winnings of national lotteries and other games of chance. On the other hand, the exclusive domain of the regions (Article 97) is a small one including fees for land usufructuary rights and royalty for use of forest resources. The bases of income tax, sales tax and excise as well as some of the fees are shared between the federal and the regional governments, as described below: * In the case of public enterprises, ownership deterrnines jurisdiction; income of employees, sales, profits, dividends and production of federal enterprises are taxed by the federal government only, while those from public enterprises owned by the regional government are in the domain of the regional government. - Non-corporate private enterprises are taxed according to the base as well as form of organization. The income of the employees of all private enterprises are taxable by the regional governments. But the sales tax, profits tax and excises are levied by the regional governments only if the entity is a sole proprietorship or a farmers' co-operative; otherwise (say in the case of partnerships), they are taxed by the federal government. - In the case of companies, the profits and dividends are taxed jointly by the federal and regional governments; similar joint taxation is constitutionally prescribed in the case of large scale mining income, all petroleum and gas operations and royalties on such operations (Article 98). * Income of individuals like farmers and selfemployed persons, income from water transport services provided within the regional boundary and rental incomes from properties not owned by the federal government, are taxed by the regional governments. Other incomes are taxable by the federal governrnent. * Fees are collected by different units of government as per the constitutional domain in which the relevant service falls. 1.39 The Constitution also provides that any residual taxation power should be assigned by the House of Federation and House of Peoples' Representatives in a joint session. This provision is significant in that neither the federal nor the regional governments have residuary powers of taxation; assignment of any new tax is completely flexible. Since the Constitution does not even exhaust the list of commonly used taxes levied across the globe (e.g., value added tax, motor vehicle tax and entertainment tax are not specifically assigned in the Constitution), one could expect a substantial degree of evolution of the tax system in the medium run. Regionalization and the Design of Economic Constitutions in Ethiopia 17 Centralization of Revenues 1.40 The present distribution of tax revenue is heavily tilted towards the federal government in Ethiopia. As Table 1.2 shows, in 1997-98, the federal government collected 82 percent of the aggregate tax revenue and 83 percent of the aggregate non-tax revenue. This is mainly due to the concentration of larger tax bases within the federal domain. For example, the entire revenue from foreign trade taxes, the largest revenue earner (41 percent of the country's aggregate tax revenue in 1997-98) accrues to the federal government. Further, indirect domestic taxes that account for another 24 percent of the aggregate, accrue largely to the federal government, its share being 85 percent. Even in the case of direct tax revenues, where the federal share is the smallest, it is as large as 67 percent. 1.41 There are, however, certain developments which, in combination with the unique division of tax bases, are likely to shift the distribution of taxes from the federal to the regional level. As noted earlier, the tax base of personal income tax has been divided between the federal and regional governments and privatization of federal public enterprises, being an ongoing process at present,9 is likely to expand the tax base of the regional governments. Other factors that may contribute to this process include possible improvements in the tax administration in the regions and the persistent governmental efforts to boost the agricultural sector, which should raise agricultural incomes and thus, expand the tax base of the regions. Whether this process will substantially eliminate the imbalance in the tax collections remains to be seen, but it is likely to exhibit a significant impact only in the long run, as (a) the above mentioned factors can take effect only gradually, and (b) the sheer extent of centralization of revenue ensures that it will take a few years to substantially reduce it without completely disrupting the continuity of the system. Sharing of Joint Revenues 1.42 Joint revenues are not significant at present, and hence their distribution (after being collected by the federal apparatus) is not yet a settled procedure. Currently, of the total collections from these taxes, direct taxes are shared between the regions and the federal government on a 50:50 basis, while for indirect taxes the ratio is 30:70, respectively. So far not every regional government has received a share of this revenue as the process of distribution among regions is under discussion. 1.43 At present royalties from mining operations are small. This is a potentially large revenue yielding area, contingent upon survey, exploration and exploitation of mineral resources. Hence, it may be worthwhile to consider a suitable scheme for the distribution of such revenues. Although received wisdom points to a federal levy on these sources of revenue, an argument could be made that exploitation of mineral resources may impose a 9 Over 150 public enterprises were privatized by September 1997, while the process for another 13 had been started. The extent of privatization can also be gauged from the fact that the estimated proceeds from privatization in 1996-97 and 1997-98 amount to Birr 700 million. 18 Chapter 1 substantial cost on the regional government under the jurisdiction of which it takes place. It should therefore be compensated for this cost. Thus, a part of the revenue from such activities should first be set aside for the concerned regional government; the rest can form a divisible pool which can be distributed according to an equitable formula between the federal government and all the regions. The formula for determining grants would, correspondingly, require suitable adjustments. Decentralization and Tax Reform 1.44 At present, the regions in Ethiopia are rightly focusing on improving the efficiency of tax administration so that they can tap the full potential of their assigned tax base with the existing tax instruments. In the future, there are at least two possibilities of rearranging taxation powers in favor of the regions which would result into a lower dependence on federal transfers. The first one does not involve any new fiscal measures but the second one does. 1.45 Privatization of federally owned public enterprises and the regional tax base. The Constitution gives the regions the power to levy income taxes on private enterprises and individuals. To the extent that privatization of federally owned public enterprises will generate income to the private sector, the tax base of the regions will be automatically expanded. The degree of this expansion is not clear, given the existing revenue sharing provisions in the Constitution. In any case, this process does not involve changing any fiscal instruments as this is a simple case of tax base transfer. 1.46 Regionally administered VAT. The reliance of regional governments on federal transfers can be reduced and their own taxing powers can be enhanced by creating suitably chosen new tax instruments. The most promising area for reallocating the tax base from the federal to the regional governments would arise with the introduction of a comprehensive value added tax (VAT). The constitution does not explicitly deal with the assignment of VAT revenues, and, as noted above, neither the Federation nor the Regions have residuary transaction powers. Consequently, decisions regarding both the administration and assignment of VAT could lead to significant changes in the balance between the federal and regional tax base. Moreover, since VAT is a non-progressive tax, it is not subject to the difficulties which may arise from base splitting (Box 1.5). For example, a regionally administered comprehensive VAT with necessary safeguards for harmonization purposes could be studied. The VAT is a state tax in Brazil but a central levy in Mexico, where it is administered by states. India is studying the possibility of converting the inefficient system of state sales tax into a comprehensive VAT at the state level. Given that Ethiopia is already in the midst of exploring a comprehensive VAT, this particular option - regionally administered comprehensive VAT- should be looked at. VAT is an administratively complex tax, however, and given the limited administrative resources in the country decisions regarding tax administration have to bear in mind the potential gains from centralized tax administration. Even if a decision is made to implement VAT centrally, the possibility of assigning such receipts to regional governments may be worth considering to reduce their dependence on the budgetary subsidies from the federation. The introduction of VAT would have to grapple with the Regionalization and the Desgn of Economic Constitutions in Ethiopia 19 issue of sharing to prevent over-centralization of taxes. It would be highly desirable to set up a tax reform committee to go into these questions as well as the desirability of splitting tax bases and other administrative issues in detail. El~~~~o 1.:Sli xBae Apliation of a pro~~~ressive~~rat~ a split tax base ~~~om~ icatos tax adinisistratio.n....b...ost conntties p~~rsooal inootno is taxed at a progressis'e rate. l~tagutentin~ the total incrnte of a ta~........ .......... ineedntmcaiss s ~ hii~csiIa o opex~ddstiini tax sttuctnre whic is4i~cuU to adminster. This is whya co ebensive defiition of incomes" ennaidesed ideal be~~bre a p o~ressive mE. Nlutr;s~id l*roe~bs pltigas wie scp N lse~sOWWWO A etr~tmi ohv YOpehniE "noe a P. le nteewbs h~feeialgqvenmen In iuii thi cotidhesnppemei t~r regenalsurcatgeon te feeraltax TsI 1011t19i'is c.n asaei ni,weete.Oltn1icm a fpa.uni ee on th regonill~har o. ..I... A smpe ypthtialexini.ca ilutrteth pobemofaspi xbseTa toern And. Further~~~~~~wx asueth aetaXtu uea ot . it rgesiiybul I tmnl yuwoslabs. Asswnoarurco~~~~~~~5percensuptoanmeoweo$0nd45perc.....e......... .Bi~~~~r 36~~~~00() (l7~~~~~~~,$t)0 + 4A00 + IV)09) only...~~~~~~~~~T, ....... vetiflable by the income tax department ottbe~~~~ ....ta .ei.......... Table 1.1: Federal and Regional Shares in Public Expenditure in Ethiopia: 1993-94 to 1997-98 (percent) Category 1993-94 1994-95 1995-96 1996-97 1997-99 Total Federal Regional Total Federal Regional Total Federal Regional Total Federal Regional Total Federal Regional ______________________ (Min Brirr) Share Share (Mln BHirr) Share Share (Mlin Birt) Share Share (Min Birr) Share Share (Mln Brrr) Share Share General Administration 690.0 47.4 52.6 922.0 41.4 58.6 1018.0 41.2 58.8 1021.0 39.3 60.7 1226.0 37.4 62.6 Defence 663.0 100.0 0.0 737.0 100.0 0.0 762.0 100.0 0.0 787.0 100.0 0.0 847.0 100.0 0.0 Economic Services 2421.0 74.3 25.7 3014.0 65.8 34.2 2924.0 60.0 40.0 3597.0 63.4 36.6 4485.0 63.4 36.6 o/w Roads 374.0 80.5 19.5 572.0 74.5 25.5 368.0 68.2 31.8 902.0 63.3 36.7 1258.0 77.6 22.4 Agriculture 565.0 43.5 56.5 529.0 29.2 70.8 659.0 27.3 72.7 551.0 9.8 90.2 848.0 28.6 73.4 Social Services 1834.0 31.0 69.0 1909.0 27.7 72.3 2120.0 30.4 69.6 2364.0 25.1 75.0 2881.0 19.3 80.7 o/w Education 998.0 24.5 75.5 1132.0 22.6 77.4 1360.0 30.7 69.3 1448.0 22.9 77.1 1683.0 20.5 79.5 Health 378.0 17.8 82.8 431.0 13.9 86.1 470.0 18.2 81.8 589.0 20.2 79.8 752.0 9.0 91.0 Other 1488.0 87.1 12.9 1791.0 85.5 14.5 1697.0 84.3 15.7 1689.0 76.3 23.7 2044.0 76.0 24.0 o/w interest & charges 957.0 100.0 0.0 839.0 100.0 0.0 841.0 100.0 0.0 880.0 100.0 0.0 910.0 1000 0.0 Total Expenditure 7096.0 65.6 34.4 8373.0 61.7 38.3 8521.0 58.8 412 9458.0 56.5 43.5 11483.0 54.5 45.5 Source: World Bank database. Table 1.2: Share of Federal and Regional Governments in Revenues in Ethiopia:1993-94 to 1997-98 Categoty 1993-94 1994-95 1995-96 1996-97 1997-98 Total Federal Regional Total Federal Regional Total Federal Regional Total Federal Regional Total Federal Regional (mlnn. Birr) Share Share (min. Birr) Share Share (mln. Birr) Share Share (mln. Birr) (mln. Birr) Total Tax Revenue 3076.53 82.73 17.27 3878.77 83.18 16.82 4723.19 82.43 17.57 5322.24 81.8 18.2 5787.45 82.96 17.04 Direct Tax 945.22 55.49 44.51 1311.59 59.62 40.38 1753.84 64.47 35.53 1907.58 61.32 38.68 2030.22 61.58 38.42 Income and Profit tax 899.66 58.3 41.7 1230.79 63.54 36.46 1648.85 68.58 31.42 1748.16 66.91 33.09 1859.79 67.23 32.77 Indirect Tax 834.14 88.19 11.81 945.55 87.02 12.98 1155.43 82.09 17.91 1720.23 81.84 18.16 1376.88 85.02 14.98 Foreign Trade Taxes 1297.17 99.06 0.04 1621.63 100 0 1813.92 100 0 2144.43 100 0 2380.35 100 0 Non-Tax Revenue 832.82 80.3 19.7 1908.34 87.73 12.27 2099.44 85.74 14.26 2024.75 81.05 18.95 2242.61 81.96 18.04 Source: World Bank database. 2. FISCAL IMBALANCE, INTERGOVERNMENTAL TRANSFERS AND SUBNATIONAL BORROWING ,4 direct outcome of regionalization in Ethiopia is fiscal imbalance. Presently, the twin issues of regional and verticalfiscal imbalances are being addressed through a system of block grants from the federal to the regional governments that focuses on regional equity and the righting of past wrongs. Despite there being a Constitutional provision to borrow domestically, very few regions have exercised this power. :2.1 Regional boundaries in Ethiopia have been drawn along ethnic and national lines and the process has highlighted wide disparities in governing capabilities across the eleven regions in the country. Moreover, the differing abilities to spend public resources in a cost effective manner and to raise resources commensurate with the taxable capabilities have accentuated regional fiscal imbalances.'° The City Administration of Addis Ababa finances nearly all its public spending from its own resources. Regions such as Beneshangul/Gumuz and Gambella barely manage to finance ten percent of their budget through local resources. On average, in 1997-98, regional governments financed 31 percent of their budgets from their own revenue; the remainder was financed through a subsidy from the center. 2.2 Vertical fiscal imbalance - the disparity between revenue means and expenditure needs at various levels of government in a federation - in Ethiopia is a byproduct of the Constitutional division of spending and taxing powers between the federal and regional government. As discussed in Chapter 1, regions have many expenditure responsibilities but few revenue yielding tax powers. Most revenue yielding and buoyant taxes (e.g., foreign trade taxes) are assigned to the federal government. This vertical imbalance has increased over last several years. Currently, the twin issues of regional and vertical fiscal imbalances are being addressed through a system of block grants from the federal to the regional governments. 2.3 This chapter covers fiscal imbalance, intergovernmental transfers and subnational borrowing in Ethiopia. First, the twin concepts of vertical and regional fiscal imbalances are discussed. A review and an evaluation of the design of the intergovernmental transfer mechanism in Ethiopia follow. Next, a brief review of the theory of subnational borrowing is introduced with a discussion of current issues in Ethiopia. ° In fiscal federalism literature, this is also referred to as horizontal imbalance. 24 Chapter 2 A. FISCAL IMBALANCE: VERTICAL AND REGIONAL 2.4 Assignment of expenditure and tax responsibilities is typically considered independently, with different principles governing the two sides of the budget. More often than not, this results in some centralization of revenue sources. Subnational capacity lirnitations with respect to tax administration often reinforce this tendency in developing countries. This is one of the key reasons for the emergence of vertical imbalance. In addition, when inter-regional equity is an important consideration, as indeed is the case in Ethiopia (see below), the federal government needs to hold some resources over and above those required for its own expenditures to distribute them among the subnational units in a way that would promote equity. This also contributes to vertical imbalance. 2.5 A high degree of regional imbalance reflects regional inequality and therefore, a need for equalizing transfers. Apart from the fact that regional equity is an objective in itself, high levels of regional imbalance could also induce movement of factors of production (mainly labor), in search of better opportunities and better public services. This is often observed at the international level in the form of "brain drain" within a federation. This is a less emotive issue, but in fact more prevalent. Vertical Imbalances in Ethiopia 2.6 The following formula provides a measure of vertical imbalance. It attempts to measure the degree of control exercised by the federal government over lower levels of governments: 1-[(Rs/R)/(Es/E)] where R = state revenues; R = aggregate revenues (federal plus states); E = state expenditures; and E = total expenditures (federal plus states). By this measure, a coefficient of zero - low vertical imbalance - indicates that state governments are pretty much autonomous in their decisionmaking as their revenue and expenditure shares (in the aggregate) are the same. On the other hand, a coefficient close to one indicates absolute federal control over states as the states' revenue share (in the aggregate) is extremely small in comparison to the expenditure share. 2.7 Using the above measure of vertical imbalance, the coefficient for Ethiopia- using 1996 revenue and expenditure numbers - is .52, a fairly high number. For comparison purposes, Table 2.1 provides this measure of vertical imbalance in some other federal countries. As seen from Table 2.1, barring South Africa (.88) and Bolivia (.66), all other countries among the presented ones exhibit lower vertical imbalance than in Ethiopia. Malaysia (.37) and Mexico (.36) are said to have significant vertical imbalances, but they are significantly less than in Ethiopia. It may be interesting to note that of these countries, Bolivia, Malaysia and Mexico have more centralized systems in terms of both public expenditures and government revenues; in South Africa there is a significant degree of centralization of revenues but not of expenditures. Fiscal Imbalance. Intergovernmental Transfers and Subnational Borrowing 25 Table 2.1: Vertical Imbalance in Selected Countries Country Year Ending Subnational Subnational Vertical Revenue Share (%/6) Expenditure Share (%) Imbalance Argentina Dec. 31, 1995 40.3 41.5 .03 Australia June 30, 1996 28.2 37.3 .24 Bolivia Dec. 31, 1996 6.8 19.8 .66 Brazil Dec. 31, 1993 24.3 27.4 .11 Canada March 31, 1995 45.3 49.4 .08 China Dec. 31, 1996 51.0 54.7 .07 Ethiopia* July 7, 1996 19.8 41.2 .52 Gennany Dec. 31, 1996 24.5 28.9 .15 India March 31, 1995 28.0 39.1 .28 Malaysia Dec. 31, 1996 9.7 15.5 .37 Mexico Dec. 31, 1994 12.4 19.5 .36 SouthAfrica March31, 1995 3.8 30.9 .88 LrSA Sept. 30, 1995 39.9 32.8 -.22 * Computed on the basis of data from the World Bank database. Source: Computed from data contained in Government Finance Statistics Yearbook, 1997, International Monetary Fund, Washington DC Note: Vertical imbalance is computed as: I1-{(RW / R)] / (E'/ E)}1, where R = Total revenue and E = Total expenditure, the superscript 's' denoting states. The measure suffers from an obvious defect of not considering local govermnents, mainly because the coverage of these in the cited data source is extremely uneven across countries. 2.8 A high level of vertical imbalance could lead to efficiency problems. Because the costs of public expenditure are not fully internalized by the regions and are supported by transfers from other parts of the country, there is a reduced incentive to provide public services in an efficient manner. Moreover, the tax burden of public services is usually lower than the cost of privately supplied services. In turn, this could distort the public- private mix of services that the citizens normally demand. For example, the fiscal federalism system in Ethiopia provides an incentive for the regions to wait for a publicly financed school even when it is possible to start a private school. Given the level of vertical imbalance, the additional tax cost of a new government school (in most regions) would be much less to the community than that of a private school. A high level of vertical imbalance also involves some loss of subnational autonomy since expenditures are dependent upon transferred resources. Too large a deviation from the grantor's (the federal government) preferences would invariably cause hesitation in adopting such policies on the part of the grantee (regional governments) even when warranted, because of apprehensions regarding loss of grants that may follow. This, however, does not appear to be the case in Ethiopia; there seems to be a broad agreement on national and regional priorities. 26 Chapter 2 Regional Fiscal Imbalances in Ethiopia 2.9 Regional fiscal imbalances in Ethiopia are in large part a reflection of the fact that the four least developed regions of the country (Afar, Beneshangul/Gumuz, Gambella and Somali) are in many ways accorded special treatment by the federal government (see Chapter 1 for details). A formal budget related measure of regional fiscal imbalance is presented in Table 2.2. The measure is the ratio of regions' own revenues to total expenditures. Since none of the regions borrow (in practice), this measure of fiscal imbalance is comparable across regions. The measure also reflects the level of economic activity in the regions as higher levels of economic activity lead to a bigger revenue base, and as long as revenue effort is not markedly different among regions, the bigger base is likely to result in a greater degree of self-financed expenditure. Table 2.2: Regional Imbalance in Ethiopia Region 1993-94 1994-95 1995-96 1996-97 1997-98 (Budgeted) Tigray 18.6 19.6 28.9 26.6 23.3 Afar 8.8 13.2 9.7 22.0 4.4 Amhara 18.8 17.4 18.6 19.1 18.8 Oromniya 29.2 27.4 29.7 30.1 25.6 Somali 60.1 42.5 24.6 23.0 12.8 Beneshangul 6.6 7.2 5.9 6.8 4.5 SNNP 19.2 18.7 21.5 23.1 18.5 Gambella 8.8 10.8 5.8 8.7 4.8 Harari 33.7 15.7 18.8 19.6 8.9 Addis Ababa 63.8 67.6 75.8 99.4 76.4 Dire Dawa 89.8 68.8 50.9 53.4 33.6 Source: Computed on the basis of data from the World Bank database. Note: Regional imbalance is computed simply as a ratio of own revenues to total expenditures. Since none of the regions can run budgetary deficits, this measure is comparable across regions. 2.10 The ratios (presented in the table) for the first few years cannot be interpreted easily. They probably reflect abysmally low public expenditure levels in some of the regions (particularly Somali), which pushes the ratio up. The numbers for the last two years - after the regional budgets have better adjusted to the present federal system - are better indications of the true degree of regional fiscal imbalances. Only Afar shows large fluctuations even in these two years; examination of the underlying data reveal this is mainly due to large changes in its expenditure. If one were to separate the special city administrations of Addis Ababa and Dire Dawa, the remaining regions of Ethiopia can be divided into four groups: one with ratios consistently near 20 per cent to about 30 per cent (Tigray, Amhara, Oromiya and SNNP), second of those consistently exhibiting ratios below 10 per cent (Beneshangul/Gumuz and Gambella), third of consistently declining (Somali) and fourth fluctuating (Afar and Harari). The first group comprises the relatively developed regions, while the second group consists of less developed ones. Somali's expenditure levels are fast rising as a result of the equalization process and hence the ratio exhibits a declining trend. The trends for the other two regions have fluctuated too much to provide any meaningful explanation. These numbers clearly show the influence of Fiscal Imbalance. Intergovernmental Transfers and Subnational Borrowing 27 intergovernrnental transfers and the substantial impact of the present system of subsidies in terms of shifting resources to the least developed regions. B. INTERGOVERNMENTAL TRANSFERS 2.11 As discussed in the previous section, the current expenditure and tax assignments in Ethiopia have led to a large vertical fiscal imbalance. In the last few years, regional governments were able to finance, on average, 30 percent of their expenditures from their own revenue sources; the remaining 70 percent was financed by the federal government in the form of a transfer. In official budget documents of Ethiopia, this transfer is called the "Budget Subsidy." Intergovernmental Transfers: A Conceptual Perspective 2.12 In a federal system of governance, intergovernmental transfers - revenue sharing and grants -can correct for fiscal imbalances that result after expenditure and tax responsibilities have been assigned. Revenue sharing, which normally means tax sharing, is resorted to mainly for the purpose of reducing vertical imbalance. It can be achieved through (a) tax base sharing, (b) tax revenue sharing. Base sharing implies taxation of a common base by two or more levels of government. The examples of base sharing come mainly from industrialized countries. The system of base sharing is almost non-existent in developing countries largely due to its relatively complex design and administrative requirements. In Canada, goods and service are taxed at the federal level, but there is also an independent provincial sales tax on them. The United States has a 'dual' income tax system by which the income of its residents is taxed by state-determined rates, in addition to the basic federal tax rate. However, the system is not fully harmonized and income coverage and exemptions vary from state to state. Revenue sharing, on the other hand, is quite common in developing countries. The process involves actual collection of the tax by only one level of government, followed by a sharing of the proceeds with other level(s). Tax collection in Australia is highly centralized and revenues are redistributed through a system of revenue equalization transfers and conditional grants. In China, the proceeds of a centrally administered value-added tax are shared with provinces on an equal basis. The sharing of revenue from personal income tax and excise duty in India provides another example. In developing countries, revenue sharing is often used to address a variety of objectives including promotion of regional development and fiscal equalization. 2.13 Intergovernmental grants are universally used in federal countries. These can be of different types with different objectives. Unconditional or block grants are lump-sum grants from one (usually higher) level of government to another without any conditions. These are meant to simply provide additional resources to the recipient governments to compensate for lack of adequate revenue. There can be conditional grants as well, which are usually intended to encourage greater subnational expenditure in areas that may not be accorded the same priority by the recipients. Examples come from the case of spillover benefits of regional expenditures such as sharing of river water and protection against pollution. In other cases, there could be a requirement of some nationally determined 28 Chapter 2 minimum level of basic services such as access to primary education. The nature of grant conditions can also vary considerably. Specific purpose grants constitute an important category of conditional grants. These grants are intended only to add to the resource availability of the grantee government for given purposes when there are no matching requirements. Specific purpose grants can have matching requirements which can change the allocation of public expenditures at the level of the recipient government. In addition to all this, grants can be close-ended (with a predetermined maximum limit) or open-ended (no fixed limit). In general, most federations use a combination of different types of grants to suit their requirements. Intergovernmental Transfers in Ethiopia 2.14 There are several distinguishing features of the intergovernmental transfer mechanism in Ethiopia. The transfer comes in the form of a block (unconditional) grant and the regions are free to spend according to their own choice and preference. (Table 2.3 provides the per-capita regional numbers for 1997-98.) In the regional budget process, once the budget subsidy is known, the regional governments decide their own spending plans for the year. Until now, however, there does not appear to be a major divergence between the preferences of the two levels of governments. Table 2.3: Budget Shares and Budget Subsidies for Regional Governments in Ethiopia: 1997-98 Region Budget Share Budget Subsidy Per Capita Subsidy Share (per cent) (million Biff) Subsidy' (Biff) (per cent) Tigray 7.7 255.7 76.1 7.8 Afar 5.0 225.8 199.6 6.9 Amhara 19.2 725.1 49.1 22.2 Oromiya 25.3 826.2 41.3 25.3 Somali 7.6 282.3 142.6 8.6 Beneshangul 3.6 161.1 326.8 4.9 SNNP 15.9 554.1 50.1 17.0 Gambella 2.8 126.1 646.7 3.9 Harari 1.8 77.0 534.7 2.4 Addis Ababa 7.3 0.0 0.0 0.0 Dire Dawa 3.8 32.0 115.5 1.0 Total 100.0 3265.3 58.6 100.0 Source: Basic data are from MEDAC and World Bank database. Note: Population figures used are for 1996-97. 2.15 The budget subsidy to each region is based on three factors: (i) the region's population share; (ii) its index of a set of development indicators relative to the national index; and (iii) the region's projected share of revenue in its budget relative to the planned performance of all other regions. Since its inception in 1995-96, this formula has been revised two times (See Box 2.1). In its last revision prior to the fiscal year 1997-98, the index of development indicators were changed. In addition, the relative weights of the three factors were changed to 60, 25 and 15 percent, respectively. Fiscal Imbalance, Intergovernmental Transfers andSubnational Borrowin?g 29 Box ;1. EvoltXion of:the ou orFederal Grants (Subsidies)to drl tran se to regions in E:thiopia first bega in 1992-93, when rants Jwere,'d hoc in are, s on approved i.prqje ,ct .ots of the regions and:an asmssment of required assistance for the in divdua projet. Form- la deter-ined grants ,,,, bck only to 1994-95, -whe a fairly cmplicated frmulawasute a get i. as frcapital expendituresoftherens were determiined odititor-s: population (30. ,-e,,ent);5I-4istce representIing a rion's relative development based on eigh tftr a T a: ypewalregion(25pecenttaxeffort(20 capital expenditure in the previos y. pe pe a na-ce (10 percent)- Rdeurrent expenditures grants were deteormied on the sis of the numberof admin iisativ units -at the sub-regional levels, the strauctu ofbureauxand offics and own even es f the r .- 1995-96 saw considerable ification of tire formula tdetennine the tota I diture fo t regions. Only three fintors population, 1-distance and budgeted sEtat rev enues were used, vaul basing equad .99 7, the fomulaubsti t -distance wiha developet in:dex comprisinigfive indicat ( of th lsn of imar ssn of telephone lines, electricity consumption an d leng In 19974f, the formulaw w d ters of weiahting, with population gven a weiho 60.percet, development indx a- weight of -25 percnt and the revenue effort vaable awet f 15 pe t.The development index was-re-estimated on the basis of six underlying factors: educatio leve, helhsc indicator, road density, electriity cnsumption, waterisupply and telephone coverage.. -iTRx -60x PGPiPQP+ .025xDd + I5x ,t- Di,t-l t-1.t.1)- ....:-:.---i - --Transfer to-region 1TR = Tot of t e er all egions. --- POPi - Popula-tio of reio i.- --:- P Tota ol population over all regions. ;Di- - -- AX c site inverted devent index, which for g iis q .6/(Ei+Hl+Ri+ELi+'Ei+Wi) wThe official title for this ind 'nS f combin icators of -the: level of velopment of a region".-: D -- :--== S- o-.tvalues ot Di over all regions. D: .n . .- 6. lii = Index ~~~~t'orcombie niaosfreducation in region i(aeir f he equa-:- lly weighted'sb-componenits: average class:size pui-teace rt'io, pm and secondary sch oo participation rate as percentage of ag.e. coho -- . dex for c indicators for healthi int region i (nma up of six e y: weight sub-coonents: number of linics, numb of tors f nurses -num o he health asitants, nuiber:ofhospialbeds,a t rat of undr-es).- -; = Index o road density in region i (road length per km2). .Li = : nd oflex 1ctricn power-sales regionfi (KHper thousand pople). TFRi - nd - : of tlephone line-densit in region -i (telephone liis pe tu dpole) Wi Index of coverage of safed gwater in region i (percentae of p: -t h saf drinkifng w ater). d I: -- REV t-l: PlIannedown revenue aisd byregion iinthepriorye (since its"Own"- revenue it e s th region' s:hare of joint revenue w t Fe even mn t: . e t). i BIJDi,t-I --Plannd budget for region i th e year prior to theYear REVt-l anned own revenue of all rei-ons in therio ye :---BUDt-l Sum ofplneret of-all re in the pior year. 2.16 The own revenue performance factor in the budget subsidy formula is designed to provide incentives for the regions for better revenue mobilization. However, the region's own reveniue is deducted from the gross budget subsidy in determining the net transfer. 30 Chapter 2 While in principle this mechanism may reduce the incentive to increase own revenues or accurately report revenues - underestimating the revenue in order to get a bigger share of the budget subsidy - for the regions, it is understood (based on discussions with several regional governments) that currently this type of incentive is not a relevant issue. It is being outweighed by a 'national pride' factor which is associated with being able to stand on one's own feet without having to depend on the federal government. 2.17 The inter-governmental transfer system involves a "Budget Offset" provision by which reductions in the budget subsidy to the regions are made for an amount that is equal to external assistance. The rationale behind this offset mechanism is to ensure that the overall resources - domestic and foreign - available to the country are shared in accordance with the national priorities, rather than being decided by the donor community. Intergovernmental Grants Mechanism: An Evaluation of the Ethiopian System 2.18 The Ethiopian intergovernmental grants mechanism can be evaluated along several dimensions: 2.19 No true incentive for revenue effort. It appears that the revenue incentive factor built into the budget subsidy formula (see Box 2.1) is largely cosmetic and does not create a direct incentive for regions to generate own revenue. In reality, the formula provides a disincentive for revenue effort. To see this, suppose x is the weight attached to revenue effort. An increase of M units in the revenue effort results in a rise of only xM units of expenditure, while the loss in grants is (I-x)M. When x is as small as 0.15 as it is now, the loss in grants is clearly larger than the rise in expenditures. 2.20 The revenue incentive clause built into the budget subsidy formula needs to be revisited. While the current presence of national pride may be leading towards less reliance on the subsidy from the federal government, the provision of the net subsidy mechanism is clearly not one which promotes better revenue collection. In a democratic system such as Ethiopia, the regional governments will have to face the electorate periodically, and the taxpayers may not be willing to pay higher regional taxes if their marginal tax payments raise the level of public expenditures only by a small fraction. 2.21 The "Budget Offset" Issue. The "Budget Offset" system protects the country from donors influence in determining the national priorities, but at the same time it blunts any incentives on the part of the regional governments to look for external sources of finance. There is no additionality at all in such funding, as obtaining funds from this source entails an equivalent reduction in the federal grant. Although such funds could bring in technical expertise and better procurement or reporting practices than normally adopted, it also entails much more paperwork and detailed planning as well as some uncertainty regarding the actual flow of funds. Discussions with regional officials revealed that donor projects are not aggressively sought because they simply crowd-out federal financing of projects. If anything, donor projects, particularly those from the multilateral institutions, are seen as excessively cumbersome (in terms of the regulatory requirements, Fiscal Imbalance, Intergovernmental Transfers and Subnational Borrowing 31 book-keeping, reporting etc.). Very few regions indicated better technical expertise that could be a part of the donor projects as a factor that would influence their decisions to seek external financing. This was particularly true in the case of social sector projects. It is therefore not surprising that there is a marked disinterest among the regional governments in obtaining donor funding. 2.22 Evidence from the regions suggests that the strategy of keeping donor agencies at an arms length in influencing the national priorities is working. The system maintains equity among regional expenditures by avoiding distortions induced by the flow of external funds among regions. This may very well be a desirable policy option of the Government of Ethiopia. However, one possible downside of this strategy is that the total pot of external assistance is exclusively controlled by the federal government and it blunts any incentives on the part of the regional governments to look for external finance. It may be possible to maintain additionality without adversely affecting the equity of the system through the use of some innovative mechanism. To do this, donor funding must continue to be allocated by the federal government keeping regional equity in view, but this can be (lone outside the federal grants (or subsidy) system, so that additionality is maintained. This may not sit comfortably with the donor community as they normally like to choose the targets of their assistance. But such a system can be made interactive with proposals invited from regional governments and installing a process of choosing proposals that involves all the concerned parties including the donors, without violating regional equity. Post-allocation interaction among the donor agency and the recipient regional government also can be accommodated within this system. 2.23 Exclusive Reliance on Block Grants. At present, there seems to be a wide support for the system of block grants in Ethiopia. The exclusive reliance on block grants"1 may also merit a fresh assessment and there may be a scope for introducing incentive based conditional grants in the future. 2.24 Generally in countries like Ethiopia which are starting afresh with a low level of development, priorities are not very difficult to determine. It is therefore not very surprising that at present there is a remarkable level of agreement on development priorities. This is clearly reflected in the three sectoral development programs (education, health and roads) that the country has embarked upon. As long as regional preferences more or less correspond to national ones, the case for conditional grants is quite weak. However, as the country moves along the path of economic and social development, this uniformity of preferences may weaken over time, and in that situation, conditional grants would be legitimate tools for the federal government to influence the spending pattern of regional governments. (It should be noted, however, that in many federations, this is interpreted as a reduction in subnational autonomy and hence resented.) To ensure that national priorities are included in the regional development plans, it may therefore be worthwhile to start looking at conditional grants as a future option. There is a wide 11 The subventions from the Road Fund is an obvious exception, since the funds so transferred can be spent in the road sector only. 32 Chapter 2 literature available on the experience of federal countries - developed, developing and transition - in instituting conditional grants. 2.25 Revenue sharing as a mechanism of resource transfer is similar to block grants in terms of its impact on subnational budgets and has no special role in the present system. However, with the maturing of Ethiopian democracy, it can become an important tool for intergovernmental transfers as it has the advantage of imparting buoyancy to the transfers, it would also be less prone to year-to-year fluctuations, thus providing a steady source of funds for the regions. It would also provide the regions with a stake in taxes levied by the federal government, which is important in the context of competitive federalism (regions trying to maximize the welfare of their own residents, and by implication their own revenues, even at the cost of other regions or the country as a whole). 2.26 Institutional Mechanism. Finally, the issue of an institutional mechanism to determine the level, form and distribution of grants merits serious consideration. At present, MEDAC and the Ministry of Finance at the federal level formulate and execute the entire system of federal grants to regions. Since the present system of fiscal federalism in Ethiopia can be characterized as 'cooperative" federalism - few disputes and widespread co-operation between the federal government and regional governments on the one hand, and between regions on the other - this mechanism can work smoothly. However, international experience shows that cooperative federalism often changes into "competitive" federalism - limnited cooperation, larger number of disputes, and increasing degree of difficulty in achieving an amicable settlement - over a period of time and in such an eventuality, the present system may need to become more institutionalized to protect the current consensus.12 2.27 One possible way to limit the notes of competitive federalism is to set up an independent Grants Commission to ensure that the inter-governmental transfer mechanism is seen to be fair by most, if not all, stakeholders, and is not subject to a frequent tinkering exercise involving minor changes here and there. Such a Commission would have a technical committee - consisting of domestic and international experts (if needed) - attached to it in order to come up with different formula options on the basis of the needs of the country's overall development. Also, in keeping with the spirit of federation, the Grants Commission could have a balanced structure representing ethnicity and nationalities of Ethiopia. Among the industrialized countries, Australia has such a system where a permanent independent agency (Commonwealth Grants Commission) consisting of professionals determines the intergovernmental grants, on the basis of formula(s) worked out through continuous research in this area. There are, however, federations where intergovernmental grants are determined and disbursed exclusively by the Ministry of Finance; one example is Canada. There are also cases of hybrid systems such as the one prevalent in India, where two agencies (the Planning Commission and the Finance 12 Elazar (1991) discusses the case of the United States; a similar change has actually taken place in India over the last forty years with the relationship between the central and the state govermments gradually changing from being partners in development to almost an adversarial one. Fiscal Imbalance. Intergovernmental Transfers and Subnational Borrowing 33 Commission) decide on the quantum of federal grants to states, besides some amount of discretionary grants provided by the various departments of the central government for specific purposes. While the Planning Commission is a permanent agency of the central government, the Finance Commission is a constitutionally mandated institution that is independent of the central and state governments, set up every five years to give their recommendations regarding tax sharing and grants, after which it is wound up. Thus, there are alternative models to choose from, and an informed choice would go a long way to avert future problems. C. SUBNATIONAL BORROWING 2.28 There are no theoretically derived guidelines for an optimal structure of borrowing by subnational governments. However, a general rule of thumb that is applicable to any borrowing per se is that a debt is useful only when it is invested to yield a rate of return that is at least enough to repay the principal along with the interest due. In the case of a government, the required rate of return may be defined as the social rate of return. However, unless the increased public welfare translates into greater availability of resources for the government, incurring the debt, the servicing of the debt in purely financial terms, can become a major problem. 2.29 In Ethiopia, the subnational governments are technically allowed to borrow from domestic sources with the concurrence of the federal government. (Article 51 of the Ethiopian Constitution states that regions in the federation can borrow from domestic resources under the terms and conditions set by the federal government.) Nonetheless, in practice very few regions have exercised this power. Several regions, however, have used their borrowing privilege in an indirect way, by incurring contingent liabilities. A number of regional governments have provided credit - with the concurrence of the federal ministry of finance - to farmers in the form of loan guarantees to commercial banks (see Box 2.2). 54 Chapter 2 22:Cntnet Liabilite of the Rei,m Govenmets d to Arcltural Lu - 0- t0 22 C 'nchemes improvig agneulStural poctivity isE enormos p-sn agiulue using low -ev-s of - 0:;~seeds ad ote moder agrclua ainp ol haves an mprat pacto devel op. etenti" ' i - . ....... T m.c e i ew 0 lilo Wd aa 0000 a- g- ,. eB ....... .. .. ... .... ..... 3,.3 .2 |:: -:S,.:s .'!g:1 '3.................................. 33 t;;~~~~~~~~~~~~~~~~~~~~~~~~~~... .. .,$ .0E '. kk the -xas.O.gr .r.cd.I Rural Regitona Ethoviern edits tnolthemerurl sher wagriculualcredpiorto 1991me halthseghrth statnacs. red f omewhat, re ionwide 'hi'annuaov dgeis as ao d -.r fo ti the lons m exotoes dovnot hav aess to asthef servis e. o t cropfalu.ornet, ot bis ;ras hete, o s ot.h we000 cear todefaul by te averag foreoSub Sahara goveia regnt anot as of Bp a t 59d uthsi oft258 deayedbranchespwere inadcuningAbasa.ms. The past,estato whikh thae dant.t.ge.s agriculteralonetor,nbulithesliffortshfailbdet reachze iunleargeNumbersthflesans, avaldl dadnta hav 2pab begional governments. suglvemen pt i the agricultural crentsc involveral ret. Craditg lavailabilit hsgiicncreised I dr ticAmall ndra tegisn prgamks lrom ab81 32illionBirn draathe 1996ks. 7eao to 2 ro meion, bui inly9e9o- e an red 2 mllioni in lealv- r the rofons regovern ment tiopay I I di o Birr, tnd perenhv of rhe credi tmall f aomouts of rbdit. Thei rexpnionalgoeredit,omincraed the agressive exeson services treand thehi.aebssiby ceutreach,h widespread default by the farmers. Moreover, regional governments are not adequately compensated for this risk; they collect 1.5 percent interest rate on the collateral posted, and this is often delayed due to poor accounting systems. The extent to which the dangers of these contingent liabilities have been realized is unclear. Nevertheless, available data on crop season credit recovery, suggest that the regional governments' involvement in repaying loans poses significant risks. In the Amuhara Region, banks lent about 32 million Birr in the 1996-97 crop season, but only recovered 21 million, leaving the regional government to pay 11 million Birr, or 35 percent of the credit, from its own budget. The regional government increased the pressure on farmers to repay their debts by cutting the zones' budgets according to the default rate in each area. The zones, in turn, cut woreda budgets, thus reducing the money available for communities to build schools, health posts, an~d other necessities. While this strategy has reportedly enabled the regional government Fiscal Imbalance, Intergovernmental Transfers and Subnational Borrowing 35 to collect most of the year's outstanding debt, it may have also penalized communities who were hit by crop failure or other unavoidable reason for default. 2.31 A potential reform agenda. The appropriate role of regional govemments should be to facilitate the development of credit and insurance schemes in order to help the farmers and not assume what could prove to be costly contingent liabilities. Protection (in the form of loan guarantee) should be provided only to an appropriately targeted group of poor farmers. The Amhara Credit and Savings Institution (ACSI), a non-bank financial entity provides an example of a private sector alternative to public guarantee of credit. The institution is geared to provide credit to promote agricultural (and non-agricultural) activities among the rural poor and introduce modern savings and loan practices in the region. A five-person screening committee, made up of community members, kebeles administration officials, and a credit officer identifies potential beneficiaries based on criteria such as poverty, freedom from debt, enthusiasm, hard work and active participation in community development. As of October 1997, the ACSI was functioning in 46 woredas with 15 branches, 67 sub-branches and has disbursed over 17 million Birr. Of the 27 thousand loans, nearly 80 percent were given for agricultural purposes. The institution has had a repayment rate of 99 percent. In addition, there are other private, informal credit systems that could be tapped. Traditional rotating savings and credit associations, the Ekubs, informal provident funds and the Idirs are common. Strong Performing Regional Governments and Domestic Borrowing 2.32 As Ethiopia continues to make progress on the social and economic development fronts, the regional governments - particularly the strong performers such as the City Administration of Addis Ababa, etc.- should be encouraged to borrow to invest in productive capital assets. World-wide experience indicates that maturing economies have successfully used this financing instrument (though with varying degrees of success). 2.33 The current thinking in Ethiopia seems to be that subnational borrowing is not desirable as it is known to create major problems in macroeconomic management. In particular, irresponsible subnational borrowing can often force federal government bail outs. These fears are not unfounded, as some recent cases like Brazil show (see Box 2.3). Tanzi (1995) cites among other instances the case of Sao Paulo, which has accumulated an unsustainable debt of around $40 billion. In India, unsustainable borrowing by state governments has also become a major cause for worry. While discussing the Latin American federations, Hommes (1995) emphatically states: "Of paramount importance are rules that limit the indebtedness of all levels of government, and in this case the lower levels. When borrowing by the subnational governments within a monetary union creates negative externalities, it must be restrained." 36 Chapter 2 2.34 Of course, with any entrepreneurial activity there is an element of risk, and borrowing by regional govebment would be no exception of the rule. What is needed is a system that ensures screening out of uneconomical and unsustainable debt being raised by the regions. Subnational defaults have to be permitted to exert discipline on lenders. If such a system can be designed, it will provide the necessary flexibility to the budgets of the subnational govercnents. Clearly, a system like the one in Canada (where provinces can borrow any amount from any source including foreign sources; the only constraint is when a federal guarantee is required) or even India (unlinmited borrowing powers of states within the country, with the Center's concurrence if they are indebted to the latter; extemal borrowings have to be routed through the center with compulsory central guarantee) may not be suitable to Ethiopia. Perhaps the former Australian model (where a national loan council allocates loans to the states after considering factors like need, absorption capacity and repayment capacity) could be adapted profitably keeping the particular requirements of Ethiopia in views. In order to avoid moral hazard problems and minimize the risks of a bail-out by the federal goverfnent, adequate regulations would need to be drafted and enforced. To this end, the borrowing amount and the debt service could be linked to the revenues of the regional governments. This will ensure that necessary preconditions are met before a region is able to borrow. 13 Australia changedth the nthe Loan Council in 1992 to one of an inforation exchange, which enables creditors to understand the risks entailed in such lending. 3. ROLE OF MUNICIPALITIES UNDER DECENTRALIZED DEVELOPMENT IN ETHIOPIA Atunicipalities fall outside the system of intragovernmental transfers, and must rely on self-generated revenues to fund the important services they provide. They are handicapped by a murky legal framework and an ambiguous position within regional g9vernment structures. An important dimension of the regionalization agenda is to ensure that municipalities provide urban services more effectively and contribute more broadly to economic growth. A. IMPORTANCE OF MUNICIPALITIES AND RECENT EFFORTS TO STRENGTHEN THEM T'he Urgent Need for Strengthening Municipalities 3.1 Ethiopia's ambitious fiscal decentralization program initiated by the 1994 Constitution did not explicitly deal with municipalities, a form of local government that was established in major urban areas and small towns in the 1940s. Since the legal recognition of municipalities at that time, there have been few laws or programs designed to enhance the functioning of these entities, with the exception of some Derg-era legislation intended to expand their role in the political arena. 3.2 The impact of this long-term benign neglect of municipalities may have been less apparent in the past. The majority of Ethiopians live in rural areas, and the municipalities have not historically been particularly important political entities. The situation is changing, however, and the consequences of continued neglect of municipalities are, for several interrelated reasons, potentially substantial. 3.3 In recent years, there has been a renewed recognition that rural development strategies can only have a limited impact without adequate urban development. Although there is presently little information on the economic significance and potential of urban centers in Ethiopia, theory and empirical evidence from other countries point to the critical role of urban centers in supporting rural development and broader economic growth through dynamic rural-urban interaction14 14 There is a substantial and diverse literature that focuses on the benefits of decentralization and its role in rural and urban development. See, for example: Hicks, U.K. (1961) Developmentfrom Below: Local Government and Finance in the Developing Countries of the Commonwealth. Oxford: Clarendon Press; Oyugi, W.O. (1978) Local Government and Development, Discussion Paper No. 131. Brighton: Institute of Development Studies, University of Sussex; Rondinelli, D.A,. (1981), "Govermment Decentralization in Comparative Perspective: Theory and Practice in Developing Countries," International Review of Administrative Sciences, Vol. 47, No. 2; Belsky, E., and G.J. 38 Chapter 3 3.4 Rural-urban growth policies are predicated on the fact that many commercial and specialized service activities in a region are typically based in urban settings. Even small urban centers often function as communications and transport centers for their regions, providing critical links to rural areas and serving as concentration points for marketing activities. Nurturing rural-urban linkages is thus a key component of a successful local economic development strategy. Municipalities are often uniquely well placed to understand this and to design supportive interventions. 3.5 The urban population in Ethiopia has recently been growing much more rapidly than the rural population (5.6 percent vs. 2.7 percent), and this disparity is projected to increase significantly in the next few years.'5 International experience suggests that more attention needs to be paid to urban areas when they are growing quickly in order to effectively meet expanding service demands (see below). 3.6 Effective management and development of smaller municipalities can also have an impact on national population policies and urbanization patterns by helping to curtail the excessive growth of population and economic pressure in primary urban centers. The process of continuing urbanization in smaller municipalities creates new local wealth and public resources through increases in economic activity, income, and land values, drawing still more people and employment to these areas. 3.7 Municipalities in Ethiopia have principal responsibility for important public services (detailed below) that could support private sector development and improve the welfare of urban residents. These services have been poorly provided in most municipalities. If municipalities cannot respond to growing local service demands that support private firms and meet citizens' needs, increased unemployment, declining living standards, and a higher incidence of poverty may result. Karaska (1982), Urban Functions in Rural Development (Worcester, MA: Rural Marketing Centers Project, Cooperative Agreement on Human Settlements and Natural Resource Systems Analysis (USAID), Clark University/Institute for Development Anthropology; Rondinelli, D.A. (1982), Small Towns in Developing Countries: Potential Centers of Growth, Transformation and Integration (Worcester, MA: Cooperative Agreement on Human Settlements and Natural Resource Systems Analysis (USAID), Clark University/Institute for Development Anthropology); Cheema, G.S., and D. Rondinelli (1983) Decentralization and Development: Policy Implementation in Developing Countries. Sage Publishing; Evans, H. (1986), Rural Trade and Production Centers: Towards a Strategy for Implementation (Nairobi, Kenya: Resource Management for Rural Development Project, Ministry of Planning and National Development, Republic of Kenya); Baker, J., ed. (1990), Small Town Africa: Studies in Rural- Urban Interaction (Uppsala: The Scandinavian Institute of African Studies); Lewis, B.D. (1991), "An Enquiry into Kenyan Small Town Development Policy," Economic Geography, Vol. 67, No. 2, pp. 147-153; Evans, H. (1992), "A Virtuous Cycle Model of Rural-Urban Development: Evidence from a Kenyan Small Town and its Hinterland," Journal of Development Studies, Vol. 28, No. 4, pp. 640-667; Gaile, G. (1992), "Improving Rural-Urban Linkages Through Small Town Market-Based Development," Third World Planning Review, Vol. 14, No. 2; Smoke, P. and H. Evans (1993), "Institutionalizing Decentralized Development Planning Under Kenya's Rural Trade and Production Centers," Third World Planning Review, Vol. 15, no. 3; Smoke, P. (1994), Local Government Finance in Developing Countries: The Case of Kenya (Oxford University Press); Evans, H. (1997), Guidelines for a Rural-Urban Linkage Approach to Poverty Alleviation," (United Nations Development Program). 15 Central Statistical Authority (1996), Ethiopia Statistical Abstract 1995, Addis Ababa. Role of Municipalities Under Decentralized Development in Ethiopia 39 3.8 It is also important to note that many of the services provided by municipalities generate important externalities across jurisdictions and involve complementarities across sectors for which different levels of government have responsibility. Municipal streets link up to roads provided by higher levels of government, forming part of an integrated road network. Improved water and sanitation services under municipal jurisdiction can help to reduce demands on health services delivered by the states. Better municipal nursery schools enhance the chances of success when students graduate to state-provided primary and higher level schools. Good service provision in municipalities can also benefit rural residents who travel to municipalities for services unavailable in their villages. Thus, effective municipal service delivery is an important component of the overall system of public responsibilities in both urban and rural areas. 3.9 Ethiopian municipalities have legal access to a number of important revenue sources (detailed below), but these have typically been poorly exploited, often due to centrally imposed restrictions on the base and rates charged on local revenues. Equally important, poor service delivery undermines incentives for local residents to pay taxes and charges due to municipalities. 3.10 Enhanced revenue generation would be possible in most cases if municipalities were given more revenue autonomy and basic services could be delivered more effectively. This would begin to build a base for greater responsiveness of municipal government officials to their constituents, enhance municipal cost recovery and fiscal discipline, and reduce inevitable future pressures on the central and state governments to provide substantial subsidization for the improved delivery of important municipal services. 3.11 Perhaps the most important reason for strengthening municipalities is that they are the institutions that are closest to the people at the urban level. Well-functioning municipalities can increase the level of local participation in the development process and, by definition, enhance the degree of self-government. Municipalities are also typically in a better position than the central government to tap into local private firms, community organizations, and NGOs to serve as partners in development. Such innovative strategies have been successful in many places in recent years. 3.12 Although redistribution is generally considered a central function, effective municipalities can help with poverty alleviation. Good service delivery can enhance economic growth, as discussed above. Municipalities are also in a better position than higher levels of government to identify and target the most disadvantaged segments of urban society, and they have a greater incentive to work to improve linkages between social programs provided by municipalities and those provided by higher levels of government. In addition, as noted above, some services provided by municipalities often benefit residents of rural areas, who tend to have lower incomes and to be less well served by public services then urban residents. 3.13 The reasons for strengthening municipalities outlined here are obviously highly interdependent. Improved local service delivery can foster dynamic rural-urban growth, improve the welfare of municipal residents, and help to create jobs and basic services for 40 Chapter 3 the urban poor. Enhanced local revenue generation can create a greater sense of municipal fiscal responsibility, create an efficiency link between services provided and revenues collected, and reduce fiscal pressures on higher levels of government. If municipal services and revenue generation can be enhanced in a context of improved accountability to local constituents and better municipal relationships with the private sector and NGOs, the potential develo,pment impact would be substantial. Positive Steps Towards Municipal Reform in Ethiopia 3.14 There is clearly increasing 'overnment awareness of the need to pay more attention to municipalities, and a number of positive steps towards reform have already been taken. First, the Federal Minist'- of Works and Urban Development (MWVUD) recently prepared draft legislation tu deefine criteria for municipal status. This legislation has been sent to the regional governments for review, and it will be revised and submitted to the Parliament as soon as possible. 3.15 Second, the MWUD has been supporting the development of projects through its Urban Development Support Services (UDSS) to enhance the financial and managerial capacity of municipalities through system development and on-the-job training. Both the World Bank and GTZ have provided substantial support for these efforts. This support is expected to continue and to increase in the future, and other donors may get involved as well. 3.16 Much of the initiative for enhancing municipal performance has come from regional governments, sometimes with support from UDSS, sometimes independently from the regional Bureau of Works and Urban Development. Some regions, such as Oromiya, have developed regulations to define the basic structure of decentralized authorities under their jurisdiction. including municipalities.'6 Others, such as Addis Ababa, have issued regulations governing particular sources of revenue.'7 A number of regions, including Amhara and Tigray, have also developed or are in the process of developing basic classification systems for municipalities. 3.17 In addition to steps already taken, several regions have been planning, sometimes with UDSS assistance, research activities to help them to define more fully the legal status, roles and responsibilities of municipalities in future state legislation. Oromiya for example, is drafting detailed regulations to standardize or set guidelines for municipal institutional relationships, service authority, tariff setting and management procedures. Other regulations are being prepared to definie urban development standards for physical plans, employment terms, budgets, etc. Amhaia is also about to begin a study designed to help develop a new legal status and organizational framework for municipalities, and Tigray is 16 Regional Proclamation No. 4 of 1993 To Establish Municipal Authorities and Rural Kebeles of the National/Regional Self-Government of Oromniya. 17 Region 14 Administration Regulations No. 3 oCf 1994 on Urban Lands Lease Holdings Regulations and Region 14 Administration Regulations No. 4 of 1994 on The Rent Holding and the Fixing of Rate of Rent of Urban Land. Role of Municipalities Under Decentralized Development in Ethiopia 41 in the process of examining its fiscal cadastre system, with the intent of improving its land tax administration and yield. 3.18 The development of regulations in some states and the research supporting them can provide lessons and examples for other states on regularizing the status of municipalities and providing them with a structure and procedures more conducive to good performance. Of course, much more needs to be done. Major reforms, however, do not happen overnight, especially in a country like Ethiopia, where municipal capacity is often weak, expectations of municipalities are low, and local resources are limited. The question of strategy, of how to move forward in such a difficult environment, is examnined in the last section of this chapter. Before that, we turn to a more detailed analysis of the history, current status and fiscal performance of Ethiopian municipalities. B. THE PRESENT MUNICIPAL SYSTEM AND PERFORMANCE CONSTRAINTS IJrbanization and Municipal Status 3.19 Ethiopia is among the least urbanized countries in Africa.'8 According to the 1994 C'ensus, there are 855 urban centers '9in the country, accounting for only 13.7 percent (8.2 million of 54.9 million) of Ethiopia's population.20 The majority of the urban centers are concentrated in four regions, Ororniya (373), Amhara (208), SNNP (149) and Tigray (74), while the most urbanized regions (in terms of population) are Addis Ababa (98 percent), Dire Dawa (66 percent)), Harari (58 percent), and Tigray, Oromiya, and Garnbella (each with 15 percent). 3.20 Addis Ababa counts as only one urban center, but its population exceeds two million and it has special status as a city government.2' Among the four regions with the largest number of urban centers, 372 of their 804 urban centers are classified as townships, 250 of which have officially recognized municipal status. Only 41 municipalities have a population greater than 20,000, and in just two cases (Bahir Dar in Amhara, which has been given special zonal status, and Nazreth in Oromiya) does the population exceed 100,000. Outside the four regions containing the majority of urban centers, only Dire Dawa, a special administrative zone, has an urban population greater than 100,000. Thus, in addition to being underurbanized in the aggregate, most of Ethiopia's urban centers directly support very small populations. 18 African Development Indicators, 1996. 19 An urban center is defined by the Central Statistical Authority as a settlement with: (a) a minimum population of 1000, at least 50 percent of whom are engaged in non-agricultural activities; (b) existing urban dwellers associations (UDAs), grassroots organizations defined by Proclamation No. 104 of 1976; and (c) status as a seat of government administration at least at the woreda level. 20 The total figure does not include the Somali Region. 21 This status was granted by Proclamation No. 87 of 1997 of the Federal Democratic Republic of Ethiopia. 42 Chapter 3 Legal Basis of Municipalities= 3.21 Most of the major legislation governing municipalities dates back to the Imperial Administration. A 1932 decree assigned a few basic services to Addis Ababa and provided for the taxation of urban land. Decree No. 1 of 1942 allowed for the creation of municipalities with appointed councils under the control of the Ministry of the Interior, and it also defined some basic municipal functions and revenue sources. Further elaboration of municipal functions and central control mechanisms was provided in Proclamation No. 74 of 1945. Legal notices pursuant to this proclamation were issued in 1971 to refine and update regulations on municipal taxes, fees and dues. These still serve today as the principal basis for municipal revenue generation. 3.22 The Derg issued three major proclamations that affected the organization and management of municipalities. Proclamation No. 47 of 1975 nationalized urban land, and Proclamation No. 4 of 1976 provided for the creation of various levels of urban dwellers associations that supported the Derg's philosophy of governance. Proclamation No. 206 of 1981 essentially turned over the administration of municipalities and urban centers to the urban dwellers associations, but this was done in the context of heavy centralized regulation and control. 3.23 After the fall of the Derg in 1991, the Transitional Government of Ethiopia embarked on a massive process of decentralization. The government structure framed in the 1994 Constitution focuses on the development of regional governments and woredas. There is no mention of the municipalities in the Constitution or in the main Government proclamations defining the structure and responsibilities of decentralized entities.23 The present Government has to date issued only three proclamations directly relevant to municipalities and urban areas. Proclamations No. 41 of 1991 and No. 4 of 1995 gave the Ministry of Works and Urban Development broad powers and duties for urban development and assigned key responsibilities for municipalities to regional governments. Most recently, Proclamation No. 87 of 1997 chartered Addis Ababa as a city government and defined in detail its organization and functions. 22 More detail is provided in: Sisay Ashenafi (1 996),"Legal and Institutional Aspects of Municipal Administration in Ethiopia," prepared for Urban Development Support Services, Ministry of Works and Urban Development, Addis Ababa. 23 Proclamation No. 7 of 1992 To Provide for the Establishment of NationallRegional Self-Governments; Proclamation No. 33of 1992 To Define the Sharing of Revenue Between the Central Government and the National/Regional Self-Governments; Proclamation No. 41 of 1993 To Define the Powers and Duties of Central and Regional Executive Organs of the Transitional Government of Ethiopia. Role of Municipalities Under Decentralized Development in Ethiopia 43 Institutional Framework in Which Municipalities Function24 3.24 The institutional framework in which municipalities currently function is complex, ambiguously defined at all levels of government, and inconsistent across regions. As might be expected from the legal framework summarized above, the role of municipalities in the decentralization efforts is particularly unclear. Even the Regional Affairs Department in the Office of the Prime Minister, which has substantial overall responsibility for the decentralization process, has no formal link to municipalities. 3 25 The Federal Ministry of Works and Urban Development (MWUD) formerly had substantial responsibilities for municipalities, but new regulations briefly outlined above assign them only a few major related tasks. They are expected to undertake studies related to broad urbanization patterns, to provide an environment conducive to urban development (including the development of appropriate training), and to define standards for the classification of urban centers. 3.26 The National Urban Planning Institute (NUPI), which is under MWUD, has been heavily involved in preparing urban physical development plans, and they are likely to continue to play some role in this area because of poor local and regional capacity. Another key MWUD player is the Urban Development Support Services (UDSS). With support from the World Bank (Market Towns Development Project) and GTZ (Urban Management and Capacity Enhancing Project), UDSS has been developing on-site training in general and financial management in selected municipalities in a few regions. In addition, they are now assisting, as noted earlier, with studies in a number of regions (Amhara, Oromiya and Tigray) to help clarify the status of municipalities and to undertake specific technical assistance tasks designed to improve municipal performance. 3.27 More substantial responsibility for municipalities is assigned to the regional government level. There are two key players at this level. First and most important are the regional councils, which have administrative responsibility for all government entities within their jurisdiction, including the municipalities. In practice, municipalities may be required to report to lower levels within the regional administration, i.e., the zone or the woreda, rather than to the regional council. 3.28 The second major actor at the regional level is the Bureau of Works and Urban Development (BWUD), which is directly responsible to the regional council and is intended to have subsidiary departments organized at the level of zone administration. These agencies are formally in charge of technical matters related to urban development, including in non-chartered municipalities. Although their technical and supervisory mandate is broad, covering preparation and implementation of physical plans, budget administration, revenue collection and personnel administration, they have in practice focused primarily on activities related to project design and implementation. In addition, 24 More detail is provided in Ashenafi (1996), and in Gebhard von Katte, Sisay Ashenafi and Shewaye Tesfaye (1996), "Report of the Reconnaissance Mission on the Approaches to and Problems of Urban Management and on the Legal and Institutional Framework of Municipalities," Urban Development Support Services, Ministry of Works and Urban Development, Addis Ababa. 44 Chapter 3 not all of the regions have zonal departments, and in a few cases, the regional BWUD have additional responsibilities for the mining and energy sectors. 3.29 At the local level, there are also important institutional issues. First, as noted above, many urban centers are not designated as municipalities, depriving them of municipal responsibilities and sources of revenue. Although there have been some standards for categorizing municipalities since MWUD issued them under the Derg in 1988, their application to ensure consistent treatment of municipalities in terms of responsibilities, structure and staffing has been inconsistent. Second, substantial variations in lines of upward accountability hinder the development of an integrated system of government. As indicated above, some municipalities report primarily to the regional BWUD, a technical agency, while others look primarily to one of the levels of regional adrninistration. Third, there is a lack of clarity in the relationship between elected municipal representatives and appointed members of the municipal staff. This can lead to disagreements and misunderstandings that slow down and/or damage municipal performance. Fourth, there is, at least in some cases, substantial interference in municipal affairs by the regional BWUD and/or higher levels of administration. This can occur in the form of payment authorizations, restrictions on the use of legally allowable municipal revenue sources, appropriation of municipal responsibilities and/or property, etc.25 Such interference not only undermines local autonomy, but also constrains the development of municipal capacity. Municipal Functions 3.30 The major responsibilities of municipalities were defined in Proclamation No. 74 of 1945 and Proclamation No. 206 of 1981, which were introduced above. Allowable municipal functions (subject to appropriate central approval) include: * Preparation of budget proposals * Assessment and collection of allowable municipal revenues * Preparation and implementation of development plans * Provision of internal roads and bridges * Provision of markets, slaughter houses, termninals, public gardens, recreational areas, and other public facilities * Regulation of cleanliness and provision of solid waste, water, sewerage and drainage services * Delivery of miscellaneous services, including fire protection, libraries, public toilets, street lighting, nursery schools, ambulance services, etc. 25 Shewaye Tesfaye and Tigist Negash (1995), "An Assessment of Municipal Financial Management in Southern Region," Financial Management Support Services, Urban Development Support Services, Ministry of Works and Urban Development, Addis Ababa. Role of Municipalities Under Decentralized Development in Ethiopia 45 3.31 Noticeably absent from this list are services that involve major interjurisdictional externalities or that have important distributional implications, such as health, education, major roads, etc. In fact, the assignment of functions to municipalities in Ethiopia appears to be generally appropriate both conceptually and relative to common international practice. 3.32 In reality, however, municipalities rarely cover this full range of functions or adequately deliver those services they do take responsibility for.26 There are four key constraints on their performance. First, anbiguities in the overall institutional structure and the weak status of municipalities discussed above allow higher levels of government to control some services and functions that municipalities are legally allowed to undertake, such as water. Second, the majority of municipalities have neither adequate funds nor enough human resource capacity to allow for even minimal service provision. Third, poorly developed and inconsistent organizational structures, managerial processes and procedures hinder effective municipal performance. Finally, a weak incentive structure, including poor employee remuneration (municipal salary scales are considerably lower than those applied to other civil servants) and a lack of accountability to municipal constituencies, seriously undermines service delivery.21 Municipal Revenue Sources 3.33 The basic revenue sources of municipalities are defined in Proclamation No. 74 of 1945. The main sources include: * Land rents and building taxes * Licenses on trades and professions * Market fees (for stalls and use of markets) * Fees for municipal services, including: sanitary services, slaughter houses, fire brigade services, mortuary and burial services, building plan approval, property registration and surveying, and use of municipal equipment, transport or employees 3.34 As in the case of assignment of functions, these are generally appropriate sources of revenue for municipalities, and some municipalities have received permission to levy various other types of local revenues as well, such as goods entry taxes (similar to the Indian octroi) and vehicle loading or unloading fees. Again, however, the reality falls short of the model. Revenue generation from municipal sources is highly inadequate relative to the fiscal potential of municipalities, and there is extreme variation in the 26 Shewaye Tesfaye (1997), "Performance and Financial Capacities of Municipalities in Ethiopia," prepared for a Workshop on Urban Management and Local Economic Development, Institute of Housing and Urban Development Studies, Rotterdam, pp. 10-14 and pp. 21-22. 27 Some interesting observations on these issues are provided in: Minas Hiruy (1997), "A Four City Rapid Appraisal of Governance and Management," prepared for the Urban Development Support Services, Ministry of Works and Urban Development, Addis Ababa. 46 Chapter 3 relative importance of these revenue sources across municipalities that seems to be unrelated to differences in their relative resource bases.28 3.35 There are at least five key reasons for this poor municipal revenue performance. First, revenue administration and financial management procedures are poorly developed and unevenly applied, such that municipalities may not even know the full extent of their revenue base or the extent to which they have collected the portion they are aware of Second, the same weak human resource capacity and poor incentives that plague service delivery also constrain effective revenue administration. Third, the rates set by the central government for most municipal taxes, fees and charges are obsolete, some dating back as far as the 1971 legal notices that updated the revenue provisions of Proclamation No. 74 of 1945.29 Fourth, municipalities have only very weak legal mechanisms to support enforcement of revenue collection. Finally, poor or nonexistent service delivery undermines the willingness of residents to pay any taxes or fees to their municipalities. 3.36 In addition to the weakness of their own-source revenue provision, municipalities receive no regular resources from higher levels of government. There is no tax sharing program, and the system of transfers that flows from the central govemment to regional governments does not benefit municipalities. In addition, municipalities have no access to credit from domestic financial institutions. The few resources municipalities occasionally receive from higher levels of government, usually for particular capital projects, are distributed on an ad hoc basis according to arbitrary and inconsistent criteria. 30 Municipal Budgeting 3.37 Although basic responsibilities for budget preparation clearly lie with municipalities, the system under which budgets are prepared and reviewed is not standardized. In the absence of formal guidelines for budget preparation and approval, there is wide variation in practice, even across municipalities even in the same region. No formal studies of the municipal budgeting process have been undertaken to date, but some anecdotal evidence illustrates the situation. 3.38 In some cases, such as Wolisso municipality (Oromiya Region), the budgeting process starts with a determination of expected available revenues. 31 Each municipal department submits a budget request to the accounting department, which assembles the requests and proposes a budget, in consultation with the departments, that falls within the budget constraint. This is discussed with the municipal council, presented in a public meeting, and after final amendments, is subrnitted to the zonal department for approval. 28 Tesfaye (1997), pp. 17-21; Excellence Management Consultants (1997), "Municipal Tariff Study," prepared with the Amhara Bureau of Works and Urban Development with GTZ assistance, Bahir Dar. 29 There are a few more recent pieces of relevant legislation, such as Proclamation No. 80 of 1993 To Provide for the Lease Holding of Urban Lands. 3 Tesfaye (1997), p. 19. 3' The examples provided here are based on fieldwork conducted by the Bank mission that prepared this document. Role of Municipalities Under Decentralized Development in Ethiopia 47 3.39 In contrast, other municipalities, such as Nazreth (Oromiya Region), each department prepares a budget request independently. These requests are discussed in a large management meeting of department heads and the mayor. After this meeting, the planning department prepares a full budget. Until recently, the budget was submitted for review to the relevant sectoral departments at the zonal level. After input from the zonal departments, the zonal administration reviewed the budget, made amendments, and returned the final budget to the municipality. Under a new regional proclamation, the municipal budget will now be reviewed by the woreda administration. Given technical capacity constraints, however, the woreda is likely to have to continue consulting zonal departments during the municipal budget review process. There does not seem to be any public meeting to review the budget, as there is in Wolisso. 3.40 Discussions with various people involved with municipalities suggest that the types of variations illustrated by these examples are common. Different municipal departments take responsibility for budgeting in different cases. The process through which budgets are prepared has different starting points (the revenue side vs. the expenditure side), involves the municipal departments and mayor/council members at different stages, and relies on various levels of public review. In addition, different levels of regional administration review and approve budgets in different cases. Some variations in the process may eventually be justified on clear grounds, but they are at present largely based on ad hoc decisions and historical accident. Governance :3.41 Two types of governance are critical to the effective performance of municipalities. The first involves accountability to higher levels of government, while the second involves accountability to municipal residents. As discussed above, municipalities do have to answer to higher levels of government, but the institutional structure in which this happens is fragmented and inconsistent. Equally important, there are few guidelines and standards to ensure that there is a reasonable basis for municipalities to make decisions about revenue, expenditure and employment, or for higher levels of government to review and approve or require amendment of these municipal decisions. Those guidelines that do exist, such as the tariff guidelines discussed above, are often many years out of date. 3.42 The second type of governance involves accountability to municipal constituents. The nature and extent of local accountability is generally weak, but this varies considerably across and within regions.32 A legal basis for various types of regional and local councils has been defined in various proclamations mentioned above. The councils of the regional governments and their lower levels are selected and function in a fairly standardized manner, but the practice for municipalities varies considerably. In general, municipal council members are directly elected by popular vote at the kebele level, but mayors are selected in different ways, and some municipalities use more inclusive and democratic procedures than others. 32 Von Katte, Ashenafe and Tesfaye (1996); Tesfaye (1997); Hiruy (1997). 48 Chapter 3 3.43 At one unusual extreme is Wolisso, a town of about 32,000 that has a council of 700 members with a term of five years. Each of the seven kebeles in the municipality elects 100 representatives, and the council members in turn elect the mayor, the vice- mayor, and an executive committee. As noted above, this municipal council is significantly involved in municipal budgeting decisions, although the council is relatively weak in comparison with the smaller woreda council, which also has representatives from the kebeles. 3.44 Most municipalities, even those with larger populations, have much smaller councils than Wolisso and much less participation in municipal decisions. In some cases, the mayors are appointed by higher levels of administration rather than elected by council members. These realities, combined with the historical tendency of municipal officials to take direction from higher levels of government, point to the need for substantial development of greater accountability to municipal constituents. This would, however, require clarification of the institutional ambiguities noted above and concurrent development of improved managerial competence at the municipal level. Summary of Key Concerns Regarding Ethiopian Municipalities 3.45 Municipalities in Ethiopia suffer from a number of serious problems. First and foremost, they operate. without clear legal status in the context of the broader decentralized public sector structure being developed by the Government. This results in arbitrary and inconsistent treatment by higher levels of government, which have to date had few clear standards to govern their dealings with municipalities. Municipalities are also plagued by weak financial and human resource capacity that constrains their ability to provide basic urban services. Local organizational structures, job descriptions, and managerial processes and procedures are poorly defined and inadequate, and the degree of accountability to local constituents is generally weak. 3.46 Collectively, these problems hinder municipalities from raising desperately needed resources, and they also undermine the ability of municipalities to provide important public services, which have the potential to stimulate dynamic rural-urban growth, to support private sector development and the creation of jobs for a rapidly growing urban population, and to enhance the well being of urban residents, among whom the ranks of the poor appear to be rapidly growing. C. MOVING FORWARD: A PROCESS FOR STRENGTHENING MUNICIPALITIES The Need for Deeper Municipal Reform 3.47 As noted earlier in this chapter, the Government has already taken some steps to exploit the potential for accelerating municipal development, but these efforts have been sporadic, limited in geographic focus, and not very closely coordinated. The discussion in the previous section suggests a more comprehensive and integrated set of reforms will Role of Municipalities Under Decentralized Development in Ethiopia 49 need to be pursued to improve the performance and to increase the development impact of municipalities. Key suggested reforms to consider would include: * Clarifying of the legal status of municipalities, including how they fit into the broader decentralization strategy being pursued by the govemment and how they should relate to higher levels of regional administration * Developing and applying objective criteria for the creation and classification of municipalities as a basis for more clearly defining their detailed roles and responsibilities * Creating guidelines for the organizational structure, job responsibilities, revenue bases, and managerial process and procedures to be used by various types of municipalities * Developing mechanisms to revise municipal tariffs and salary scales on a regular basis * Enhancing accountability to municipal residents through reforms to standardize the electoral process and opening channels for greater citizen information on and participation in municipal decision making * Enhancing the capacity of municipal councils and staff to undertake their responsibilities more effectively * Raising awareness of urban and rural economic development issues and the role of municipalities in this regard * Developing creative partnerships with private sector firms, community groups, and NGOs. 3.48 Obviously such a massive reform agenda cannot be defined or implemented rapidly. In addition, before substantial progress can be made, more detailed and accurate information is required on current municipal performance and the potential for municipalities to enhance local economic development and rural-urban linkages in the Ethiopia case. In order to move forward, a gradual and pragmatic strategy for moving forward needs to be developed, as discussed in more detail below. Some Examples of Municipal Policy and Procedural Reform in Developing Countries 3.49 Many countries have researched and experimented with decentralization policies, municipal capacity building programs, and rural-urban linkage strategies in recent years. Some brief examples will set the stage for proposing a process for moving forward with such reforms in Ethiopia. The examples are categorized into three types: limited focus municipal reforms; strategic overall approaches to decentralization and local government reform; and, programs to stimulate rural-urban linkages and local economic development. 50 Chapter 3 3.50 Successful Limited Focus Municpal Reform. A number of cases of successful limited-focus municipal reforms are emerging from international experience. A few types of examples are described here for illustrative purposes. Most of these include options for relying on citizen consultation groups and/or private sector firms arid/or NGOs as advisors and/or local service partners. 3.51 The city of Pune in India, with assistance from the Center for Development Studies and Activities, recently created a multi-actor local consortium to help identify and implement changes in local management, particularly with respect to enhancing the quality and use of local input in decision making. This consortium includes representatives from the local government administration, local business groups, local media, local public interest groups, and local academic institutions, each of which supplies particular types of expertise. Although this is a new innovation, those involved have high expectations for the role the consortium can play in defining and implementing reforms.33 3.52 In a similar, but service-specific case, the city of Santo Andre in Brazil has been experimenting with a citizen advisory group. The Association of Public Transport Users was created by a group of concerned citizens in the wake of a municipal electoral victory by an opposition party that replaced a long-entrenched and insulated city government. In response to substantial pressures from this association, changes were made in the way city officials solicited citizen input on transportation services, and several improvements in service quality resulted.34 3.53 In a number of countries, there have been successful experiments with using urban community groups to assist directly in service provision. In Jakarta, Indonesia, for example, the city government has used various partnership systems to involve neighborhood groups in trash collection services. These groups carry solid waste to a central collection point for transport by the city government to a dumping area. The groups also collect user fees, a portion of which they can keep for their own community activities as an incentive to do a good job. The details of their responsibilities are specified in enforceable contracts between the city government and the community groups, which also helps to ensure good performance.35 Similarly, one city in Cote d'Ivoire recently improved municipal sanitation, in part using community groups to provide certain types of 33 Nadkanii, A. (1995) "Participate (Pune): Multi-Actor Collaboration in Developing Communities in Urban India" in Wilson, R. H. and R. Cramer (1995), International Workshop on Good Local Government. Austin, 'IX: Lyndon B. Johnson School of Public Affairs, University of Texas, pp. 45-52. "4 Morales, C.A. (1995) "Santo Andre: Innovation in Transport Manageinent in a Medium-Sized City" in Wilson and Cramer, pp. 23-29. 35 Johnson, J. (1992) "Contracting Out, Privatization and Community Organizations in the Solid Waste Sector in Indonesia," unpublished masters thesis. Cambridge, MA: Department of Urban Studies and Planning, Massachusetts Institute of Technology. Role of Municipalities Under Decentralized Development in Ethiopia 51 services and to put pressure on municipal employees and residents to keep the city cleaner.36 3.54 Some municipalities in Uganda have been experimenting in recent years with the private-public partnerships in the management of municipal markets. Private firms are awarded contracts under competitive bidding to manage markets, and they develop relationships with market user associations to assist with both market fee collection and certain types of maintenance of the facilities. In most cases where this model has been followed, market revenues have dramatically increased, to the benefit of both the municipalities and the private management firms. A recent study suggested that at least in some cases, market users - both buyers and sellers - are substantially happier with the operation of the market facilities since the new arrangements were adopted.37 3.55 There have been quite a few cases of local governments in Africa tapping user charges and sales proceeds from certain commodities as a source of revenue, sometimes in partnership with private firms.32 In South Africa, municipalities use substantial service user charge surpluses from water and electricity services provided by municipal corporations as a general source of revenue. In Nigeria and Kenya, market and bus park fees are a significant source of revenue that helps to subsidize non-revenue-generating services. In Zimbabwe, municipalities have a monopoly over the sale of alcoholic beverages and tobacco, which accounts for a large percentage of their own-source revenues. Market-oriented economists may see some of these types of revenue innovations as undesirable, but they do contribute substantially to the budgets of municipalities that have access to few other productive sources of revenue. 3.56 The number of examples of these types of phenomena reported in the development literature to date is relatively small. There is, however, enough evidence to suggest that it is possible for local governments to behave innovatively and to develop constructive relationships between local governments and private partners (including NGOs) that improve certain aspects of local service delivery and revenue generation. Such approaches merit further consideration as Ethiopia considers its own options for municipal reform. 3.57 Strategic Overall Approaches to Decentralization and Local Government Reform. Many decentralization and municipal government reform programs in developing countries have been not been particularly successful. There are a number of reasons for this. First, they are often based on western conceptual frameworks that may be inappropriate or incorrectly applied. This does not mean that these frameworks cannot be 36 Galectic, A. (1998), "How One City in Cote d'Ivoire Successfully Tackled Its Sanitation Problem: Building Popular Support and an Incremental Approach," unpublished masters thesis. Cambridge, MA: Department of Urban Studies and Planning, Massachusetts Institute of Technology. 37 Birungu, H. et.al. (1998), "What Makes Markets Tick? Local Governance and Service Delivery in Uganda," a study prepared for the Management Governance and Development Division of UNDP by the Institute of Social Research, Makerere University, Kampala, Uganda. 38 Olowu, D. and P. Smoke (1992), "Determinants of Success in African Local Governments," Public Administration and Development, Vol. 12, No. 1. 52 Chapter 3 used as a starting point, but they must be interpreted in the carefully assessed context of a particular country. Second, they are typically inattentive to the complex institutional realities that logically govern the extent and pace at which decentralization and municipal reform can occur in a particular country. Attitudes about the way the world works do not change as quickly as reformers might like. Third, these efforts are usually far too comprehensive or far too limited in scope. Comprehensive programs overwhelm capacity and political tolerance, often leading to failure. Limited programs, unless carefully structured with vision, may not have a strong effect, even if, as in the case of a few examples cited above, they work well. Finally, many decentralization and municipal reform programs have been driven by uncoordinated and competing donors, often with much external technical assistance but few clear incentives to encourage capacity building. 3.58 As these types of problems are increasingly recognized, a number of countries are beginning to take alternative approaches to municipal reform that are comprehensive but strategically and gradually implemented. One interesting case is Kenya's Small Towns Development Program (STDP), a capacity building project for small urban governments jointly funded by the Kenyan and German governments (GTZ) and based in the Ministry of Local Government (MLG). Under this project, there have been some broad common reform goals among the local authorities participating in the program, but the exact steps each took and the pace at which they took them differed. 39 3.59 The strategy in this project was to employ a mobile team of experts to work individually with each participating local government, often with private sector support, to design a unique package of realistic reforms that best met their needs and capacities. In addition, a mobile technical assistance team worked periodically with the local governments to implement the negotiated program of reforms gradually, and, as necessary, to modify it. Local citizen participation mechanisms were required as part of the reform process, and some training opportunities were provided for local officials. Local governments that met the goals they agreed on were rewarded in various ways, such as greater access to funding and reduced MLG interference in their decisions. In most of the participating local governments, their performance - in terms of revenue collection and service delivery - has improved considerably. 3.60 The STDP has also had several other benefits. First, virtually all of the technical assistance personnel have been local rather than foreign consultants, raising acceptance and the credibility of the initiative. Second, STDP efforts have generated information that has assisted the MLG to create a variety of procedural manuals and training materials for different types of local governments. These have helped to get central and local officials to accept the notion of using standards and models in making basic revenue and expenditure decisions. Third, and perhaps most importantly, the continual presence of the STDP over a period of years has raised the visibility of local governments among central officials and the general public and demonstrated that, with adequate assistance and 3' The Kenya experience is examined in: Ministry of Local Government (1993), "Review of the Small Towns Development Project," Nairobi; Eigel, U. (1994), "Technical Assistance to Small and Medium-Sized Towns: A Practical Example from Kenya." The Courier (GTZ review). Role of Municipalities Under Decentralized Development in Ethiopia 53 incentives, they can improve their long-poor and long-maligned performance. Such proven success, even if initially limited, is a critical ingredient of any decentralization and municipal reform program. 3.61 The Kenyan government, with very substantial World Bank funding and technical assistance, is now defining a broader program of reform - the Kenya Local Government Reform Program (KLGRP) - largely based on the successful experiences of the STDP and lessons learned from less successful past donor efforts. Following the lead of STDP, the KLGRP is being built gradually and systematically in a series of manageable, mutually reinforcing, cumulative steps. There has also been an explicit realization that the close cooperation of a variety of government agencies is required, so the Task Force that provides policy guidance on KLGRP is interministerial. In addition, most of the KLGRP staff members are regular government employees who have continued to perform many of their regular duties in addition to the KLGRP responsibilities. This creates a situation in which KLGRP is not seen as separate from government activities, but an integral and mutually reinforcing part of them. 3.62 This Kenya program, though still to be fully proven, strongly suggests that a gradual approach and individually tailored packages of reform, supported by carefully targeted technical assistance, can have substantial benefits. This approach also recognizes that decentralization involves slowly changing fundamental attitudes about what central and local governments should do and building their capacity to do it. As Ethiopia considers possibilities for designing a municipal capacity building program, some valuable lessons might be drawn from the Kenya experience. 3.63 Programs to Stimulate Rural-Urban Linkages and Economic Development. Yet another class of reform programs goes beyond simply trying to make municipal governments perform better in revenue generation and service delivery by trying to, directly support private sector development. Such programs are often based on the rural- urban linkage model briefly described at the beginning of this chapter. From the late 1980s until the mid-1990s, for example, Kenya experimented with an innovative Rural Trade and Production Center (RTPC) program.40 Under this program, the Government of Kenya (GOK), with technical assistance from USAID, identified underexploited rural- urban systems throughout the country. These consisted of small urban centers and their economic hinterlands. A coalition of central government/local government/private sector actors prepared plans for stimulating dynamic rural-urban growth in these RTPCs, focusing on the development of institutional and physical infrastructure. 40 Lewis, B. D. (1991), "An Enquiry into Kenyan Small Town Development Policy," Economic Geography, Vol. 67, No. 2, pp 147-153; Evans, H. (1992), "A Virtuous Cycle Model of Rural-Urban Development: Evidence from a Kenyan Small Town and its Hinterland," Journal of Development Studies, Vol. 28, No. 4, pp. 640-667; Gaile, G. (1992), "Improving Rural-Urban Linkages Through Small Town Market-Based Development," Third World Planning Review, Vol. 14, No. 2; Smoke, P. and H. Evans (1993), "istitutionalizing Decentralized Development Planning Under Kenya's Rural Trade and Production Centers," Third World Planning Review, Vol. 15, no. 3. 54 Chapter 3 3.64 Some of the RTPCs were chosen for political reasons, so not all of them took off. Several of them, however, provided substantial benefits in the form of new infrastructure and improved local economic performance. The biggest problems with the scheme were a limited role for the private sector and a tendency, at least initially, to focus on physical infrastructure and to pay inadequate attention to important institutional reforms and support, such as enhanced credit access and training. 3.65 One particularly positive example from the RTPC program occurred in Kutus, an urban center in the rich agricultural land surrounding Mount Kenya. Kutus had an active agricultural market before the RTPC interventions, but the market was informally organized, and links to some of the surrounding fertile areas were poor. The institution that designed the RTPC interventions, the District Development Committee, which includes representatives from the central government, the local authorities, and the private sector, designed and implemented a package of investments to relieve the bottlenecks on growth. A new market and a new bus park were built, concentrating transport and marketing activities in good facilities, and both had the added bonus of greatly increasing local government revenue collections. In addition, access roads and a bridge were built to improve all-weather access to the surrounding rural areas, and the small local farmers were assisted with organizing and securing credit and technical assistance. There have been problems in the Kutus RTPC, and the overall program has floundered recently due to political and economic strife in Kenya, but it has been successful enough to justify considering the use of similar strategies elsewhere. 3.66 Even more ambitious rural-urban development schemes, including components that directly strengthen municipal governments, are coming into the policy limelight. For example, a new UNDP/UNCHS funded program in Indonesia, Poverty Alleviation through Rural-Urban Linkages (PARUL), taps government programs to provide supportive urban infrastructure and forges public-private partnerships in strategic sectors to stimulate economic development and export growth.41 The program is too new to evaluate, but there are some interesting possibilities emerging. In one of the PARUL pilot provinces, for example, the program is focusing on tapping into nascent public and private efforts to stimulate the cashew nut industry. 3.67 The Governor of South Sulawesi Province last year initiated a program to double agricultural production by the year 2000. The Governor's program created a number of provincial level committees with responsibility for targeting particular sectors or subsectors. There are also general (non-commodity or sector specific) committees established at the local government level. The structure is a bit top heavy and bureaucratic, but rather than bypassing it, the PARUL program plans to tap and develop it. The Director of the Provincial Government Planning Board has agreed to a PARUL proposal to create a cashew nut sub-committee of the provincial committee that deals with tree crops. This committee includes a small number of representatives from the public sector, processing firms, traders and farmers groups involved in the cashew nut industry, 41 Evans, H. (1997), Guidelines for a Rural-Urban Linkage Approach to Poverty Alleviation," (United Nations Development Program). Role of Municipalities Under Decentralized Development in Ethiopia 55 and it is to be chaired by the private sector. The intention is for the committee to connect with initiatives currently being developed by the private sector. The Indonesian Association of Cashew Nut Producers, for example, is considering a proposal to require all of its members to contribute a small amount of the price of each kilogram of cashews towards a local industry development fund. This fund would be used, for example, to help small farmers to increase productivity and to contribute to public sector efforts to develop improved infrastructure. 3.68 There is no question that using local governments and firms as part of a coalition to stimulate economic development is a difficult and time-consuming process. But this type of creative multi-actor win-win approach offers great promise, both to revitalize the performance of municipal governments and to bring about improved economic development performance. The approach taps private sector initiative, while building a base of support within present government structures to increase the possibilities for accessing whatever public funding and technical assistance might be available. The approach may also pave the way for increased government trust of the private sector to help design services and activities that provide broader economic and social benefits to municipalities and their surrounding regions. A Strategic Approach 3.69 It would be premature to recommend a specific program of municipal reform for Ethiopia, but the examples discussed above provide a basis for the Government to consider the types of reforms it may wish to pursue. Such reforms, however, should be designed in a way that ensures broad acceptance by beneficiaries and that provides a strategy for implementation. As noted above, experience from other countries tells us that many well-researched and normatively desirable reforms are never realized because reforms are undertaken without building adequate support from major interested parties with different points of view and objectives. In addition, attempts are often made to do too much too quickly-in ways that create conflicts among concerned agencies at the central level, overwhelm local capacity, and fail to take account of political tolerance for reforms. Local fiscal reforms should not be seen purely as deliverable outputs; rather, they should in most cases be part of a process that builds local technical and managerial capacity and stimulates local accountability as they are implemented. 3.70 With this view of the world in mind, the Government might consider how to move forward through a process for enhancing the performance and impact of municipalities that incorporates at least some of the following key elements: First, recognizing that developing a broader reform program will take time to define and implement, some efforts could be made to move forward as quickly as possible with reforms that are non-contentious, easy to implement, or unambiguously required. For example, in the Ethiopian case, there seems to be general agreement that the land rent and building tax will remain a major source of municipal revenue and is a priority target for improved performance in many municipalities. Thus, while there may be some contentious issues to 56 Chapter 3 resolve before concrete reforms could be put into place, this could be selected as one of the initial priority reform areas. * Second, the Government should consider the possibility and options for directly assisting local economic development as part of municipal reforn programs. A municipality cannot drive private investment activities, but it can work directly and closely with the private sector, as in the Kenya and Indonesia cases cited above. This implies the need to explore options for government partnerships with the private sector and NGOs, both for municipal revenue generation and service delivery, as well as for programs targeted more directly at private productive activities. * Third, any serious reform process should consult relevant actors in a way that highlights differences in their perspectives, begins to build some degree of consensus, and allows for strategic compromise in defining and implementing reforms, as in the Kenya case discussed above. It is better to choose a "second-best" reform program that moves the system in the right direction over time than to attempt to undertake in the short term a "normatively superior" set of reforms that is likely to meet crippling opposition during its implementation. This is important whether developing an overall system of reform or specific subcomponents, such as property tax reform or determining the extent to which and the pace at which a particular service ought to be decentralized to municipalities. It is important, however, to note that "negotiated" reform processes do not have to be arbitrary and subjective. On the contrary, significantly transparent processes with objective criteria are required to create a sense of fairness, so that municipalities understand the reasons why they are being treated in certain ways. * Fourth, there should be appropriately differentiated treatment of municipalities, sectors and revenue sources. Some local governments have more capacity than others, some sectors are more complex to decentralize than others, and some revenue sources are more suitable for certain places and uses than for others. Explicitly recognizing these differences allows reforms to be targeted in ways that maximize the probability that the reforms will be successful. In addition, transparent criteria for differentiation, as suggested above, make clear to local governments what they have to do in order to be treated more favorably (i.e., given more responsibilities and/or resources) in the future. * Fifth, the sequencing of municipal fiscal reforms should be gradual and strategic. This is important for a number of reasons. Moving step by step allows for simple successes that build a basis for further achievements and enhances the credibility of reform efforts. In addition, careful sequencing provides a mechanism for integrating components of a reform program, e.g., such that the adoption of new financial procedures or local sources of revenue is tied to possible grant or loan program reforms. Reform bundling and strategic sequencing also create an opportunity for changing behavior and Role of Municipalities Under Decentralized Development in Ethiopia 57 building capacity. For example, a reform process might require the adoption of new municipal financial management procedures in return for receiving training or a project grant from the state government. This should help to institutionalize important new skills, and future reforms can build on the modest ones that are put into place at earlier stages. Similarly, requiring even a weak municipality to borrow for a small part of an infrastructure project begins to create a mentality of fiscal responsibility that can help to build up a history of creditworthiness over time. Sixth, it is critical to develop and institutionalize a system and procedures for monitoring reform progress, probably at the state level in the Ethiopia case. Only if this is possible can the government be sure that the goals of municipal reform programs are being met. This also provides a mechanism for ensuring that municipalities ready to move on to the next reform step can be identified and fairly treated. 3.71 Moving forward with a reform process requires assigning some government staff or hiring a consultant to map more fully the current municipal position in the intergovernmental fiscal landscape and to consider options for public-private partnerships for local economic development more fully. This would include, for example: • To catalogue more thoroughly the existing formal fiscal responsibilities of municipalities and the extent to which they are actually being undertaken. * To identify more fully the fiscal (service and revenue) interactions between relevant central government departments and municipal governments, as well as between states/woredas and municipal governments. * To document the current and potential role of municipalities in stimulating economic growth in both urban and rural areas. * To take stock of all existing policies, regulations and programs related to fiscal decentralization and the strengthening of municipal service functions and revenue sources. * To understand the major differences in perspective on municipal government roles and responsibilities among key central, state and local government actors. * To identify and classify (in terms of purpose and scope) major government and donor initiatives designed to build fiscal, technical, managerial and political capacity in municipal governments. 3.72 With this more fully developed base of knowledge, the Government of Ethiopia could begin to think more systematically and strategically about how to include municipalities in the broader ongoing process of decentralization and intergovernmental fiscal reform. 4. REGIONALIZATION AND GENDER ISSUES A. INTRODUCTION AND BACKGROUND Despite their contribution to the country's economy, in both the agricultural and the informnal sector, and the importance of their social role, women in Ethiopia do not have equitable access to economic resources. As the responsibility for the design of development programs and sound laws and regulations which shape the role of women in society shifts to local governments, the opportunities for more gender sensitive development patterns will expand Fulfilling this potential will require the formulation of cohesive programs to strengthen the voice of women in society. 4.1 The gender dimension of poverty in Ethiopia manifested in the low level of women's access to education and health services, employment opportunities, and participation in the social and political life of the country, are reflected in the welfare indicators provided in Table 4.1. These indicators are also validated by the participatory rural appraisals (PRAs) that have been conducted in most regions in Ethiopia.42 Table 4.1. The Gender Dimension of Poverty in Ethiopia Indicators Ethiopia Sub-Saharan Africa Literacy Rate 35 62 Aggregate Female Enrollment* 25 39 Female Life Expectancy at Birth 51 53 Fertility Rate 7 5.7 Contraceptive Prevalence Rate 4 15 Maternal Mortality Rate 1,528 700 Infant Mortality Rate 120 104 Child Malnutrition (stunting) 60 39 Pop. per Primary Health Care 19,171 8,046 % of Female Candidates at The Last 2.1 na National Election (1995) Access to Safe Water 18 37 Pop. Below the Poverty Line** 34 na * Defined as % of male enrollment. ** The poverty line is set at US$1 per day. Source: Data From World Development Report 1997, Human Development Report 1995 and Internal World Bank Documents. 42 Findings from PRAs conducted in 1997 for the Goveunment/Bank WID) report. 60 Chapter 4 4.2 There is no doubt that investment in women needs to be increased in Ethiopia; it is critical for poverty reduction. Moreover, setting aside equity issues, it is largely accepted that such investment produces significant social returns. Moreover, there returns are actually greater than returns on similar investments in men and this is unlikely to be different in Ethiopia. However, effective and efficient investing in women is an extremely complex task. Women face greater economic and regulatory hurdles in relation to men, and are also burdened by significant cultural and social barriers that must be crossed or addressed. The complexity is exacerbated given that Ethiopia is a large country with more than seventy ethnic and socioeconomic groups. While the constraints facing women are relatively similar (lack of access to economic resources, education, health, services, credit, information, markets, etc), the PRAs verify that these problems cannot be addressed through similar programs or delivery mechanisms. Program content and delivery methods of development interventions need to be sensitive to the traditions and customs of beneficiaries if they are to be effective. Also, it is clear that addressing gender issues cannot be the responsibility of a single sector or agency. Actions have to be jointly planned and coordinated so that there is a holistic tackling of the myriad problems facing this vulnerable group. 4.3 Another important, but often ignored, fact is that if development of women is to be sustainable, it must come at the initiation of both men and women. People at local levels, both men and women, value a way of life where the husband or a male member is regarded as the head of the household - an attitude that is steeped in the religious beliefs and codified in the law. If the demand for change is to come from the proper constituency, women must be strengthened to organize and identify their priorities and interests, empowered to bargain with other players, furnished opportunities to increase their economic assets, and provided information on their rights and obligations. External assistance should facilitate these processes and provide opportunities for women to attain self-sufficiency and reduce their dependency on other household members, necessary conditions for long-term development. 4.4 The National Policy for Women, 1993 (Women's Policy), that was formulated concurrently with the regionalization initiatives acknowledges that neglecting the concerns of women have resulted in social and economic costs, and underscores the need to move towards a more institutionalized approach through enhanced policy measures. The Policy articulates a comprehensive strategy to institutionalize the socioeconomic and political rights of women and promote gender-sensitive policies and interventions (see Box 4. 1). 4.5 A broad institutional framework for implementation is also envisaged (see Chart 4.1). At the federal level, a Women's Affairs Office (WAO), under the Prime Minister's Office has been established with the task of coordinating and monitoring the implementation of the Women's Policy. WAO is also responsible for coordinating operations of the donors and NGOs. Women's units or departments have also been established in eleven key federal ministries, including Ministry of Economic Development and Cooperation, Agriculture, Labor and Social Affairs, Health and Education. Regionalization and Gender Issues 61 'g-.'^ > "'-... 9 + ' ' ' ' '' W~~~~~~~~~~~~~~~~~~~~~~~~....... ...w ' ' . ? . ..... , Box~wi U..VyE nt~ h o~e' ok ,-'?a~~~~~~~~~~~. i. '. ..?r¢ .. J- 01 ' ' -X ... , ,,g,.,, '., ? ,,,, E .,; ................... , -: . > . - '....... ... ..:.... ........ ........'........ . ..... re-ordering ofit p~~~~~~10:ii: ritiesM .......... ... ...ain ~m~ 1a Char)'t 4.1: n sititutional Arandegemaioentsse for thle e ntat anf th e Ntoa oiyo oe t.-Y ou~~~~~rgiom . ..... V* ivi4veme t iA al dv me r i Chart 4.1: Institutional Arrangements for Implementation of the National Policy on Women I , , ~ ~~ ~ ~~~~~~~~~~~I , . vedea ornment Executiwl Orpans ------ ~~~ Wme' Affairs omc I Stat Concik I I I Zonal WA Bureau, =Zoal E I I I I | |ZnlW ucu ZnlEeuirCmite I I L_ __… _ _ _ __ I |Woreda WA Bureaus [Wredu Executive Commitees A~~~~~untabilitv ~~~~~~~~~~~~~~~~~~L ---- Grassroot orp-anizaons lKebeleDevelop. C=-mft l Accountability -_--____ ConsultationlCollaboration 4.6 In line with the regionalization policy, the Women's Affairs Bureaus (WABs), and the Women's Affairs Departments in sectoral bureaus are the key players in the implementation of the Women's Policy at the regional levels. They are expected to contribute to the gender sensitiveness of the regional development plan, as well as act as promoters, catalysts, and watchdogs for women's issues concerning programs and projects in the various sectors. 43 43 WABs (Part B, Section 3 of the Women's Policy), are responsible for: (i) creating awareness on the organizational structure and the objectives of the Women's Policy; (ii) coordinating the women's focal points within the regions, as well as facilitating governmental and non-governmental activities related to women; (iii) encouraging research and other activities that may be required to raise the visibility on women's issues in the society and formulating policy proposals at the regional level; (iv) making efforts to increase women's active participation in development activities 62 Chapter 4 B. REGIONALIZATION: POTENTIAL AND CONSTRAINTS FOR GENDER-SENSITIVE DEVELOPMENT 4.7 The regionalization policy is welcomed in the country as a more participatory form of governance and a promising step for addressing gender issues. The new institutional framework has the potential for enhancing women's role in regional development and incorporating their concerns and needs in the regional development plans. A regionalized approach could result in more responsive solutions to needs and priorities of local communities, improved efficiency and efficacy of service delivery, enhanced participation of beneficiaries in local development activities, and better capacity to identify, and gradually address, region-specific socioeconomic factors accountable for the gender imbalances. 4.8 It is a difficult task to clearly assess the overall impact of the regionalization policy on the implementation of the Women's Policy. The process is quite recent, the information base is limited, and consistent regional gender-disaggregated data is not available. This section, therefore, analyzes the potential for gender-sensitive development interventions in the new institutional framework, and the constraints that have been identified. The discussion is premised on the assumption that regionalization should lead to at least two improved outcomes for women: * Sustainable and responsive development through gender-sensitive development planning and improved participation of women in development interventions; and - Equitable development resulting from a legal framework that provides women with access to economic resources. Sustainable and Responsive Development 4.9 Sustainable and responsive development for women needs an overall strategy and gender-sensitive plan for each region, resources for implementing the development plan, as well as the participation of women at some critical stages, to ensure that such planning responds to the regional priorities and needs of women. 4.10 Gender-Sensitive Development Planning. The recent changes in institutional arrangements have succeeded in bringing gender issues to the table at the regional levels. Each region has established WABs and most regions have appointed coordinators for women's activities in two to three of the important sector bureaus. In some regions, women's focal points have also been established in the lower government echelons, i.e., zones and woredas. Many of.these offices are dynamic, and are striving for change but due to a variety of reasons, discussed below, the new arrangements are unable to function effectively. and interventions; and (v) monitoring that women's concerns and issues are integrated and properly addressed in the regional policies, strategies, plans and programs. Regionalization and Gender Issues 63 4.11 Administrative Functions and Responsibilities. To function effectively, decentralized entities need clear guidelines on aspects related to their organization, as well as on the rules governing interaction with other entities in the broad institutional environment. Placement of WABs in the Regional President's Office was meant to confirm the importance ascribed to the inclusion of women's needs in the development process. However, their relationships with other institutions, also working on critical areas related to gender, are unclear. WABs are not linked in any formal manner to the other bureaus, especially with core bureaus like Planning and Finance, except in Amhara and the South, where some restructuring has taken place.44 There is no formal relationship between women's departments in sectoral bureaus and WABs; the former being accountable to their bureau chiefs. Consequently, most WABs function in isolation, unrelated to those involved in mainstream development planning. Vertical linkages and downward linkages, i.e., with zones and woredas are also weak. Thus, even though significant investment has been made in these institutional arrangements, in most regions, the potential for WABs to play effective roles in planning and coordinating gender- sensitive development is underutilized. 4.12 Preparation of regional development plans is also undertaken based on gender- neutral assumptions.45 Typically, the guidelines issued by the Planning and Economic Development Bureaus to initiate the process of regional development planning do not require or provide any incentives to encourage regional or sub-regional authorities to address gender issues in the development planning process. The Planning Bureaus are also not mandated to consult with the WABs, though they may include the departmental gender coordinator in the planning process. Also, a sample review indicates that Regional Planning Bureaus do not scrutinize project proposals for gender-disaggregated impact.46 The majority of projects and programs are targeted to the household, which is considered as an homogeneous unit, and socioeconomic and cultural constraints hindering women's participation are not considered. Thus, there is no formal integration of gender concerns into sector or other development plans. 44 It is noted that the women's departments themselves are usually isolated from the mainstrearn work of the bureau. For example, 1he women's unit in the planning bureau is not involved in project appraisal or monitoring unless the project is specific to women. 45 In weaker and remote regions, like Afar and Gambella, the development plan is fonnulated essentially by the regional planning bureau with little, if any, feedback from the woredas, kebeles or the village-level entities, while in other regions, i.e., Oromiya and Southern, planners adopt to the extent possible a bottom-up approach. 46 For example, in the Southern regions, appraisal is based on three criteria: the project should be in the priority sector; it should be asset forming; and the organization should possess the capacity to implement the project. In Afar, criteria include whether the activities are within the regional pnorities; the financial performance of the implementing sector bureau; the implementing capacity of the sector bureau; and the other demands on the budget. 64 Chapter 4 Table 4.2 Women's Representation in Decision-making Bodies Region Bureau Heads Regional Council Reg. Executive Committee Total No No. of Total No. of Total No. of women* No. women No women Afar 19 0 44 2 15 0 Amhara 12 0 274 17 21 0 Garnbella 13 0 40 0 19 0 Southern 16 0 239 6 21 0 Oromiya 19 2 353 14 21 0 * Does not include WABs 4.13 Capacity Issues. The capacity of WABs varies from region to region; it is a particular issue in the remote regions. It is clear that capacity needs to be built up particularly in participatory development, but a word of caution would not be out of place. Unless other institutional aspects (administrative and budgetary issues) are addressed, merely enhancing capacity of these bureaus will only increase the frustration of the staff, particularly in the larger regions, which do have certain levels of expertise. 4.14 One way of testing the readiness of decentralized institutions for pro-active gender-sensitive participation is to assess the number of women in leadership or managerial roles. A larger number of women in the top levels of management could perhaps assist in bringing gender issues to the fore, and would certainly help change perceptions of women. In Oromiya, for example, there are no women representatives in the regional council or executive committee. There are some women in Kebele, district and zonal councils. In almost all the regions, except for WAO, heads of sectoral bureau are all male, even though these are political appointees, which provides considerable scope in appointing women. The Constitution permits the use of affirmative action as a tool to address gender inequality. In order to jump-start and bring women into decision making roles, regional governments should examine the implications of formulating an affirmative action program for women, if not in administrative positions in the regional government, at least in the local council. The staff of the regional governments should also be targeted as primary recipients of gender awareness and sensitization measures, and a firm consensus should be built at higher echelons of regional governments on the need to improve women's living condition and their role in the society. 4.15 Capacity of staff is also limited by the lack of gender-disaggregated data, at regional levels and below. WABs do not have access to gender disaggregated information to enable sensitive policy formulation, and do not receive timely feedback on performance and problems encountered in the field. Also, systems for monitoring and evaluation are currently very weak, and these impede the assessment of implementation progress and impact of regional development plans, projects and programs from a gendered perspective. Sector and planning bureaus are currently not required to incorporate gender-sensitive monitorable and measurable targets so as to be able to signal both successes and failures in reaching women. Regzionalization and Gender Issues 65 4.16 Remuneration and incentives of staff will also need to be addressed along with capacity issues. Although the pay structure for WABs and other women's units is similar to that of sectoral bureaus, WABs do not provide opportunities for upward mobility. The undefined advisory role of WABs discourages officials used to clear job descriptions, powers and responsibilities, and deters them form joining these bureaus. Also, there is a distinct reluctance, especially at zonal level, to budget for staff working in women's focal points in a timely manner. 4.17 Fiscal Issues. Review of decentralization processes in other countries indicates that adequate funding for lower-level institutions was the single most important factor that undermined many of the decentralization attempts of the 1970s 47. It is critical that decentralized institutions have resources to carry out their mandate. Also, their budgets should be predictable and determined in a transparent manner, otherwise effective planning is impossible. With the exception of the WAB in Southern Region, women's bureaus do not currently receive any share of the capital budget, and they do not have any independent access to resources. Generally the budget they receive covers only a small amount of their recurrent costs. They are therefore forced to turn to sympathetic donors who provide grants based on the donor's perceived priorities and needs for the region. This significantly affects their ability to operate and leaves them with little or no leverage or bargaining powers. 4.18 Budget allocation is undertaken at the regional levels, utilizing either their own income or subsidies transferred by the federal government48. The current intergovernmental fiscal transfer formula is based on gender-neutral criteria such as regional population size, relative level of regional development and capacity of the regions in generating fiscal resources. The subvention formula also does not provide any incentives to ensure that regions allocate adequate budget for the implementation of the Women's Policy. Moreover, revenue collected by each region is a criterion for determining budgetary transfers to the regional governments. This reduces the budget availability in poor regions where the needs of women are significant, but may be marginalized during the allocation of scarce resources. 4.19 Women's Participation in Development. Regionalization has the potential of enhancing equity in investment in the delivery of services resulting from wider representation and transfer of power to local persons who would, in a more centralized arrangement, be excluded from any decision making on matters of development.4 However, an active civil society with capacity for participation is important in contributing to an effective and well-functioning decentralized system. 47 Zyl, Barbosa, Parker, Sonn, 1995 Policy Research Working Paper No. 1948. Agriculture and Natural Resource Department, World Bank, Washington D.C. 4S Govermnuent of Ethiopia. Proclamation No. 33/1992. 49 Silverman, Jerry M, 1992 Public sector decentralization: economic policy and sector investment programs. World Bank Technical Paper No. 188, World Bank, Washington D.C. 66 Chapter 4 4.20 Participation of Women in Development Activities. Women's participation in the planning and implementation of development interventions is currently limited and weak. The high illiteracy rate, the low levels of educational attainment, lack of skills, the absence of leisure, given their household and other responsibilities, combined with lack of information on developmental opportunities, contribute significantly to this lack of participation. Therefore, the majority of women are unable to speak for themselves. It then becomes important to have strong women's groups or representatives to speak on their behalf However, PRAs indicate that such groups are rare, and are scattered among urban or semi-urban areas. Women's voices are, therefore, not effectively heard in development planning, particularly in the rural parts of the country. 4.21 Common groups that are currently involved in development interventions include Peasant Associations and Service Cooperatives. These were conceived as autonomous and locally accountable units for governance and delivery of services to their members. However, in reality, they lack capacity and resources, and are controlled by men who rarely consider gender issues in decision-making. Although not barred legally, women are not actively encouraged to become members, membership being limited to a few women heads of households. Further, the requirement for individual membership fees is a disincentive for registering more than one family member. The net result is that women rarely participate in formal institutions that are currently used for delivery of a number of important services. 4.22 At the same time, the PRAs indicate that women participate in a variety of indigenous self-help groups, but such groups are only marginally involved in development activities.50 These self-help groups hold tremendous potential as vehicles for women's participation. Traditionally accepted, they are based on peer support and pressure. The rules in such groups are more flexible and responsive to women. They are less demanding in terms of time, since informal meetings can be conducted at the market or on the way to collecting water. However, while these constitute attractive features, the lack of legal status perhaps is one single feature that currently deters the participation of these groups in development interventions. The absence of registration means that these groups are often difficult to identify or organize into service delivery agents. 4.23 Some argue that there is a need to ensure a minimum level of formality, particularly in terms of accountability when interacting with third parties, if these self-help groups are to be effectively used as instruments through which to deliver services. Others argue that introducing such formality would reduce the flexibility and attractiveness of such groups for women. The 1994 Proclamation on Agricultural Cooperatives5" seems to provide regional governments with a way forward. While the proclamation was issued to create the new regulatory framework for establishment of primary rural cooperatives, it 50 These include idirs (a funeral association), iqubs (a rotating saving and credit association), debo (labor sharing arrangement), maheber, senbete and tertim (religious associations). Indigenous associations also exist in pastoralist areas though they show a lower degree of institutionalization due to the nomadic or seminomadic life style of the local conmmunities. 51 CGovemrnent of Ethiopia. Proclamation No. 85/1994. Regionalization and Gender Issues 67 also allows for the creation of smaller secondary cooperatives below the Peasant Association level. This provision could be utilized to assist women in forning small legally recognized groups that are registered and monitored at the village level and can be used as units for the delivery of services to women. These groups could continue to maintain the flexibility enjoyed by informal associations. Regional governments have not yet used this provision for the establishment of small self-help groups. Such a step should assist in increasing the participation of women, if not as planners at least as beneficiaries. 4.24 Participation of Non-Governmental Organizations (NGOs). The lack of institutional capacity within the government institutions on one hand, and the growing demand for services on the other hand, increases the scope and potential for involvement of NGOs, and the private sector as well, in the provision of services to poor women. Broadening collaboration between governments and NGOs would assist in increasing coverage of services, including capacity building activities, to poor women particularly in less accessible communities. NGO's have a comparative advantage and important role to play in giving a voice to women in situations where traditional social arrangements are formal government structures leave them underepresented in decision-making. Given the weak capacity of women, and the >'irtual absence of formal women's groups, NGOs can play a critical role in enhancing the impact of regionalization on women. 4.25 Currently, however, NGO role in service delivery is insignificant. For example, only six percent of NGOs are operating in the education sector, and most of them are concentrated in Addis Ababa, Oromiya and Southern Region. It is estimated that 95 percent of existing kindergarten services in the country are run by NGOs, but such a service is currently available to only one percent of the children of kindergarten age. 4 26 The regulatory environment is also not conducive for NGOs and their increased participation in the provision of public goods.52 The existing framework is non- transparent and does not facilitate their efficient functioning. For example, there is no consistent rule for registration. NGOs are required to repeatedly obtain clearances from the central, regional, woreda, and kebele level entities. Also, since the requirements are not clear, they are asked to return with different types of information, increasing the tediousness and costs of registration. 4.27 As a result, NGOs do not currently play a significant role in women's development. For example, both in Afar and Gambella only one NGO is currently working with women. On the other hand, given the poor state of the infrastructure, in particular roads and markets, NGOs also prefer to operate in urban and semi-urban areas, with very little and often no presence in remote areas where the regional governments need more support. There is, therefore, the need for creating mechanisms and incentives to both increase chances that remote and poorer areas are increasingly targeted, and to permit the various government agencies, NGOs and the mass organizations to work more closely in the design and implementation of plans and projects for local development. 5X A guideline for NGOs operation in the country has been issued by the Relief and Rehabilitation Commission. The guideline is known as "Chapter Six" of the National Policy on Disaster Prevention and Management. 68 Chapter 4 C. IROVED AND EFFICIENT DELIVERY OF SERVICES 4.28 Regionalization and Delivery of Social Services. Low access to social services, particularly in education and health, is a critical factor hindering the participation of Ethiopian women in the country's socioeconomic development. Within the broad health and education policy framework established at the federal level, regional governments are the key stakeholders responsible for improving access to health care and making education for women a planned and sustained process. The regionalization process thus brings to the forefront issues of technical and fiscal capacity of regions to ensure the required level of investment in health and education, and improve the delivery of services to the local population, particularly to women. 4.29 Current coverage and quality of educational and health services is poor throughout the country (see Table 4.3 and Box 4.2). Aggregate available data also hides the existence of a significant interregional gap, which in turn reflects the relative level of regional infrastructure development. Inequitable access to health facilities across regions also exists with Afar having one of the highest ratio of population per doctor and per health care facility. In regions with greater availability of services, such as Gambella and Southern, the inadequate road network and the absence of modem transport/communication systems make most of the existing facilities inaccessible to the rural population.53 All this means thus that regional governments are required to invest a considerable amount of resources in order to meet the increasing regional demand for health and education services. The task is daunting yet necessary for attaining sustainable long- term development. 4.30 The regionalized arrangements provide significant advantages for enhancing effectiveness in the provision of social services. Each region may now adopt or develop approaches that are region-specific. This appears to be critical in regions where access to available services is highly dependent on underlying socio-cultural attitudes towards girls schooling and women's health care. Regional authorities would need to decentralize responsibilities further and where possible to the facility level, in order to strengthen the potential for women's participation in the design and delivery of services, something envisaged in the National Health Policy. Under the Government's Ethiopia Social Rehabilitation and Development Project (ESRDF), communities are involved in project design and implementation. However, in the provision of health and educational services, involvement during the post-implementation stages is difficult, since the facilities have no scope for independent revenue generation. Both men and women are, therefore, restricted to serving on committees that oversee implementation. In the ESRDF Project, it is mandatory that 30 percent of all Community Project Committee members should be women. It has been difficult to ensure even this level of participation, but is increasingly found to be acceptable to men. Increased exposure to the outside domain, with the implicit sanction of the community elders, is the first phase of change in gender roles for 53 For example, in the case of Southern Region, due the lack of roads and transport facilities, only 47 percent have access to services (SNNPRS Regional Government, 1996). It is also estimated that rural population need to travel an average distance of 27 kilometers to receive primary health care (World Bank, 1998b). egionalization and Gender Issues 69 many poor women in traditional and rural societies. However, project experience indicates that even this degree of involvement is found to significantly raise community ownership of the facility, and greater and deeper involvement would be possible, if the policy environment is conducive. 4.31 Also, accountability has moved closer towards the ultimate beneficiaries, making it possible to better understand the impact of services on poor women. With training and provision of technical assistance, it will permit the establishment of gender-sensitive monitoring and impact indicators, particularly in the areas of family planning and reproductive health care. This would provide lessons in cost-efficiency and fine-tuning policies to enhance the impact of development interventions on women. Further, given the marked differences between regions, it would be useful to examine whether some regions should receive greater support than others, in order to lessen existing disparities. This is particularly important for regions with a limited tax base and scarce economic resources. .. ... . ... . . ... ..... . ... .. ..... . . ,, ~. .. ; ... .... . . . . .... ... .... . . .---.4,2. W-$s- fa.l. -h an Ed -io St:t in ' hioph .t.p.aryschl.... . . .. ..il is .nly ..... . . The .et enrollmn ra tio. stnsa 13 prcn vi. -I pecet fo by. Theagreteda aX.o als th eisee of a - nntepart at seonar and. te-:ar s-o As, codn to daapoie -yt~Mnsr of- E.aetin i thaadicya1994195, les Tha 23pr:to disos odge r~ awr . ..-.. .... ? . ' ? Besides, g~rls perform p d -o r t ;ute luh erenag o shol roons |mportant 4et na-s of the poohealtsfas of 'omen 'e regon-speificy soc2 a and cultral vaib,.e.., Th,,,,,,,,,.,,¢,urpretofwmnmn be,,re thea of.: 1 .5? an oe hn56 peren:'r' ?befr th g 1, thea exoigwmnto th health inctin of muiple. prganies ndblfid: - - -? f * P e ? ? . -W '2 Mm,'t Qi Sore f dXcatiou~ Fducafion ich a f--disAbaba. ... . .y.....1.9 - : ............ {,-tS eX ..-............... .. ... .9 .--an -p .: .pe:r-.-:-.. .- ~ ~ ~ ~~~1 . th . ..,. ...,.. ,.,>.... .. ... .. .... ........ .. f ? .,---,.- .. . Table 4.3: Selected Socioeconomic Rgional Indicators Regional Indicators HSurvey Demographic and Household Ethiopia Tigray Afar Amhara Oromiya Somali Benshangul SNNPR Gambella Harari Dire Dawa Addis Ababa Indicators Area 1060 60 77 189 360 216 47 112 26 0.3 1.6 0.4 Population 55,765,000 3,358,000 1,131,000 14,769,000 20,012,000 1,979,000 493,000 11,065,000 195,000 144,000 277,000 2,342,000 Urbanization (89) 14.4 2.9 9.5 11.1 7.4 7.3 7.2 16.6 58.8 68.7 99.2 Population below 15 . __ Life Expecteancy atBirth 50 51 50.4 . 49 54 _ _ Male 53 50 49 48 54 Female 47 52 52 50 55 Literacy and Education Level Ethiopia Tigray Afar LiteracyRate 26 7 18 22 24 30 ._ _. male 36 9 23 29 34 39 female 17 6 12 16 15 20 Gross Primary Fnrollment 29 53.1 9.1 19.3 23.1 14.3 41.1 32.2 60.6 54.4 43.1 86.7 male 36 60 10.9 20.7 30.2 19.9 57.6 44.4 74.5 57.6 45.5 85.9 female 22.1 46 7.2 18 15.6 8.2 23.3 19.2 44.7 50.8 40.7 87.4 Gross Secondary Enrollment 26.1 43.7 8.4 17.9 21.2 11.6 35.4 28.8 53.9 53.4 41 84.9 male 32.2 49 10 18.9 27.2 16.2 49.3 39.5 67.8 55.3 42.8 84.4 female 20 38 6.8 16.8 14.9 6.6 20.2 17.4 38.2 51.4 39.1 85.3 Gross Tertiary Enrollment 6.5 4.4 1.5 5.2 5.6 0.5 2.4 5.1 5.3 31.6 19.4 40.8 male 7.5 5.5 1.9 5.5 6.5 0.8 3.1 6.6 8.2 35.9 21.6 45.6 female 5.5 3.4 1.1 4.8 4.7 0.2 1.7 3.5 2.3 27.6 17.2 37.2 Health Status Infant Mortality Rate 120 118 116 118 128 99 Under 5 Moratlity Rate 174 170 173 189 137 . Fertility Rate Assisted Births* 0.5 4.2 5.3 4.8 10.5 Family Planning 1.5 4.7 3.1 0.9 Access to Services _ Hospital per Pop. 279,833 565,500 1,342,636 800,480 659,667 246,500 1,229,444 195,000 28,800 92,333 146,375 Health center per Pop. 239,857 282,750 328,200 317,651 282,714 1,941 325,441 65,000 0 277,000 146,375 Health Station per Pop. 23,816 33,265 33,264 23,683 29,985 7,145 26,097 5,735 28,800 15,389 6,00 5 Doctor pcr Pop. 39,050 75,429 58,608 63,735 46,013 16,989 43,391 12,172 3,418 6,970 8,401 Distribution of Schools 10,836 83 2808 4178 2142 81 No of Primary Schools 9276 68 2503 3601 1823 67 No of Secondary Schools 1230 12 229 468 256 10 . No of Tertiary Schools 330 3 76 109 63 4 Per Capita Expenditure for Health 15 16 6 7 10 17 8 34 55 17 20 Per Capita Expenditure for Education ._ _ Labor Market Participation Economically Active Population 1,476,129 620,406 7,647,025 9,299,158 260,839 4,328,319 74,424 48,763 89,574 614 male 786,216 402,678 4,174,356 5,167,678 1 135,609 2,601,129 43,480 29,331 54,600 408,758 female 689,913 217,728 3,472,669 4,131,480 125,230 1,727,190 30,944 19,432 34,974 232,388 Source: (a) PHRD Ethiopia Social Sector Study. Access to and Supply of Health Facilities and Services. Addis Ababa. September 1996. (b) Ministry of Education, Education Statistics. Addis Ababa, July 1996. Regional Development in Ethiopia. CSA, 1994 Population and Housing Census... etc. Regionalization and Gender Issues 71 Agricultural and Related Services 4.32 The agriculture sector constitutes the mainstay of the Ethiopian economy contributing for about 45 percent of the GDP and employing more than 80 percent of the country's labor force. Eighty-nine percent of Ethiopian women live in rural areas, they constitute almost 50 percent of the rural labor force, and the agriculture sector is their largest employer.54 4.33 Devolution of powers to the regional levels to implement the agricultural policy is very important since there are significant interregional variations in degree and extent of women's participation in the agriculture sector. In regions where hoe culture prevails (such as the Southern and part of the Oromiya), women are involved in virtually all stages of crop production, and about 70 percent of their time is spent working in the field. In highland regions, such as Amhara and Tigray, where the plough dominates, women are not directly involved in farmiing but play a significant role in related activities like food processing, dairy husbandry and backyard cultivation. In pastoral societies, like Afar, herd management is a male task but women are involved in different animal husbandry activities. They are responsible for feeding and looking after goats and sheep, taking them to watering and grazing places, milking livestock (except camels), processing milk products, as well as processing hides and skins. Despite their contribution, women are regarded as marginal workers and do not receive equitable access to critical inputs of production such as land, fertilizers, credit facilities, and are not targeted as potential recipients of extension and other support services. All this results in their productivity being significantly lower than men's across regions. 4.34 While access to land is an important issue and is discussed later, it is equally important to ensure that women farmers have access to extension services so as to enhance their capacities for raising land productivity. The current extension system, however, poses a number of constraints for women. Extension workers are mostly men who seldom perceive women as potential participants and beneficiaries, and services focus mnostly on male-related activities. For example in Amhara region, women constitute less than 24 percent of the extension workers or development agents (DAs). In Gambella, all the 39 Contact Farmers assisting in service delivery are male farmers. Women's participation, if any, is limited to home economics, weed control, and poultry production. An additional factor is that the main thrust of the agricultural extension strategy is geared to better off farmers. Although there are significant implicit subsidies in which to obtain fertilizers and improved seeds, a 25 per cent down payment is required. Poor farmers, especially women, find it difficult to afford the initial payment, since it falls due at a time when other debts are pressing. This suggests a need to review the poverty and gender impact of the present arrangements. 5" Central Statistical Authority, 1996. Agricultural Sample Survey 1995/96. Vol.4: Report on Crop Land Utilization. Statistical Bulletin 152. Addis Ababa. 72 Chapter 4 4.35 As reiterated in the gender strategy prepared by the Women's Affairs Department in the Ministry of Agriculture for incorporation in the National Agriculture Policy, there is an urgent need for designing more gender-responsive contents and mechanisms for the delivery of extension services. Regional governments now share this responsibility. While the broad policy framework for extension advises and services package is formulated at federal level by the Ministry of Agriculture, the management and implementation of the policy is left at the regional levels. Regional government actions are, therefore, critical in ensuring that women are effectively targeted as service beneficiaties. Equitable Development 4.36 Regionalization permits increased attention to governance issues, a necessary pre- condition for equitable development. A predictable legal framework is essential, and should provide for equitable allocation of resources between men and women and a satisfactory dispute settlement method. This section briefly discusses selected aspects of the existing legal framework governing women's access to critical economic resources, focusing on four areas: allocation of resources within a household; women's participation in the formal and informal labor market; women's access to land and natural resources, and mechanisms for dispute resolution. 4.37 Equitable Access to Intra-Household Resources. Equitable allocation of intra- household resources is necessary for sustainable development. The legal framework must allow women to control and manage all types of economic resources. Currently, despite the Civil Code (1960) that has invalidated all customary laws and practices that govern allocation of resources within a household, customary practices are observed by the majority of rural people. Customary laws are inequitable in the allocation of resources, and control is mostly with men, women being unaware of the more progressive provisions of the Civil Code that deal with division of resources. Moreover, the 1994 Constitution complicates the matter by introducing a clause which permits the applicability of customary laws to personal disputes, should both parties to a dispute agree to such application.55 Customary practices also differ across regions, and therefore in amending the legal framework, attention needs to be focused on these regional differences. Critical regional stakeholders need to be involved in addressing these issues in manner where legal reform will be effective. This aspect is discussed in detail in the joint Government/Bank WID Report. S5 Enabling regulation is necessary, but is still under discussion by the Government. Reg!onalization and Gender Issues 73 4.38 Equitable Participation in Formal and Informal Sectors of the Economy. The next few years are expected to be characterized by rapid expansion and increasing economic opportunities for Ethiopians. To maximize the returns to these investments and to ensure sustained economic growth, it is essential that human resources are deployed and utilized effectively. It then becomes critical that women's concerns are systematically incorporated into economic policies and strategies for economic development. 4.39 Regional governments now take primary responsibility for ensuring that in the process of providing private operators with incentives for investing in the regional economies, women's concemns are considered and incorporated. Since regional economic growth driven by private initiatives is likely to increase the share of women in wage employment, regional governments would need to develop monitoring systems to ensure that rights and employment conditions of the female labor force are protected. For example, though consistent data are lacking, it is reportedly noted the existence of a marked gender differentials in the current wage structure (see Box 4.3). This trend would need to be closely monitored by regional authorities. Weconol Be*,Opp !XM adEpkW-en ~ ~ active ..: p 14on 34 : (:84: Wr w znn . am:t- tI~ a~ unmploed. ~mem Pcout fravr ml hr fepoim n eev ne snij1er of W oum,bein enae nlwdcso-ma1 i n o anngjh.I set1a ihat 4uiix~~g the I9~~0s, .ltoug .....u o 0 ecn o h or oc i aiCatdg ~iterpises ima1ewSes wepi Jn-ly 21preto h oAl ae. h 1drtgd~ ~n~empoycd re illterat hoa (34 : pecn)adIowh aeciltdg:ei 2 ecn) ~Pa& rleaed y teSiisy of Labo.n oilgp1 los thtans.0pr~to h fcniate~ jo seekers (43~659) ar&gir1~ betwen uS and 1. While 4O percent..fgo...n.. . cimplycesare wen.8 71pret fte (5 .eren..fx.....econenratd t ~~., those earning l~s than 40(~ Bitt per mouths Almost 6 percent of the men earn salaries ~~bovr. Mlr 1)0;ls fithar i HOD&ApretV h oinaei h sicslr ag.Tit ecn of t WO npQes anshre abo- K!..ir 60; Nly. ie ftewme.We nlzd WIrsa eVr,thatq,*,'i .7.aei o~ ndtv atvte ang eti efraepre to ot~ cemiua Sv~nu~; entA& LttsaboAu~riy 1994t9 srbmo::dohr,192FdrlCii.5r 4.40 Female Entrepreneurship and InomlScor h noml etr hc acoutsfow9iprcn of4;5 privae mauatuig uretyemly alrenubro women. recen surve condcte acossrgosb h eta SaitclAtoiyi 1997. inicte tht6Vecn ftoeoeaigi h ml adcat auatrn secto (lagelyinfomal)are emals. Aso lteraure nd epirial eidene.sugest.tha the har of emae inutsto te dmesitaeismr than5pectinvlm,ahog les i valu duelf to M ine ficece inrti_taeadwomn'W ivlvmetnlo-pi semet.-heCA ure aolso shwta huhntr inwnme,a ela xeto women's participation, may vary, all regions are affected by this phenomenon.~~~~~~~~~~~~~~~~~~~~~..... . 74 Chapter 4 4.41 Polices aimed at promoting entrepreneur development cannot, thus, overlook the significant role played by women in the micro-enterprise/informal sector. The PRAs indicate that more and more women, even though strapped for time, are gradually moving into economic activities outside their homes to increase their access to economic resources and for survival. Given the lack of earning oppormnities in the wage sector, women in both urban and rural areas turn to the informal sector. 4.42 The scarcity of financial resources needed for financing either business start-up or expansion of already existing activities is the most severe obstacle to raise women's productivity and income. Almost 70 percent of the small manufacturing establishments assessed by the CSA survey initiated their businesses vwith very little capital (250 Birr or less). Yet, the lack of capital was reported by 60 percent of them as being the most severe constraint. Only 0.22 percent of the establishment were financed by formal financial institutions. To its credit the government has recognized the importance of creating an environment conducive for the promotion of female micro and small entrepreneurship, by issuing a regulation on the rnicrofinancial sector. The merit of the new regulation is that stresses the importance of expanding easily accessible and sustainable rural schemes for the rural poor. However, it could also impact women adversely. Currently, a number of NGOs are successfully implementing microfinance schemes for poor women.56 If these NGOs have to register as companies, and be subject to the other requirements of the new Proclamation, it is possible that they will wind down their operations, reducing women's existing access to credit. 4.43 Lack of a place for working or trading, particularly as this seems to be a requirement for both credit and licensing, represents another major constraint. In Addis Ababa, the provision of premises for trading is the responsibility of the administrative councils at the various levels - kebele, woreda, zonal and regional. A permanent trading place costs Birr 2-3 a day, which is difficult when these women have a maximum income of Birr 3-4 on their good days0' Licensing procedures are also cumbersome.5' The federal government has recently taken a positive step through the promulgation of the "Commercial Registration and Business Licensing Proclamation" aimed at streamlining procedures for licensing and eliminating licensing requirement for enterprises with an initial capital of less than Birr 1,000. The impact on poor women will, however, be clear only when regional governments pass their own regulations, and to date regional governments have not yet regulated the licensing of such enterprises separately. 4.44 To create an enabling environment for microentrepreneurs, the Ministry of Trade and Industry is reportedly formulating a Micro and Small Entrepreneurship Development Policy in collaboration with the regional governments. The Policy is expected to establish a Federal Agency with branches at regional levels, which, in consultation and collaboration with other relevant stakeholders, would provide non-financial services. Such services 56 The above mentioned CSA survey reportes that 20% of the establishment had their initial capital provided by external assistance, either donors or NGOs. 57 Further an indication of the difficulties entered by these women as they try to shift to the formal sector is the fact that only two percent of the establishment had a license. Central statistical Authority, 1997 Regionalization and Gender Issues 75 would include: training and skills building, access to information, management support, as well as support and assistance in finding a working or trading place, marketing facilities, transportation facilities. It is hoped that such policies would aim to ensure that regional governments respond with regulations that remove constraints faced by women in the area of microenterprise; and provide women with supportive services that can enhance their productivity by decreasing transactions costs and easing access to markets. 4.45 Equitable Access to Land and Natural Resources. Eighty-nine percent of the female population live in rural areas and largely depend on agriculture-related income. Yet the PRAs indicate that very few women have independent control over land. Except for a few female-headed households, women's access to land has accrued from their position as wives, or mothers of male children. According to a survey on Crop Land Utilization recently conducted by the Central Statistic Authority only 17.6 percent of the existing landholders are female. The rest of women either work on land primarily cultivated by the male members or cultivated separate plots of land assigned by the husband. 4.46 The 1994 Constitution reiterates the principle of equal rights to women with respect to use, transfer, administration and control of land, but it confers the responsibility for implementing laws and administering land and other natural resources to the regional states. The Federal Ministry of Agriculture has the power to enact laws for the utilization and protection of land and other natural resources, whereas the responsibility to administer land rests with the regional states. A proclamation on rural land administration has also been issued. Regional governments, therefore, now have the authority to ensure gender equity in distribution of rural land and are actually obliged to do so, given the federal law and the 1994 Constitution. To date the only regional state that has issued a regulation on land is the Amhara region. 4.47 The Amhara Land Proclamation aims to achieve a single and final redistribution of land. Under this law, the Peasant Associations would provide each new landowner with a legal certificate of land ownership. The Proclamnation defines land tenure rights to include the right to sell, exchange or bequeath property. According to the regional Womenls Affairs Bureau and the Amhara Council, the land proclamation has benefited women, i.e., by 1997 a total of 554, 889 family heads have received rights to land, of which 129,677 or about 23 percent are women. 4.48 The Amhara Land proclamation provides some lessons for regions that wish to pass land administration acts. A tradition that is maintained in Amhara, is that village authorities or kebeles are conferred the power to administer land. Their decision-making authority is significant, and is not subject to further judicial scrutiny except in certain limited cases. This system holds great potential in that decisions can be taken in a participatory manner, and women could be given opportunities to have their voices heard. kebeles, however, are male-dominated institutions and these groups tend to interpret the regulations in a customarily accepted manner. Another point is that under the proclamation, the term "household" is defined as a "social unit that includes people who .live under one roof and depending on the situation who are under the responsibility of a 76 Chapter 4 male or female household head."58 Clearly, female-headed households are included and can now access land. However, the position of women in polygamous households is not clear. The definition of "eligible household" does not clarify the rights of the different households in polygamous families. Households are defined as a "social unit under one roof " In such cases, the question arises whether "wife-headed households" are entitled to independent land, or whether polygamous households are allocated land as "a single social unit" that should then be divided by the male head amongst his wives. In the latter case, married women become extremely vulnerable and dependent on the spouse. This would be of importance among communities that practice polygamy. 4.49 The Land Proclamation is silent on the criteria for land distribution, which is in the hands of kebele elders. In the past, this has been known to create problems. It has been alleged that allocation of land has varied due to favoritism and inaccuracies in measurement.59 The Women's Affairs Office document on gender problems and priorities also indicates female headed households have smaller pieces of land. Social status and economic resources were also found to be important in influencing the pattern of land distribution.60 Thus, the failure to provide guidelines to determine equitable criteria could operate against the interests of women and other vulnerable sections, particularly in regions where customary practices do not give them any rights to land. 4.50 The Amhara Proclamation fails to specify how the land should be registered. In the absence of clear directions, and since both according to the customary law and under the Civil Code, the husband is deemed to be the head, it is likely that in all cases land will be registered in the name of the husband, and that he will be in full control of the land. Thus, this new proclamation could protect the female head of a household, but reiterate the dependency of the married woman on her husband. Moreover, this also means that in the event of divorce or death of the husband, women are easily dispossessed of any rights to use land. Women would have been much better protected had the proclamation '8 Persons eligible to land entitlement include: * Heads administering families; * Couples who have been married before 1996 and have been living together in their own house; * Pensioners who have decided to live together with their children or those who pensioned them; * Persons who have attended marriageable age earlier and have been married, or due to lack of money couldn't make marriage and have been working as servants in the house of others in their respective dwelling rural localities; * Women living in the rural locality and make money for living in rural localities and their means of subsistence is based on agriculture; * Small traders who live in rural areas and whose livelihood is based on agriculture; * Town dwellers whose main source of livelihood is based on agriculture; * Pensioners who live in rural areas and has held land formerly for their living;; * Publicly elected representatives whose basic means of existence depends on agriculture, but are serving in positions of responsibility in urban centers; * Persons who moved away from other rural Peasant Associations and have resided no less than five years in the Peasant Association where they now live, and who have decided to remain in the same place, and who can provide documentation that they do not have rural land holdings in their name where they formerly resided; and * Persons who without being separated from agricultural activities has been engaged in smallscale craft activities in rural locality. 59 Land Distribution and its Implication for Peasant Differentiation in Wogda, Northern Shewa, Ethiopian Journal of Development Research, April 1995, 60 Id at 7. Regionalization and Gender Issues 77 required kebeles to ensure that land is registered in the name of both spouses and to enforce a rule that neither spouse alienates such land or deprives the other of it, without the consent of the other. 4.51 Settlement of disputes relating to land is conducted by arbitration at the kebele level, and such decisions are final. Kebele elders are likely to be all male, and therefore reflect traditional biases towards ownership of land by women. For example, the Proclamation provides that "persons who without being separated from agricultural activities have been engaged in small-scale activities" can have access to land. It is not clear what "without being separated from agricultural activities" means. If interpreted traditionally or narrowly, this could have an adverse impact on women, because in Ethiopia, agricultural activities is normally defined within the context of land practices that equated "personal cultivation" as oxen ploughing, a male activity. The kebele committee may interpret such terms, not defined at law, consistent with traditional practices, thereby limiting women's access to land. Moreover, the decision of the kebele-level arbitration is final. Thus, if the kebele arbitrators refuse to apply the Civil Code provisions, which confer greater property rights on women, and instead chose to apply more inequitable customary laws, women can not appeal against such a decision. 4.52 Thus, even though the Amhara Proclamation does not discriminate on the basis of gender, the impact of some of its provisions in a gender distorted context is inequitable. Two issues become important here. If regional governments are committed to gender equity in access to land, it is important to study the impact of the Amhara Land Proclamation, before drafting their own regional enactments. Also, it would be essential to provide kebeles with some gender-sensitive guidelines in determining land criteria and allocation. Otherwise, women will not have equitable access to land, and despite the marginal increase in ownership that such proclamations may result in, there will be little or no difference in their legal access to land for the majority of women. 4.53 Access to Natural Resources. Given the allocation of tasks within the household, and the dependency on natural resources for survival, women have a strong vested interest in natural resources management and conservation (NRM). As stated in the National Environmental Policy (NEP) (1997), national resources are the foundation of the economy. Smallholder peasant agriculture, including forestry, is the dominant sector accounting for about 45 per cent of the GDP, 85 per cent of the exports, and 80 per cent of the total employment. Yet natural resources are fast deteriorating. Estimation of annual deforestation, mainly for rainfed agriculture, vary from 80,000 to 200,000 hectares. According to the NEP, there is an estimate that some 2 million hectares of pasture land will have been destroyed by soil erosion between 1985 and 1995. The permanent loss in value of the country's erosion in 1990 was estimated at Birr 59 million. The environment is in a critical state; degradation has adversely impacted women who rely on natural resources for food, firewood, livestock fodder, medicinal plants and other raw materials either for domestic use or sale of surplus to supplement household subsistence economy. In 61 This holds true also in urban areas. It is estimated that biomass resources supply 95 percent of the country's energetic requirements (of which 90 percent is for households consumption), and women fuelwood carriers in Addis Ababa supply 78 Chapter 4 particular, environmental degradation increases their burden as primary collectors of fuelwood and fodder and women now spend most of their waking hours searching for food, fuel, and fodder (PRAs). 4.54 The responsibility for NRM has been devolved to regional authorities under the new regulations. The NEP specifically requires that conditions be created that will support community and individual resource users to sustainably manage their own resources. It also requires that women, as key actors in natural resource use and management, shall be equally treated and empowered. Social equity is mandated in resource use. In terms of management of forests, the main proclamation is the one issued in 1995 (Proclamation No 84). This proclamation also requires that state and regional forests should be managed in a manner that ensures that the inhabitants would be beneficiaries. Central govemment, regional organizations or concessionaires may manage state or regional forests in accordance with the management plan approved by the Ministry or the appropriate regional body. At the same time, the proclamation states that no person shall without a permit form the appropriate body take any forest product, graze domestic animals, or keep bee hives to extract honey in any of these forests. As experience in other countries demonstrates, this has potential for abuse by petty bureaucrats. 4.55 Increasing women's participation, as both users and managers of natural resources, and enhancing their sense of ownership would also constitute a strategic element for the success of regional conservation strategies. The potential to manage regional and state forests, provides regional governments with an excellent opportunity to increase the conservation of natural resources in regions through the involvement and participation of communities at the local level. Thus, regions would need to expedite preparation of gender-sensitive and equitable management plans, and have them approved by the competent authorities. 4.56 Dispute Resolution. Appropriate and accessible remedies are critical, if women are to enjoy their rights. Delays and obstacles in enforcement can negate women's rights. In Ethiopia, there is a dual system of dispute settlement. For all disputes concerning betrothal and marriage, the civil code vests powers in family arbitrators. For all other matters, the conventional or formal judicial system has jurisdiction. 4.57 The practice of family arbitration was inspired "by the idea that judges who are appointed by the state are not perhaps the best placed and the best qualified to resolve disputes of a family nature." Also, it is a well-established and respectable form of dispute settlement, based on the age-old Ethiopian custom of shimagale. The system of family arbitrators is meritorious in that it provides women with greater access to legal remedies than if their only recourse were to the formal judicial system. It also avoids the need to drag personal matters before an impersonal arbiter. However, a number of issues limit the effectiveness of family arbitrators. These include their lack of legal training, their scarcity almost 35 percent of it Regionalization and Gender Issues 79 in urban areas, their lack of power to enforce a decision, and their deliberate postponement of a decision to obtain increased remuneration. 4.58 The system is, therefore, under scrutiny. However, in reforrning the system, it is important that the baby is not thrown out with the bathwater. The family arbitration system provides a dispute-settlement system that is relatively more accessible to poor women in Ethiopia than in countries where the only recourse is the formal judicial system, and it is certainly a culturally acceptable mode of dispute settlement. The system needs to be revised so that it does not compromise its advantages and is more gender-sensitive (some regions are reportedly reviewing the possibility of establishing family courts to tackle the issue). Regionalization now provides an excellent opportunity to deal with the constraints of the existing system. It is possible to develop a decentralized system that would combine the advantages of the family arbitration system, and at the same time maintain a greater level of formality in terms of review that the applicable law is applied in a gender-sensitive fashion. Regional governments would need to take a driving role in this endeavor. D. CONCLUSIONS 4.59 Regionalization holds tremendous potential in increasing the participation of women and bringing them into the mainstream. However, this requires the regions to be proactive in considering gender issues seriously. In suggesting a few priority recommendations, focus is on the proper utilization of the potential inherent in the regionalized framework, as well as on those inputs that are likely to be most critical in achieving the goals of the National Policy for Women. * Regions should formulate a region-specific strategy for the development of women, involving NGOs and the private sector as close partners. Regional governments should employ a "package approach" addressing a number of critical gender issues through a carefully thought out and cohesive program defining clear responsibilities and accountabilities for the various actors. * Strengthen Regional Institutional Arrangements. It is highly recommended that WABs be provided with authority to clear development plans in the regions for their gender content. Such a key role will require greater capacity building and gender-sensitization of all actors. * Improve Collection of Region-Specific Gender-Disaggregated Data. As part of the effort to improve overall data collection, regional policy makers and implementors need gender-disaggregated data to make informed decisions and monitor and evaluate the impact of development interventions on men and women. Data collection systems need to be urgently established. * Provide Incentives to Regions for Gender-Sensitive Interventions. Regional governments should provide sectoral bureaus with incentives to mainstream gender concerns in their activities. Matching resources could be 80 Chapter 4 provided to bureaus undertaking priority activities for women. Also, the Federal Government may wish to consider the sub-criteria or pointers that constitute the basis for transfer of subsidies to the regions on a gender- disaggregated basis. MAP SECTION IBRD 29284 35 ~ 4/ K \ L ~~~~~~~~~~~~~~~~~R d REPBLI 15 -~~~~~~~~~~~~~~~~~~~~~~~~-~~~~~~~~~~ / ~~~~~ 9LICERITR~~0 SU D A N -~~~~~~~~~~~~~~ OF Khodooro ~~~~~~~~~~~~~~~~~~~~~~ /~~~A--lor*> REP. OF YEMEN S ~~~~~~~U D A1 N ror \Ho dro~ Ad e ,o4.' Y E M ESUDN )7Y 4 G.d- AMHARA~~~~~~~~~~/ AA/D,~j (A-Tj i o oot ~BENSp'%GUL 7'" ~~~~~ ~~~ SOMAIIA Ad!, Zerne A4-MFAA ARirbo T~ ~ ~~~~~b N S.N.N.P.R.S. . bod4 obo 0910 B~A, Torro~ KENYA. orb U ~~~~~~~~~Kob, I"- OROMIVA~~~~~~~~~~~~~GAD SOAL 7-a t Bore (or,, ho 'Th -*-- I'~~~~~K. Y~~~~~~~~~ B., ETHIOP~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~IOOIAI L o~~~~Io Welbo. ~~~~or ~~~ s F '~~~, - - '#MIsrokGoshomo ~~~~~~~~ ASPHALT, ALL-WEATHER ROADSTHERROAD IT~~~~~~~~~~~~~~~?oo GRAVEL, AULLWEATHER ROADS ___ bo~~~~000er~~~ NA9~~ObV, G,rir N S5,~~~~l N~~ Gelodj -"~~~~ ESRAL AND! OURL RA DRY-WEATHERATHROADSD 0-..* Mieon sAd, N Woko ' - - "'E' - Kebri - WordoS5~~~~~~~~~~~~~~~~ 0 SELECTEDEECTTOWNSN 'N Abo K do o~~N- STA- VTE/ RFGION [UNOFFICIAL 1 994 BOUNDARIES) S U DA N &-24NevroFjArb~~~~~~~~~~~~~SAorcb ~~~~~ -- , Sh obo ERNATIONAL BOUNDARIESRNAIONL BUNDRIE 5 U DkA N ~ - GA' ''-~ F' Ho~~9S-A~r I Moryorr, N ,~-- .Mo~HA\/ 0 M A L IA Tobelo - 0 tU~~~~~~~~~~~~~~erro Sh,I,odi ,~~~~~~~~~~~~~~ ~ ~ ~ 100~ l 200 JO0N ,0~~~' ~~ (' NNOSO / TVO mop roro yrcdoced by ,do N4Np De~~~~~~~~~~~~~~~~~~~~~~~~~gO 1,Th0,p,d-obyLh.he W-o,N-dT BW-o,k- UGANDA ~//,7 O E NorreI0,bod0~ JULY 1999