Report No. 20810-ET Ethiopia Public Expenditure Review (In Two Volumes) Volume 1: Main Report August 31, 2000 World Bank Country Office in Ethiopia Country Department 6 Africa Region Docment of the Worki Bank GOVERNMENT FISCAL YEAR July 8 - July 7 CURRENCY EQUIVALENTS Currency Unit: Ethiopian Birr (Br) Official Rate: US$1.00-Br 8.239 (August 12, 2000) ACRONYMS AND ABBREVIATIONS ADB African development Bank MEFF Macroeconomic and Fiscal Framework ADF African Development Fund MTEF Medium-Term Expenditure Framework ARM Annual Review Meeting MEDAC Ministry of Economic Development and Cooperation BRDM Budget Reform Design Manual MoE Ministry of Education CDF Comprehensive Development Framework MoH Ministry of Health CIDA Canadian International Development Agency MoF Ministry of Finance CSA Central Statistical Authority NBE National Bank of Ethiopia CSR Civil Service Reform NFSP National Food Security Program DFID Department for International Development NGO Non-Governmental Organization DPPC Disaster Preparedness and Prevention Commission O & M Operation and Maintenance EC European Commission OECD Organization for Economic Cooperation and Development EFY Ethiopian Fiscal Year PEM Public Expenditure Management ESRDF Ethiopian Social Rehabilitation and Development Fund PEP Public Expenditure Program EMCP Expenditure Management and Control Program PER Public Expenditure Review ESDP Education Sector Development PFP Policy Framework Paper EU European Union PIP Public Investment Program FIS Financial Information System PMO Prime Minister's Office FY Financial Year PIM Program Implementation Manual FSP Food Security Program PRSP Poverty Reduction Strategy Paper GDP Gross Domestic Product RFB Regional Finance Bureau GNP Gross National Product SDP Sector Development Program GoE Govermment of Ethiopia SIP Sector Investment Program IIICES Household Income, Consumption and Expenditure Survey SNNPR Southem Nations, Nationalities and People's Region HID Harvard Institute for International Development SPA Special Partnership with Africa HIPC Highly Indebted Poor Countries TA Technical Assistance HSDP Health Sector Development Program TWG Technical Working Group IDA Intemational Development Association UNDP United Nations Development Program IMF Intemational Monetary Fund USAID United States Agency for Intemational Development IPF Indicative Planning Figure WMS Welfare Monitoring Survey JRM Joint Review Mission WMU Welfare Monitoring Unit Vice President Callisto Madavo Country Director Oey Astra Meesook Sector Manager Frederick Kilby Task Team Leader Duvvuri Subbarao PREFACE This report is part of a collaborative effort between the government and a multi- donor team. From the government side, the task was led by Ato Hailemelekot T/Giorgis, Vice Minister of Finance, while coordination and logistics were handled by Ato Asrat K/Work, Adviser in the Ministry of Finance. From the donor side, the task was coordinated by the World Bank with Duvvuri Subbarao (AFMET) as the Task Team Leader and Frederick Kilby (AFTM2) as the Lead Specialist. The government nominated counterpart officials drawn from the Ministry of Finance and MEDAC to work full time on the mission (Listing at Annex A). The donor segment of the mission included 29 members representing a broad spectrum of 5 multilateral institutions and 10 bilateral development partners (Listing at Annex B). The mission members were sub-divided into six teams, each working on a different topic of the PER, as follows. Topic Coordinator Overview and context D. Subbarao (World Bank) Review of fiscal performance M. Pinon-Farah (IMF) S. Olanrewaju (AfDB) Information systems for strategic expenditure S. Lister (UK - DFID) management Aid management and integration of aid flows K. Comelis (EC) into the planning and budgeting process G. Kayira (UNDP) Public expenditure - shifting attention from F.C. Sebregondi (EC) inputs to results PER Process . J. Verheul (Netherlands Embassy) The main PER mission was in the field during January 24 - February 5, 2000. In addition to meetings in Addis Ababa, the mission members divided into six groups to visit five different National Regional States (Amhara, Tigray, Gambella, SNNP and Oromiya) and Addis Ababa. The mission also included a full day workshop on February 2, 2000 when the approach and preliminary findings of each of the teams were discussed with a larger audience comprising federal officials, regional officials (Finance Bureau representative form Amhara, Oromiya, Somali and Dire Dawa Administrative Council), a cross-section of the donor community and also, for the first time, two representatives of the budget committee of the Parliament. The workshop provided an important opportunity to debate and brainstorm on the preliminary findings of the mission. The assessment of fiscal performance was based on the government's latest pre- actual figures for FY98/99 and budget execution for the first 4-5 months of FY99/00, data for which were available at the time of the PER mission in Jan-Feb 2000. Subsequently there were two major developments accentuating the fiscal stress. First, the border conflict erupted into a full scale military operation, and second, the drought situation escalated to crisis proportions. The fiscal outturn for the full year is therefore likely to be considerably worse than the projection of this report. A brief update has been attempted factoring in these developments. This report was prepared under the general guidance of Oey Astra Meesook (Country Director) and Nigel Roberts (World Bank Country Office Representative). Peter Miovic was the Sector Manager until January 2000, and Frederick Kilby thereafter. Vinaya Swaroop (DECRG) was the Lead Adviser while Benno Ndulu (AFCO4) and Allister Moon (ECSPE) were the peer reviewers. The extremely competent assistance rendered by Yeshi Gizaw, Fekerte Getachew and Merege Getachew of the Ethiopia Country Office, Roboid Covington and Tanisha McGill in Washington Headquarters by way of document processing and production is gratefully acknowledged. TERMINOLOGY CALENDAR AND FISCAL YEAR The Ethiopian calendar is approximately seven years behind the standard western calendar (thus 11 September 1999 was, for Ethiopians, New Year's Day 1992). The year consists of 12 months of 30 days each (in sequence; meskerem, tikemt, hidar, tahsas, ter, yekatit, megabit, miyaziya, genbot, sene, hamle, nehase) with a 13th month (pagume) of 5 days (6 in leap years). The Ethiopian fiscal year (EFY), however, commences with hamle, and the year from 1' hamle 1991 through 30th sene 1992 is, for example, reckoned as EFY 1992. Thus, in terms of the western calendar, the Ethiopian fiscal year begins on July 8th and ends on July 7t. CHANNELS OF AID FLOW Aid disbursements are classified into three main channels, as shown below. Funds disbursed through the federal Ministry of Finance, which are invariably captured in the budget, are categorized as channel 1. This channel is further sub-categorized as channel IA - unearmarked and channel 1B - earmarked donor funding. Channel 2 represents donor aid negotiated with and disbursed through sector bodies. Funds disbursed directly by donors without the intervention of any government agency comprise channel 3. (See also explanation of the federal structure below). The government has an express preference for channel I on the ground that aid flows through channels 2 and 3 undermine the integrity of the planning, budgeting and accounting systems. CHANNEL 1 CHANNEL 2 CHANNEL3 (via Finance bodies) (via Sector bodies) (direct) Ministry of Finan.e . --.. Sector Ministry Regional Finance ReBional Sector Bureau Bureau Zonal Finance ............. Zonl e Department Woreda Finance .........o.... >c.Woreda Sector Office Offce FEDERAL STRUCTURE Under Ethiopia's constitution, the federation comprises nine national regional states. They are listed below, with population figures from the 1994 census which reflect a tremendous variation in scale: Tigray (3.lm); Afar (1.lm), Amhara (13.8m), Oromia (18.7m), Somali (2.3m), Benishangul-Gumuz (0.5m), Southern Nations Nationalities and Peoples (10.4m), Gambella (0.2m), Harari (0.13m). In addition, two separate city administrations have a similar range of responsibilities, and are often informally included in references to "the regions": Addis Ababa (2. lm) and Dire Dawa (0.25m). Regions are further divided into zones and woredas (districts). The following table shows the terminology applied to different levels of administration. Administrative level Finance bodies Planning bodies Sector bodies (Examples) Federal Ministry of Finance Ministry of Economic Ministry of Education Development and Cooperation Region* Finance Bureau Bureau of Planning and Education Bureau Economic Development (BOPED) Zone Finance Department Planning Department Education Department Woreda (disttict) Finance Office [not present] Education Office* CURRENCY AND EXCHANGE RATE As of August 12, 2000, the exchange rate for the Ethiopian Birr (Br) was US$ 1.00 = ETB 8.239. CONTENTS PREFACE ................................................. TERMINOLOGY ..........................................................III CALENDAR AND FIsCAL YEAR ..........................................................IlI CHANNELS OF AID FLow ..........................................................III FEDERAL STRUCTURE ......................................................... IV CURRENCY AND EXCHANGE RATE ......................................................... IV EXECUTIVE SUMMARY ................................................. The Context For Identification Of Focus Topics...1.......................... ........ ii Review Of Fiscal Performance ......................................... . ........... ii Information Systems For Strategic Expenditure Management .... .................................................. iv Aid Management ...................................................... v Public Expenditure - Shifting Attention From Inputs To Results ...................................................... vii PER Process ...................................................... viii 1. OVERVIEW AND CONTEXT .1 INTRODUCTION.1 Context.1 Approach.3 Areas Of Focus And Emerging Themes. 4 Status Of Previous PER Recommendations ......................................... , . , ., 6 2. REVIEW OF FISCAL PERFORMANCE. 9 FISCAL TRENDS: . 9 FiscAL PERFORMANCE DURING 1998/99 AND PRoJEcTIoN FOR 1999/00 .11 Fiscal Performance During 1998/99 .11 Projection On Fiscal Outcomes For 1999/2000 .11 Risks/Alternative Scenarios For 1999/2000 .12 FISCAL IMPACT OF THE BORDER CONFLICT ................................... 13 Direct Fiscal Costs ................................... 13 Second Round Effects Of The Conflict ................................... 15 Assistance To Displaced And Returnee Population ......................................... ............................... 15 Cost Of Re-Routing Road Freight Via Djibouti .............................................................................. 16 Slow Down In Private Investment ................................................................................... 17 SUSTAINABILITY OF CURRENT FISCAL STANCE ................................................................................... 17 ECONOMIC AND FUNCTIONAL CLASSIFICATION OF EXPENDITURE ............................................... ........ 18 Econo nic Classification ................................................................................... 18 Functional Classification ................... ........................................................... .... 19 Trends In Social Sector Expenditures ................................................................................... 20 3. INFORMATION SYSTEMS FOR STRATEGIC EXPENDITURE MANAGEMENT ............ 25 INTRODUCTION ................................................................................... 25 STRATEGIC PLANNING OF EXPENDITURES ., 27 Context ................................. 27 Strategic Expenditure Planning ................................. , 30 Aid Dialogue And Expenditure Strategy ...................................... 31 Expenditure Strategy And Decentralization ...................................... 34 FINANCIAL MANAGEMENT INFORMATION ...................................... 37 Context ....................................... 37 Budget Planning And Preparation ........................... I , 37 Budget Implementation ...............................39................ , 3 9 Financial Reporting .39 Financial Management Capacity .41 Concluding Remarks .43 4. AID MANAGEMENT AND INTEGRATION OF AID INTO THE PLANNING AND BUDGETING PROCESS ................. 47 INTRODUCTION ....47 Integration Of Aid Flows Into The Planning And Budgeting Process .48 External Aid Flows And The Regions . .53 Implementation Of Donor Aided Programs .................................., 55 The Offset System. 56 Fungibility....... 58 Food Secuty ..60 Government-Donor Dialogue .................... , 63 5. PUBLIC EXPENDITURE -SHIFTING ATTENTION FROM INPUTS TO RESULTS ........ 67 INTRODUCTION ..................................................................... 67 Ethiopia's Experience In Tracking Results . . . . 67 Why Shift From Inputs To Results . ........................................,,,,,........... 69 Indicators - Key Characteristics .......... 72 Experience Of Other Countries ........................... . . 73 Improving Current Practice In Etiopia . . . .74 Next Steps In Strengthening The Focus On Results . . . . ...... 76 6. PER PROCESS .81 Progress To Date In Meeting The SPA Guidelines .81 Further Progress Made During PER 2000 Process . 81 Issues And Dilemmnas On The Way Forward .82 What Is The PER For? Who Expects To Benefit From It And In What Way? ............ ................... 82 Who Owns The PER Process? .................................................................. 83 What Should Be The Institutional Basis For Implementation Of The Recommendations? ............... 85 What Should Be The Role Of The Regions In The PER Exercise/Process? ............... ...................... 85 ANNEX A Members Of PER Mission (Government Segment) ............................................................. 87 ANNEX B Members Of PER Mission (Donor Segment) .................................................................. 88 Tables, Boxes, and Figures Tables Table 2.1 Ethiopia - Fiscal Trends, 1986/87-1999/00 .......................................................... 10 Table 2.2 Estimated Additional Cost Of Re-Routing Road Freight Via Djibouti ..................................... 16 Table 2.3 Private Investment - Approvals And Implementation ....... ................................................... 17 Table 2.4 General Government Expenditures - Economic Classification ................................................ 19 Table 2.5 General Government Expenditures - Functional Classification ............................................... 20 Table 2.6 Recent Trends In Social Expenditures .......................................................... 21 Table 2.7 Real Per Capita Social Sector Expenditures .......................................................... 22 Table 4.1 Aid In Regional Capital Budgets - FY96/97 - 99/00 .......................................................... 53 Table 5.1 Some Welfare Indicators Available In Ethiopia .......................................................... 78 Boxes Box 3.1 Information Requirements For Public Expenditure Management .............................................. 27 Box 3.2 Recurrent Costs, Dual Budgeting And Expenditure Strategy ..................................................... 29 Box 3.3 Strategy Documents For International Dialogue .......................................................... 33 Box 3.4 Budget Decentralization In Amhara Region .......................................................... 36 Box 3.5 Issues Arising From The Social Sector Development Programs ................................................ 42 Box 3.6 FIS, Interim FIS And Future FIS .......................................................... 43 Box 3.7 Summary Of Conclusions And Recommendations .......................................................... 44 Box 4.1 Aid Information Requirements For Public Expenditure Management ........................................ 50 Box 4.2 GoE-Donor Working Group On Aid Reporting And Forecasts ................................................. 52 Box 4.3 SNNPR Regional Budgets - The Story The Figures Tell .......................................................... 54 Box 4.4 The ADF Credit For ESDP Support .......................................................... 55 Box 4.5 Fungibility - What Difference Does Aid Make? .......................................................... 59 Box 4.6 Summary Of Conclusions And Recommendations .......................................................... 64 Box 5.1 Burkina Faso - Pilot Case On Conditionality Reform .......................................................... 73 Box 5.2 Mozambique - The PES And The Budget .......................................................... 74 Box 5.3 The International Development Goals .......................................................... 79 Box 5.4 Proxy And Lead Indicators .......................................................... 79 Box 5.5 Uganda - Innovative Work On Poverty Indicators .......................................................... 79 Figures Figure 2.1 Trends In Defense Expenditure .......................................................... 14 Figure 2.2 Share of Social Sector Expenditure To GNP Comparative Perspective .................................. 23 EXECUTIVE SUMMARY 1. This millennium Public Expenditure Review (PER) is the seventh in a continuous series of annual PERs for Ethiopia. In addition to the standard assessment of the fiscal situation, it has been the practice for PERs to address some thematic issues in public expenditure management of contemporary relevance and of mutual interest to the government and the donor community. The focus topics for this PER emerged out of consultations between Government of Ethiopia (GoE) and a core donor group at a workshop in Brussels (October 1999) organized by the European Commission. The workshop also resulted in a shared understanding that the PER must shift emphasis -from analysis to problem solving, and from the perspective of the federal government to the joint perspective of the federal and regional governments. This PER exercise has made a beginning in pursuing this understanding. 2. The broad contours of the context in which PER 2000 is being conducted are as follows: Economic context: The border conflict is exacting a progressively heavier fiscal toll. For some time, well into FY98/99, the government was able to contain the fiscal impact of the conflict by financing defense expenditures through budget contingencies, cuts in non-priority spending and extra-budgetary funds. These endeavors appear to have reached their fiscal limits as evidenced by significant recourse to domestic financing towards the end of fiscal year 1999 and through fiscal year 2000, and reduced capital spending, thus threatening government's objective of expanding basic social services. Relationships with aid agencies: The border conflict has strained short-run aid relationships, as exemplified most notably by the reluctance of many aid agencies to make fresh commitments while the conflict persists, and by the absence of a macroeconomic adjustment and reform program supported by the IMF. Beyond the immediate context of the conflict, however, Ethiopia has made commendable efforts to ensure that external assistance is integrated with its own development efforts. Sector programs in health, education and roads are shifting aid relationships away from a focus on donor-financed projects towards a broader dialogue between the government and the donors focused on key sector strategies which should, in course of time, expand both in breadth and depth to an effective discussion on the entire budget framework. The PER has an important role to play in supporting this envisaged expansion in scope and accelerating the transition from project specific aid to broad support for SDPs and/or free standing budget support aid. The PER provides an overall assessment of public expenditures that complements the sector-specific dialogues of the SDPs, and will be one of the main instruments for evaluating the 'track record' under the Highly Indebted Poor Countries (HIPC) debt relief initiative when Ethiopia's candidacy is revived. The PER is beginning to address the links between public expenditure and poverty reduction, and therefore will also be an important input for the preparation of the Poverty Reduction Strategy Paper (PRSP), which will replace the traditional Policy Framework Paper (PFP), and of the Comprehensive Development Framework (CDF). The Context For Identification Of Focus Topics 3. A major concern of previous PERs has been that Ethiopia should rapidly develop a medium-term framework for planning expenditures, so as to strengthen the links between policy formulation, planning and budgeting. This concern is shared by the government, and is reflected in their program to develop a Public Expenditure Program (PEP) for the programming of both recurrent and capital expenditures. The government's approach is first to develop a Public Investment Program (PIP) for the programming of capital expenditures: this is based on a Macroeconomic and Fiscal Framework (MEFF) which has a three-year perspective and presents the government with alternative scenarios for the balance between recurrent and capital expenditure and between federal expenditures and subventions to the regions. 4. The essence of a medium-term strategy on public expenditures is that there should be binding decisions at a high level (Council of Ministers and Parliament) on broad, multi-year allocations of expenditure before detailed annual budgets are prepared. The PIP/PEP approach provides for this - indeed it is required by law that the Council of Ministers and Parliament should approve such allocations. In practice, however, this has not so far occurred and the MEFF and PIP have remained shadow exercises. One argument is that uncertainty (about economic prospects, such as for example drought, and external events, and about likely flows of aid) makes medium-term planning impractical. This is perverse: a major reason for undertaking systematic medium-term expenditure planning is to address such uncertainty and the preparation of medium-term forecasts could be an important focus for dialogue with the donor community about aid commitments over the medium-term. PER 2000 re-emphasizes the importance of carrying through the government's commitment to more vigorous and strategic medium- term planning of expenditures. Towards this end, the focus topics for this PER, aimed at improving the strategic phase in the budget process, are the following: * Information systems for strategic expenditure management; * Aid management and integration of aid flows into the planning/budgeting process; * Public expenditure - shifting attention from inputs to results; and * PER process Review Of Fiscal Performance 5. Fiscal performance weakened during FY98/99, and projections, based on budget execution during the first four months of the current year (July-November 1999), point to further deterioration in the fiscal stance during FY99/00. Total revenues and external grants are projected to decline by 1.7 percent of GDP, notwithstanding an increase in tax revenues by 0.8 percent of GDP, owing largely to a 10 percent surcharge on imports introduced in December 1999. Non-tax revenues are expected to decline by 1.7 percent ii of GDP, after the high levels achieved through exceptional measures during FY98/99, while external grants are projected to decline by 0.8 percent of GDP. Overall expenditures are expected to decline by almost 1 percent of GDP, despite an increase in defense-related expenditures of 1.2 percent of GDP. The fiscal deficit after grants is projected to reach a level of 5.8 percent of GDP which, given lower expected external financing, would lead to an increase in domestic financing to 4.2 percent of GDP. 6. Consequent to the pre-emptive demands of the border conflict, defense related expenditures are projected to reach a level of 8.1 percent of GDP during FY99/00, up from 6.9 percent last year. This projection factors in the government's intention to cut defense expenditures by at least 50 percent during the last eight months of the -year, compared with the actual monthly average of the first four months. The increased defense expenditures have cut heavily into the government's capital expenditures as well as the recurrent portion of the social sector expenditures, thereby seriously threatening the government's development goals. Last year, the government was able to insulate, at least partly, the social sectors from the burden of adjustment. In the education sector, both capital and recurrent expenditures, as a proportion of GDP, were protected while in the health sector, only capital expenditures suffered a decline. During the current year, both capital and recurrent expenditures are projected to decline in both the sectors. The lost opportunities in terms of development forgone could multiply rapidly, the longer the conflict situation persists. 7. The current fiscal stance of the government is unsustainable for a number of reasons. The government's domestic borrowing is rising rapidly and the foreign exchange reserves are declining, thereby jeopardizing its commendable record of macroeconomic management in the mid-90s. Domestic financing of the fiscal deficit, at 4.2 percent this year, contrasts sharply with the discipline of the past when the government had very small or even negative domestic financing. If the monthly defense spending in the remaining eight months of the year is not cut by half as projected, the fiscal strains will intensify rapidly and result in even further rollback of capital expenditures and the more crucial recurrent expenditures in the social sectors. 8. The burden of the conflict is exacting a progressively heavier toll outside of the fiscal sector. The government's huge borrowing can potentially crowd out private investment, while the conflict related economic uncertainty has eroded investor confidence in general. The import cover provided by the foreign exchange reserves is projected to decline from 3.9 months last year to 3.2 months this year. During the first five months of FY99/00, the level of net foreign assets of the banking system has declined by over US$110 million. While this decline in net foreign assets has dampened the effects of the government's borrowing from the banking system on the monetary aggregates, it clearly points to the unsustainability of the current policy stance. Moreover, it also illustrates the unpromising prospects, in terms of rates of inflation and depreciation of the exchange rate, and ultimately, the level of investment and economic growth. iii Information Systems For Strategic Expenditure Management 9. The practical problems in information management indicated by the government included: (i) delays in receiving information needed for budget preparation; (ii) delays and inaccuracies in reporting expenditures; (iii) limited capacity to satisfy competing and overlapping demands for information; and (iv) weaknesses in horizontal and vertical communications between government agencies, and between government and aid agencies. Any approach to addressing the above problems should emphasize: (i) the need to look at both the demand and supply sides; (ii) awareness of the costs of generating and disseminating information; (iii) the information implications of decentralization; and (iv) dovetailing reforms in information management with reforms already under way as part of the Civil Service Reform Program. 10. The government has a strong sense of strategy, bottom-up procedures for defining priorities, and discipline in pursuing objectives. However, expenditure allocation is dominated by the task of fitting expenditures to available resources within a strictly annual horizon and this is done largely on a line item basis. Even for the capital budget, the system is geared to the implementation of most projects within a single fiscal year, and medium-term strategic issues are not rigorously addressed. At the federal level, there has been significant progress in developing the technical basis for genuine medium-term planning of expenditures (preparation of the MEFF and PIP, and development of a budget calendar that would allow strategic expenditure planning to precede annual budgeting each year). What is now required for moving forward is first, to give a statutory cover to the budget calendar and then to adhere to it and second, as financial legislation already requires, to make the setting of indicative planning figures the pivot of resource allocation each year. 11. Aid agencies tend to overload senior decision makers with demands for dialogue and the production of strategic planning documents, including the PIP/PEP, the Policy Framework Papers (now to be superseded by a Poverty Reduction Strategy Paper), the Comprehensive Development Framework, the various requirements for the Sector Development Programs (SDP), and the PER itself GoE itself undertakes a number of periodic strategic planning exercises, including the SDPs and, currently, the Five-year Plan. As next steps, the main aid agencies should rationalise their demands for dialogue with the government. They should put forward, for high level discussion with the GoE, consolidated proposals for a streamlined set of documents on which dialogue might focus, and propose a unified calendar for donor/government dialogue. This calendar should be synchronised with government's own planning and budgeting calendar, and would, inter alia, link the SDP, PER, PRSP and CDF processes. 12. Ethiopia's federal system has actively promoted the decentralization of planning and budgeting functions, and there are encouraging signs that this can lead to better use of resources. But effective medium-term planning and utilization of resources at the regional level is compromised: iv * by the federal government's failure to provide, even on a non-binding basis, indicative medium-term projections of the federal subsidy; such projections should be included in a timely MEFF each year; * by donor practices that inhibit the multi-year programming of aid flows; donors should review and reform such practices; and * by management arrangements for the SDPs that demand centralized information and decisions; the forthcoming mid-term reviews of ESDP and HSDP should pay particular attention to this issue, as well as to the synchronization of the SDPs with the regular planning system of Government. 13. The financial management system has remarkable strengths: budgeted funds are made available promptly, and fiduciary controls are reasonably effective. However, prompt and adequately detailed reporting of expenditures is an acknowledged weakness. The Expenditure Management and Control Program is focusing on remedying this through cost-center approach to budgeting and phased integration of recurrent and capital budgets. Aid Management 14. External aid accounts for nearly two thirds of the capital budget of the federal and regional governments put together, making aid management an important variable in overall public expenditure management. Despite recent innovations such as SDPs and some determined efforts by MEDAC to improve systems and procedures, many problems and imperfections persist. Resolving and reforming the system calls for concerted and coordinated action by the government as well as the donors. 15. The government's financial regulation requires that all external aid be accounted for in the consolidated fund. Yet a significant portion, often more than half of the total external aid, fails to be captured by the budget. Considerable uncertainty surrounds even the portion that flows through the budget. More reliable forecasts of how much aid will actually flow in and into which sectors is important at two levels: (a) so that the governments (at both federal and regional levels) can anticipate the total of aid and local resources that will be available and make appropriate decisions about their allocation; (b) so that the implementing agencies for specific projects have an early indication of sources of funding and can take into account any special conditions attached. Towards more systemically integrating aid flows into the planning and budgeting process, MEDAC should issue guidelines on which projects should be shown in the budget and how. The donors, on their part, can help by increasing the time horizon of their aid commitments and providing more accurate data on actual disbursements. 16. Smooth implementation of externally aided programs is often inhibited by the difficulties in fulfilling conditionalities and by the complexity of disbursement and implementation procedures. The move towards sector-wide approaches is meant to reduce and harmonize donor requirements, and existing constraints and administrative burdens could be further reduced by means of free standing budget support loans and v grants. This is a long term agenda, and not all donors will be able to move rapidly in this direction. A variety of actions are required, including: (a) further measures to simplify existing SDP disbursement and reporting arrangements, and to align them more closely with Ethiopia's own systems, which are generally working well, and whose application is within the reach of the administrative capacity available at the implementation level; (b) joint govemment - donor efforts to develop modalities for budget support and sector- focused aid that is not tied to specific projects; and (c) an urgent effort by donors to move to multi-year programming of their funds, so as to give implementers a longer horizon for organizing project implementation, and, at aggregate level, to increase the confidence with which macroeconomic projections can be made. 17. The budget offset system is intended to ensure inter-regional equity in the distribution of total resources for development. Although, in theory the intentions behind the offset system are rational, in practice it creates perverse incentives. First, in the. regions, there is already a decided preference for treasury funds, which are more reliable and timely and less demanding in terms of documentation and procedures, over extemal aid. The offset strongly reinforces this bias against aid flows. Consequently, the goal of maximizing the inflow of total extemal aid into the country is compromised. Second, the system gives an incentive to the regions (and at times to donors) not to declare expected aid, and to encourage channels of aid flow that are less likely to be factored into the offset calculation. The resolution of this complex issue would lie in basing the offset on recorded ex-post levels of donor disbursement, formulating clear guidelines on what will and will not included in the calculation, publishing the data on which the offset is calculated and maintaining the principle of offsetting less than 100 percent of the aid flow so as to expand the total resource envelope. 18. The mission noted that there is neither complete appreciation of the concems regarding fungibility of aid on the part of the govemment nor a total understanding of all the dimensions of the issue on the part of the donors. Fungibility concerns have come to the fore because of the border conflict, but arguably the concerns exist even beyond the immediacy of the present conflict. It is important that the concerns in this regard are discussed openly and honestly between the government and the donors possibly at an exclusive workshop within the framework of the PER process. 19. Food security in Ethiopia has three main components: supply, entitlement and emergency response. In order to ensure a more appropriate and timely donor response to emergency appeals, the government should make adequate and timely contributions to its own relief appeals. Chronic food insecurity and development related projects should not be addressed through emergency measures, but through programs - donor funded or otherwise - included in the annual budget. Finally, as a priority, the govemment should address the food security issue comprehensively, taking the current economic circumstances into account. 20. Aid relationships in Ethiopia are in a transitional phase. The government has taken a commendably strong lead in seeking to ensure that aid is properly integrated with other resources, and in developing sector-wide approaches that can avoid the weakness of vi traditional, fragmented project-focused aid. Adopting such a broader perspective requires a continuing dialogue between government and donors. Some of the mechanisms for such a dialogue have been put in place, particularly with respect to the Sector Development Programs. The PER itself is becoming an important mechanism for reviewing, and helping to resolve, issues that are more macroeconomic or systemic than those on which the SDP mechanisms principally focus. However, there tends to be insufficient appreciation by both donors and government of each other's concerns, and the mechanisms for dialogue between govemment and donors need to be better articulated. Establishment of a forum to fill existing gaps is an urgent imperative. The meetings of this forum should be built into the budget calendar and should aim at promoting a shared understanding on the relevant issues and perspectives at the appropriate stage in the planning/budgeting cycles. The mandate for this partnership dialogue could be widened, in the course of time, to a discussion on the overall budget framework and sector strategies. Public Expenditure - Shifting Attention From Inputs To Results 21. Increasingly Public Expenditure Reviews are moving from an analysis of spending pattems focused mainly on issues such as the sectoral composition of spending, or the balance between capital and recurrent spending, to a more precise assessment of the results of these programs. This reflects a trend in both developing and the OECD countries toward evaluating the benefits of public spending relative to its cost, as a means of improving both the overall cost-effectiveness of spending and the accountability of the public sector. Program evaluation is not new, but the conscious attempt by governments to assess and compare the returns on different programs by some common yardstick such as effectiveness in achieving key results is now gaining momentum. 22. Given the priority Ethiopia accords to reducing regional disparities in development and improving the welfare of the most disadvantaged groups in society, a shift in focus from inputs to results promises to be an area for fruitful dialogue. In attempting such a shift, Ethiopia has a good foundation to build on. The Education, Health and Roads Sector Development Programs specify performance indicators to be monitored and targets to be achieved under these indicators for the year 2001/02. In sectors not covered by SDPs though, monitoring of results is less advanced. In addition to the indicators used in the SDPs, indicators are also included in the development index of the government's budget subsidy formula used for distributing the budget subsidy horizontally across the regions. 23. Shifting from tracking expenditures (an input into the development process) to tracking what expenditures produce (outputs) and their impact (outcomes and impacts) is advisable for a number of reasons. The shift in focus can potentially (i) increase the productivity of public resources; (ii) become an effective instrument for enforcing the accountability of the public sector; (iii) enhance the allocative efficiency of public expenditure within and across sectors; and (iv) help in aligning external aid with the govemment's development strategy. vii 24. While some performance indicators are used in planning and budgeting in Ethiopia, their use could be widened and strengthened. There are several problems with current practice: (i) The planning and budgeting processes run on two parallel tracks right from the federal level to the zonal level with weak inter-linkages. Consequently, the use of output/outcome indicators that may take place in the planning process does not inform the annual budgeting exercise. (ii) The incremental basis on which budgets are prepared is not very amenable to factoring in past expenditure or the inter-sectoral linkages and tradeoffs. Several actions could be taken to improve the current practice. First, the ongoing Expenditure Management and Control Program has promising implications for linking inputs and outputs in public expenditure management. Second, beyond these reforms, indications of expected results should be provided together. with the resource envelope at all important stages of the planning and budgeting process. Third, the use of output and outcome indicators should be extended to sectors where this is not yet the case. 25. Beyond improving current practice, monitoring of results could be strengthened in several ways so as to increase the effectiveness of public expenditure in promoting welfare and alleviating poverty. A first step in strengthening the focus on results would be to define through a participatory process the overall results that the government wants to achieve in terms of improving the welfare of the people and the indicators chosen to monitor these results'. Second, it may be worth examining whether the current Welfare Monitoring System (WMS) provides sufficient information and whether there are cost- effective ways of learning more about results on the ground. A third step would consist of studying in some depth the linkages between public expenditures and the output and outcome/impact indicators selected to track results. Finally, it would be possible to use existing and new information as well as the results of these studies to assess whether the distribution of expenditures not just within one sector but across sectors is the most effective given the results the government wants to accomplish. 26. Donors can support these efforts to link public expenditures more closely to results in three ways: (i) they can provide technical assistance for pursuing the issues raised earlier and for developing output and outcome indicators in some of the leading sectors; (ii) they can provide technical assistance for building data bases; and (iii) they can help in capacity building at the federal and regional levels. PER Process 27. Starting with the PER in 1999, there was a conscious effort to turn the PER from being a product driven exercise to a continuous process. This effort was informed by the need to make the PER a more relevant and useful exercise for the government as well as the donors, to enhance government ownership of the process and to establish a basis for more effective dissemination of PER findings and recommendations. This process has strongly facilitated a policy dialogue between the government and the donors during an In fact, this is one of the prerequisites for preparing the PRSP. viii arguably strained period, while maintaining commendable transparency on the fiscal outcomes. Furthermore, the increasing participation of government officials and donor representatives in the PER process has contributed to better appreciation of each other's positions on a number of relevant issues. 28. The PER 2000 process further pursued efforts towards making the exercise more collaborative and participative. Most notably, the agenda for PER 2000 was discussed and finalized in consultation with the government, the PER working group comprising officials of MoF, MEDAC and donors' representatives met several times during the year to take stock of the PER process, the PER schedule has been advanced with reference to the planning/budgeting cycle so that it can become an effective input into the budget strategy, and finally, for the first time, participating donors have been systematically invited to review the draft PER within their institutions and feed their comments into the World Bank review process. 29. Notwithstanding the progress made as above, the PER process is still beset with several shortcoming and some dilemmas. There were concerns that government ownership of the process is still ineffective and incomplete, and that the whole exercise is still dominated by World Bank conventions and requirements. This lack of clarity inevitably impacts on the dissemination and implementation of PER recommendations, Furthermore, there is also no clarity on which recommendations are accepted, and on when and how they will be implemented. An interesting suggestion made was that the PER exercise should adopt the SDP model for review and monitoring - namely, a Joint Review Mission, independent of both the government and the donors, consideration of its findings at the Annual Review Meeting, and monitoring by a Steering Committee in the inter-mission period. This JRM-ARM model, adapted for the PER process, should also have an in-built institutional basis for involving the regions, starting from identification of the agenda, analysis of issues, dissemination and implementation of recommendations. Whether and how this model can be adapted to the PER process is an issue worth pursuing. ix 1. OVERVIEW AND CONTEXT INTRODUCIION 1.1 Ethiopia, among the poorest countries in the world, made remarkable progress in the mid-90s in fiscal consolidation and in reorienting public expenditure towards expanding basic social services. The challenge over the last two years has shifted to ensuring that the pre-emptive demands of defense expenditure on account of the border conflict with Eritrea do not jeopardize fiscal sustainability, and the much needed social sector spending. This Public Expenditure Review, the seventh in a continuous series since 1994, begins with a review of the fiscal situation. 1.2 In addition to the standard assessment of the fiscal situation, it has been the practice for PERs to address some thematic issues in public expenditure management of contemporary relevance and of mutual interest to the government and the donor community. The focus topics for this PER emerged out of consultations between GoE and a core donor group at a workshop in Brussels organized by the European Commission. The workshop also resulted in two other agreements. First, that the PER must shift emphasis from analysis to problem solving, and from the perspective of the federal government to the joint perspective of the federal and regional governments. Second, that the PER must shift focus from production of an annual report to a continuous process of dialogue and reform. This PER exercise has made a beginning in pursuing the above understanding. 1.3 This overview chapter (a) describes the context in which the PER has been conducted; (b) explains the common approach underlying the different topics; and (c) highlights some of the common themes that emerge. A final section summarizes the status of implementation of previous PER recommendations. Context 1.4 The way PER 2000 has been carried out has been permeated by three main contextual factors: * Economic context: The border conflict that erupted in May 1998 is exacting a progressively heavier fiscal toll. For some time, well into FY98/99, the government was able to contain the fiscal impact of the conflict by financing defense expenditures through additional non-tax revenues, budget contingencies, cuts in non-priority spending and extra-budgetary funds. These endeavors appear to have reached their fiscal limits even last year as evidenced by significant recourse to domestic financing towards the end of fiscal year 1999. The current year is witnessing even higher defense expenditures and even deeper cuts in capital expenditure and budget transfers to the regions, thus threatening the government's objective of expanding basic social services. Even as fiscal sustainability is threatened by the significantly larger domestic financing this year, the impact of the conflict is spilling into other sectors. The government's domestic borrowing has the potential to crowd out the private sector, monetary management is getting impaired and the balance of payments is coming under increasing strain. Decentralization: Ethiopia's recent Constitution has mandated a federal structure. Its institutions of federalism, however, are still evolving. As is characteristic of most federations, in Ethiopia too, there are vertical and horizontal imbalances: different regions have very different financial resource bases (the horizontal imbalance) and the regions' expenditure responsibilities outstrip their own revenue bases (the vertical imbalance), which makes them heavily dependent on fiscal transfers from the center. Ethiopia is still grappling with the issue of fiscal decentralization with a variety of practices and mechanisms for accelerating overall development while ensuring inter- regional equity. This PER is the first to make a systematic effort to extend its investigations to the regional level. * Relationships with aid agencies: The border conflict has strained short-run aid relationships, as exemplified most notably by the reluctance of many aid agencies to make fresh commitments while the conflict persists, and by the absence of a macroeconomic adjustment and reform program supported by the IMF. Beyond the immediate context of the conflict, however, Ethiopia has made commendable efforts to ensure that external assistance is integrated with its own development efforts. Sector programs in health, education and roads are shifting aid relationships away from a focus on donor-financed projects towards a broader dialogue between the government and the donors focused on key sector strategies which should, in course of time, expand both in breadth and depth to an effective discussion on the entire budget framework. The PER has an important role to play in supporting this envisaged expansion in scope and accelerating the transition from project specific aid to broad support for SDPs and/or free standing budget support aid. The PER provides an overall assessment of public expenditures that complements the sector- specific dialogues of the SDPs, and will be one of the main instruments for evaluating the 'track record' under the HIPC debt relief initiative when Ethiopia's candidacy is revived. The PER is beginning to address the links between public expenditure and poverty reduction, and therefore will also be an important input for the preparation of the Poverty Reduction Strategy Paper (PRSP), which requires, among other things, that public spending decisions be informed and assessed by their impact on poverty alleviation. 2 Approach 1.5 Recent analytical work has found it useful to evaluate public expenditure management by looking at how it impacts on three levels of outcomes: (i) aggregate fiscal discipline; (ii) strategic resource allocation; and (iii) operational efficiency. Ethiopia has traditionally had a strong commitment to aggregate fiscal discipline, and the government's effort at fiscal consolidation through the 90s, but for the conflict induced pressures, has been commendable. There is also a tradition here of preserving the integrity of the budget. Budget ceilings are honored and funded by the MoF and the Regional Finance Bureaus, while the approved ceilings are respected by the spending agencies as hard budget constraints. On the other hand, in Ethiopia, as in most countries, systematic attention to the second and third levels - allocative and operational efficiency and effectiveness of public expenditure - is relatively recent. This PER is the first to make the analysis of the results and impact of public expenditure one of its main topics. 1.6 A key insight is that the three levels of outcome are not independent of each other. Gains at one level often imply a cost at other levels (e.g. the measures applied to enforce aggregate discipline may easily lead to sectoral misallocations and operational inefficiencies). The processes of public expenditure management (PEM) are crucial in enabling potential conflicts between the levels to be reconciled - and these processes are as much political as technical. Experience from both OECD and developing countries suggests that a combination of top-down and bottom-up elements may be the most effective way to reconcile the three levels. Needs may be identified and priorities assigned in bottom-up fashion, but substantive budget decisions need to be made in the context of mutually consistent hard budget constraints: these have to be transmitted top- down, and set to conform to the requirements of aggregate fiscal discipline. In any such process, of course, there will be many iterations, but the essence of an effective PEM is that there should be binding decisions at a high level (Council of Ministers and Parliament) on broad, multi-year allocations of expenditure (Indicative Planning Figures or IPFs in Ethiopia's terminology) before detailed annual budgets are prepared. Moreover, capital and recurrent budgets should not be developed separately: capital and recurrent dimensions of expenditure for each sector and institution should be considered in an integrated fashion. 1.7 This general approach is accepted by the Government, and is reflected in GoE's Expenditure Management and Control Program (EMCP)2. Although previous PERs have included considerable debate about the best way to apply the appropriate design criteria in the Ethiopian context, the present PER assesses progress in developing the capacity for strategic expenditure planning, and also systematically considers the processes of public expenditure management in general (at both federal and regional levels) and of aid management in particular. Aims have been, first, to understand how the present systems work in practice (taking account of the innovations and reforms that are already under 2 See Chapter 3 and Appendix 2 (Volume II) for a description of the objectives of EMCP and the progress to date. 3 way), and, second, to identify practical improvements for consideration both by the government and by its aid partners. At a minimum, this approach should increase mutual understanding between government and donors. In practice, many pragmatic recommendations have emerged, including proposals for more systematic follow-up of agreed actions. 1.8 An important prerequisite for effective public expenditure management is accountability of the executive to broader constituencies such as the legislature, media, academia, business community and other organs of civil society. Aid relationships can undermine this accountability by fragmenting responsibility for budget outcomes and by substituting accountability for donor financed projects by accountability to stakeholder constituencies. On the other hand, aid relationships can enhance broader accountability by locating their aid in the context of the total budget strategy and catalysing discussion on overall outcomes. By providing an objective evaluation of public expenditure trends and outcomes, the PER can become a reference frame for enforcing such accountability. Areas Of Focus And Emerging Themes 1.9 The report is organised around the main topics agreed in the preparatory workshop and the concept paper. The broad questions addressed in each chapter are listed in order to give a preview of the content and emphasis of each topic. Review of fiscal performance (Chapter 2): What do the fiscal trends look like in a long term perspective? How was fiscal performance last year (FY98/99), and what are the projection for the current year (FY99/00)3? What has been the fiscal impact of the border conflict, and is the resultant fiscal stance of the government sustainable? How has the conflict impacted on the economic and functional composition of public expenditure? Information systems for strategic expenditure management (Chapter 3): What is the role of appropriate information at various stages in the planning and budgeting cycles? How can demand for information be rationalized, . and its supply improved? What are the information requirements arising out of Ethiopia's evolving framework of fiscal federalism, and how can systems be improved to facilitate effective medium-term planning and utilization of resources at the sub-national level? What are the implications of aid agencies' demands for information and how can this be streamlined so as to facilitate strategic planning of expenditures by the government? 3 The assessment of fiscal performance was based on govenunents latest pre-actual figures for FY98/99 and budget execution for the first four months of FY99/00, data for which were available at the time of the PER mission in Jan - Feb 2000. 4 Aid management and integration of aid flows into the planning and budgeting process (Chapter 4): What should be the institutional basis for government-donor dialogue and how can this interaction be encouraged to focus on broader strategy rather than enclaves of donor finance? What are the impediments to effectively integrating aid flows into the planning and budgeting process in general, and forecasting of aid flows in particular? How can problems in implementation, and in particular constraints in meeting project conditionalities, be resolved, or minimized? What are the merits and demerits of the budget offset system and how can it be improved? Under what circumstances is aid fungible, and how can mutual misgivings on the issue between the government and the donors be resolved? What are the issues in donor support for food security and what should be the respective approaches of the government and the donors? Public Expenditure - shifting attention from inputs to results (Chapter 5): Why is it important and relevant to shift attention from inputs to results in public expenditure management? What is the link between public expenditure and welfare? What is the status of measuring output/outcome indicators of public expenditure in key sectors? What is the way forward in designing performance indicators and output and outcome based budgeting? How can donors support these efforts? PER Process (Chapter 6): What has been the progress made in meeting the Special Partnership with Africa (SPA) guidelines on the PER process? What are the improvements effected in the PER 2000 process? What are the persisting shortcomings and deficiencies? How can the PER exercise be made more relevant and useful to the government and the donors? 1.10 Without going into the details of the recommendations and findings in the main chapters, it is nevertheless useful to draw attention to some common themes which also reflect the fruitfulness of the approach described in the previous section. Thus Chapter 2 documents how the costs of the border conflict are severely testing GoE's ability to maintain fiscal discipline, as well as the repercussions of the aggregate constraint on non- defense expenditures: allocations to social sectors have been severely affected, and so have subventions to the regions. In addition, new levels of uncertainly about resource availability are making it more difficult to achieve efficiency in short-term resource allocation and use (the latter point is further explored in Chapter 3). 1.11 Chapter 3 - following on from previous PERs' discussion of alternative models of medium-term expenditure planning - concludes that progress in this area now depends primarily on the Government moving forward quickly to put into effect the strategic expenditure planning system and fiscal calendar to which it has already made a legislative commitment. Both Chapter 3 (from the perspective of information management for strategic expenditure planning) and Chapter 4 (from the perspective of aid management) examine how existing budgeting and implementation processes work, and both, inter alia, highlight the need for a more co-ordinated dialogue between 5 government and donors. Both also pay attention to decentralisation, and highlight the need for the federal government and donors alike to review procedures that inadvertently hamper effective decentralisation and delegation of PEM responsibilities. Chapter 5 explores several themes which relate to operational efficiency and effectiveness as well as to strategic allocation of expenditures; it represents the first broad attempt to tackle the challenge of shifting from a focus on inputs (amounts spent) towards a focus on the results of public expenditure, and, like the previous chapters, identifies a number of steps for the way forward. Finally, Chapter 6 reflects on the PER process itself, and proposes ways of tailoring future PER work more precisely to meet the often diverging requirements of different stakeholders. Status Of Previous PER Recommendations 1.12 A checklist of previous PER recommendations and their current status of implementation is at Appendix 1. Issues and recommendations concerning the PER process are addressed in Chapter 6. The following is a summary of the status of implementation of other recommendations. 1.13 Public expenditure management reform: The last three PERs have all stressed the importance of developing a systematic medium-term framework for expenditure planning. At a minimum, GoE should press ahead with the program it has already adopted for the institutionalisation of the MEFF on which the federal Public Investment Program is to be based. This would require adoption of, and adherence to, the financial calendar developed by the budget reform design teams. The difficult political and macroeconomic situation over the last one year has inhibited strategic planning. There has, however, been continued progress with the basic components of budget reform (activities that fall under the Expenditure Management and Control Program are described in greater detail in Appendix 2). However, the financial calendar has not yet been made operational, and the PIP/MEFF have remained shadow exercises. During the recent ARM, donors have underlined that an acceptable macroeconomic framework and improved quality financial reporting and auditing standards are prerequisites for such a change in their implementation modalities. A joint macroeconomic working group involving MoF, MEDAC and the NBE has been set up, but the necessary information exchange and collaboration among these agencies has not yet become regular or rapid enough. Chapter 3 again reviews strategic expenditure planning requirements. 1.14 Aid management: PER- 1999 urged donors to collaborate in providing more reliable and timely information on prospective aid flows, and to plan their aid over a longer horizon. In the event, aid relationships have been dominated by the shadow of the border conflict, which (as Chapter 2 charts) has further inhibited aid. A joint GoE/donor committee to work on the improvement of aid data was recommended by PER 1999. At the time of the PER 2000 field work, this had been agreed to but was yet to be convened. Specific recommendations for multilateral donors to review their SDP implementation modalities have not so far resulted in any visible changes, although both the SDP Joint Review Missions and the present PER have found this to be a continuing concern. During the recent ARMs, donors have underlined that an acceptable macroeconomic 6 framework and higher quality financial reporting and auditing standards are prerequisites for such a change in their implementation modalities. A regular topic of previous PERs was the offset system, which was considered to be a significant disincentive to the regions' uptake of aid. For the current year, there has been a major change, with MEDAC now offsetting only 30 percent of grants and 80 percent of loans. This addresses the biggest problem with offset, and should allow the discussion to move forward to broader issues (see Chapter 4). 1.15 Federal-regional fiscal relations: During the year under review, the resource demands of the border conflict substantially reduced the funds made available to the regions, and thus overshadowed discussion of larger issues in fiscal federalism. The revision of the offset system should work to the regions' advantage, but they are yet to benefit from an earlier notification (even indicative) of the federal subsidy which is required by the financial calendar (see Chapter 3 for more detail). 1.16 Allocation of expenditures: Previous PERs recommended (a) increasing the budget shares for education and health; (b) increasing the share of O&M expenditures; and (c) conducting a simultaneous review of the sustainability and mutual consistency of the various sector programs. Regarding (a) and (b), as Chapter 2 describes, increasing defense allocations have squeezed the social sectors to the point that per capita spending has fallen, while O&M expenditures have also been squeezed. The effective shrinking of the resource envelope for non-defense activities magnifies the concern about the fiscal sustainability of the sector programs, but, with the MEFF remaining a shadow exercise, GoE has not properly addressed the issue. Thus the roads SDP is continuing on an implementation path which, even without the border conflict, was set to crowd out social sector programs at the regional level. To go beyond identifying this problem to the stage of beginning to rectify it is a continuing challenge - both for the PERs and for GoE's own medium-term planning. 1.17 Revenue mobilisation: The 1997 and 1998 PERs examined revenue collection issues in some detail and made some quite specific and technical recommendations (e.g. for strengthening of tax and customs administrations). Implementation of such measures is under way. However, neither PER 1999 nor PER 2000 has really focused on revenue issues. It may be appropriate to revisit the issue in the context of PER 2001, especially from a PRSP perspective. The task should be to project the resource availability over the medium-term for implementing the poverty reduction strategy after accounting for non- discretionary expenditure both at the federal and regional levels. 7 2. REVIEW OF FISCAL PERFORMANCE 2.1 Fiscal management all through the year was dominated by the need to finance burgeoning defense expenditures and to minimize the burden of adjustment on other sectors. In the second year running since it erupted in May 1998, the border conflict is exacting a progressively heavier fiscal toll. For much of FY98/99, the government was able to contain the fiscal impact of the conflict by financing defense expenditures through increased non-tax revenue, budget contingencies, cuts in non-priority spending and extra- budgetary funds. Whatever be the source of financing, the opportunity costs of coriflict- related expenditures, in terms of foregone development impact, are huge. Besides, even these endeavors reached their fiscal limits last year itself as evidenced by the significant recourse to domestic financing by the end of FY99/00. The current year is witnessing even higher defense spending, even deeper cuts in capital expenditures and budget transfers to the regions, and substantially increased domestic financing. This fiscal stance is evidently eroding the gains made by the government's commendable macroeconomic management in the mid-90s and jeopardizing its basic development objective of improving and expanding basic social services. A reversal of the current fiscal stance, especially through adjustment of defense expenditures, is urgently warranted. 2.2 The layout of the chapter is as follows. The next section will analyze the fiscal trends since FY86/87 and relate them to the government's policy stance, especially on defense expenditures. The following section will review the fiscal performance in FY98/99 (based on latest preliminary actual figures) and the projected performance during the current year based on the budget execution during the first four months of the fiscal year. The section also assesses risks and presents alternative scenarios for the fiscal outturn. An assessment of the impact of the conflict, covering both direct fiscal costs and second round effects, is the focus of the next section. This is followed by a brief evaluation of the current fiscal stance of the government. The chapter concludes with analysis of the economic and functional classification of public expenditure, including some inter-temporal and inter-spatial comparisons. FISCAL TRENDS 2.3 For the purpose of analyzing the fiscal trends, the period FY86/87-99/00 can be sub-divided into four distinct sub-periods: late civil war (86/87 - 90/91), stabilization (91/92 - 94/95), reconstruction (95/96-97/98) and border conflict with Eritrea (98/99- 99/00) (Table - 2.1). There was a steady reduction in the fiscal deficit, both before and after grants, over the first three sub-periods evidencing the government's strong commitment to fiscal discipline. This trend was, however, reversed during the last sub- period (98/99-99/00), largely as a result of the fiscal impact of the conflict. Table 2.1: Ethiopia - Fiscal Trends, 1986/87-1999100 . . ......................................... . .................................................... 1986/87 1991/92 1995/96 1998/99 (as a percentage of GDP) -1990/91 -1994/95 -1997/98 -1999/00 1997/98 1998/99 1999/00 Actual Actual Pre.Act. Pre.Act. Pre. Act. Pre. Act. PER Proj. Total revenue and grants 24.6 16.5 21.8 20.6 21.5 21.5 19.8 Total revenue 21.0 13.7 18.7 18.7 18.7 19.1 18.2 Tax revenue 14.3 9.6 12.4 11.7 11.7 11.3 12.1 Non-tax revenue 6.8 4.1 6.3 7.0 7.0 7.8 6.1 External grants 3.6 2.8 3.1 1.9 2.8 2.3 1.5 Total expend. (includingnet 31.7 22.4 25.4 26.0 25.2 26.4 25.6 lending) Of which: Recurrent expenditure 22.1 14.9 14.7 18.3 15.7 17.6 19.0 o/w Defense 9.6 2.7 3.0 7.5 4.9 6.9 8.1 Capital expenditure 9.6 7.5 9.7 7.7 9.2 8.7 6.6 Net lending 0.0 0.0 1.0 0.0 0.2 0.0 0.0 Fiscal balance (cash basis) After grants -7.1 -5.9 -3.6 -5.4 -3.6 -4.9 -5.8 Before grants -10.7 -8.7 -6.7 -7.3 -6.5 -7.3 -7.4 Financing 7.1 5.9 3.6 5.4 3.6 4.9 5.8 External (net) 3.0 3.2 2.4 2.2 1.7 2.7 1.7 Domestic 4.0 3.7 0.2 3.2 1.3 2.2 4.2 Banking system 3.8 3.5 -0.3 2.6 1.3 1.0 4.2 Other and residual 0.1 -1.0 1.0 0.0 0.6 0.0 0.0 -...-.I.-...... ........ ....... . ........... .... . ....... ..... ..... _.. . ........ ......... ..... ........ -.... ..... ........ ............ .-.... ... .. ............. ..-_.'.._ ... .. .......... ... .... ... . . ........ .... .. . ..... ..... ..... .... ...... . .. ........ . .................. ............ Memo items: Capital expenditure by sources 100.0 100.0 100.0 100.0 100.0 100.0 100.0 of financing Domestic treasury 50.5 60.8 71.6 52.3 63.0 54.9 49.8 External assistance 11.6 10.0 6.5 12.5 11.9 12.3 12.7 External loan 37.9 29.2 22.0 35.2 25.0 32.8 37.5 Domestic borrowing/fiscal 53.1 61.6 -8.4 58.3 36.1 45.2 71.4 deficit (after grants) Reservesinmonthsofimports N/A 5.3 7.1 3.6 4.3 3.9 3.2 Source: Ministry of Finance, World Bank Database, Projection for 1999/00 by the PER Mission (February 2000), MEDAC and National Bank of Ethiopia 2.4 The dominant factors responsible for fiscal improvement varied across the first three sub-periods. Between the first two sub-periods, both total revenue and total expenditure declined as a percentage of GDP, but expenditure declined more than revenue, owing to the transition from war to peace. Consequently, fiscal deficit before grants declined from an average of 10.7 percent of GDP during 86/87 - 90/91 to 8.7 percent during 91/92 - 94/95. Similarly, the deficit after grants declined from 7.1 percent of GDP to 5.9 percent of GDP. Between the second (91/92 - 94/95) and the third (95/96 - 97/98) sub-periods, both total revenue and total expenditure registered an increase as a percentage of GDP, but revenue increased faster than expenditure yielding a reduction in 10 fiscal deficit from 8.7 percent of GDP to 6.7 percent before grants and from 5.9 percent to 3.6 percent after grants. During 95/96-97/98, the government's borrowing from the banking system was -0.3 percent of GDP, evidencing its determined effort at fiscal consolidation. This fiscal discipline was, however, impaired during the current period of border conflict when defense expenditures increased from 3.0 percent during 95/96 - 97/98 to 7.5 percent during 98/99 - 99/00. As a result, the fiscal deficit before grants is projected to increase to 7.3 percent of GDP and that after grants to 5.4 percent of GDP. FISCAL PERFORMANCE DuRING 1998/99 AND PRoJEcTIoN FOR 1999/00 Fiscal Performance During 1998/99 2.5 Fiscal performance weakened during FY98/99 reversing the improving trend of the earlier period. Total revenue and grants remained at 21.5 percent of GDP, the same as in the previous year, despite declines in tax revenues and external grants. This was achieved through a number of exceptional measures, such as transfers from extra - budgetary funds and privatization receipts, which yielded an increase in non-tax revenues of 0.8 percent of GDP. Overall expenditures, however, increased by 1.2 percent of GDP, led largely by an increase of 2 percent of GDP in defense-related expenditures to a level of 6.9 percent of GDP. The fiscal deficit after grants rose by 1.3 percentage points of GDP to reach 4.9 percent of GDP, and led to a level of domestic financing of 2.2 percent of GDP, up from 1.3 percent of GDP during FY97/98. Projection On Fiscal Outcomes For 1999/2000 2.6 Projections for FY99/00, based on the budget execution during the first 4-5 months of the year, point to further deterioration of the fiscal stance. Total revenues and external grants are projected to decline by 1.7 percent of GDP, notwithstanding an increase in tax revenues by 0.8 percent of GDP, largely owing to a 10 percent surcharge on imports introduced in December 1999. This surcharge is expected to yield additional tax revenues equivalent to 0.7 percent of GDP during the period January-June 2000. Non-tax revenues, however, are expected to decline by 1.7 percent of GDP, after the high levels achieved through exceptional measures during FY98/99, mainly owing to decline in transfers from (a) public enterprises, (b) sugar auction sales, and (c) fuel stabilization fund. In addition, external grants are projected to decline by 0.8 percent of GDP. 2.7 Overall expenditures are projected to decline by 0.8 percent of GDP, from 26.4 percent in FY98/99 to 25.6 percent of GDP in FY99/00. This is despite an increase of defense expenditure by 1.2 percentage points of GDP. Consequently non-defense expenditures are expected to decline by 2 percent of GDP during FY99/00 to a level of 17.5 percent of GDP. Since non-defense current expenditures are projected to increase by 0.2 percent of GDP, the burden of adjustment is falling largely on capital expenditures which are expected to decline from 8.7 percent of GDP in FY98/99 to 6.6 percent of GDP in FY99/00, the lowest in recent history. Particularly worrisome is the cut in budgetary subventions to the regions, both on revenue and capital accounts. Between the preliminary actual outcomes of FY98/99 and budget estimates of FY99/00, recurrent 11 transfers are projected be cut by 6 percent and capital transfers by as much as 40 percent (Appendix Table 9). 2.8 Since total revenue is declining more steeply than total expenditure in FY99/00, the fiscal deficit is projected to worsen, reaching 7.4 percent of GDP before grants and 5.8 percent of GDP after grants. Net external financing (loans) is expected to decline by 1 percent, to a level of 1.7 percent of GDP during FY99/00. This, together with the expected higher deficit after grants, is expected to result in an increase in domestic financing from 44.8 percent to 72.4 percent of the total fiscal deficit, and in terms of GDP, from 2.2 percent of GDP in FY98/99 to 4.2 percent of GDP in FY99/00. This level contrasts sharply with the very small or even negative domestic financing observed during 1995-98. Moreover, as will be explained in the next sub-section on risks/alternative scenarios, there are several factors that may lead to an even higher recourse to domestic financing than projected under this baseline scenario. Risks/Alternative Scenarios For 1999/2000 2.9 There are several risk factors that could potentially weaken fiscal performance during FY99/00 beyond that projected under the baseline scenario presented above. Three possibilities would be higher defense-related expenditures, higher central treasury financing of expenditures, and a lower GDP growth rate than assumed under the baseline projection. 2.10 First, the projection on defense spending is based on the intention of the government, as indicated to the PER mission, to implement sharp cuts in average monthly expenditures during the last eight months of the year, compared to actual average spending during the first four months of the year. Under the alternative scenario that average monthly expenditures are reduced by only 25 percent, instead of the 50 percent assumed in the baseline projection, defense spending would exceed 10 percent of GDP. This would mean additional expenditures of about 2 percentage points of GDP, which would most likely be financed through additional domestic financing. In such a case, domestic financing could exceed 6 percent of GDP. 2.11 Second, the baseline projection also factors in the government's intention to reduce central treasury financing for capital expenditures from 4.8 percent of GDP last year to 3.3 percent of GDP during FY99/00. Given that the projection implies significant reductions in capital expenditures, including in health and education, it is possible that in an effort to protect these expenditures, government will resort to further domestic financing. While protecting these expenditures is important, and even critical in a country like Ethiopia, this should be done through neutralizing cuts in other expenditures rather than further domestic financing. Increased domestic financing would not only potentially crowd out private investment, but will also increase the debt servicing burden in future years. 2.12 Third, the government is projecting a GDP growth rate of 7.3 percent during FY99/00, up from 6.3 percent during FY98/99. This projection is based, to a large 12 extent, on the results of a recently completed survey/forecast, which estimates a 7.8 percent growth rate in agricultural production during the Meher season. This projected high growth rate is inconsistent with the government's recently launched food aid appeal to the donor community. The authorities explained that the regions in need of food aid and those with high agricultural growth rate are generally not the same. Nevertheless, it is the view of the PER mission that a lower GDP than that envisaged by the authorities is a distinct possibility. Under such a scenario, revenues would likely be lower and domestic financing needs higher, than under the baseline projection. 2.13 Additional data that became available after the PER mission in Jan-Feb 2000 indicates that the fiscal outturn was considerably worse than that indicated by the above projections and risk assessment. There were two major developments subsequent to the mission. First, the border conflict erupted into full scale military operation for over six weeks in May-June, 2000 fuelling further escalation of defense spending. Second, the drought in the southern and eastern parts of the country reached crisis proportions necessitating relief and rehabilitation on a scale much larger than initially anticipated. In spite of substantial international assistance for relief and rehabilitation of as much as 1.5 percent of GDP, responding to the drought put severe pressure on government's own resources. Although precise figures are not available, it is estimated that defense spending would have gone well beyond the projection of 10 percent of GDP in the worst case scenario above. Also, relief and rehabilitation expenditure is estimated to be as much as 8 times the budgeted figure. The fiscal stress emanating from these factors is estimated to have driven up the fiscal deficit to over 13 percent before grants, and around 10 percent after grants. These developments also necessitate a significant downward adjustment in growth projection to around 5 percent. FISCAL IMACT OF THE BORDER CONFLICT Direct Fiscal Costs 2.14 Benefiting from the peace dividend following the end of the civil war in mid - 1991, defense expenditures dropped substantially from over 8.8 percent of GDP in FY90/91 to about 2.5 percent of GDP during FY91/92-96/97. Since the outbreak of the border conflict in May 1998, this downward trend has been reversed as preliminary actual defense outturn jumped sharply to 4.9 percent of GDP in FY97/98 and further on to 6.9 percent of GDP in FY98/99. Defense expenditures are projected to rise even further to reach a level of 8.1 percent of GDP during FY99/00 (Figure 2.1). This projection factors in the government's intention to cut military expenditures by at least 50 percent during the last 8 months of the year, compared with the actual monthly average of the first four months (as explained in para 2.10). The PER mission was informed that this reduction was possible since the high levels of defense expenditures in the early months of FY99/00 contained a large proportion of one-time exceptional expenditures. Under the assumption that in a sustainable medium-term fiscal framework, defense-related expenditures would not exceed 2-3 percent of GDP, the additional cost during FY99/00 would be around 5-6 percent of GDP. 13 Figure 2.1: Trends In Defense Expenditure |~~~~~~~~~ 35 35 ~ .,... .--Defenses pendirg as share of total 3 ~~.. ~~ expend. 30 D,~ efense spendirg as shareof GDP 30 25 25 8 20 20 10 - j - 10 5 ~~~~~~~~~~~~~~~~~~~~~~~~~~~5 0* 0 86187 87/88 88/89 89/90 90/91 9792 92/93 93/94 94195 95196 96/97 97/98 98/99 99/00 F is cal year Source: Ministry of Finance, World Bank Database and Projection tor 1999/00 by the PER M-ission (Febnriy 2000). 2.15 The originally approved appropriation for defense for FY99/00 was Binr 2.5 billion. In addition, the budget contained an appropriation of Birr 800 million for contingencies. In the event, given that defense expenditures during the first four months of the year exceeded Birr 2.2 billion, the contingency line is likely to have been utilized as early as in November or December 1999. Consequently, a supplementary budget appropriation is expected to be sought from the Parliament raising the allocation for defense from the original Birr 2.5 billion to Birr 4.2 billion. Unlike last year, during the current year, the entire burden of increased defense spending is being bone by domestic financing which, as noted earlier, is rising from 2.2 percent of GDP last year to 4.2 percent this year. 2.16 By comparison, Ethiopia's projected defense spending exceeding 8 percent of GDP is clearly above the average defense spending in Sub-Saharan Africa of 2.8 percent of GDP in 1995. 2.17 The increased defense expenditures, coupled with the sharp decline in non-tax revenues, have cut heavily into the goverment's capital expenditures as well as the recurrent portion of the social sector expenditures, thereby strongly threatening the goverment's development goals. Capital expenditures are projected to decline by over 2 percent of GDP during FY99/00 to a level as low as 6.6 percent of GDP, which compares unfavorably with levels exceeding 10 percent of GDP as recently as in FY96197. A big factor in this decline is the drop in extenal financing (loans and grants) from 5.0 percent 14 last year to 3.2 percent this year. The government has also budgeted lower treasury funds for capital expenditure. It is not possible to draw an unambiguous conclusion about whether it is the lower external financing that has led to lower budgeting of treasury funds or the other way around. During the discussion of the aide-memoire of the PER mission, the government did indeed argue that they would have been able to come up with the requisite counterpart financing had the donors maintained aid flows at the normal levels. Given the fiscal constraints and the substantial domestic borrowing that the government had to resort to, it is difficult to countenance this argument and believe that the government would have been able to increase treasury financing to absorb higher levels aid. 2.18 Last year, the government was able to insulate, at least partly, the social sectors from the burden of adjustment. In the education sector, both capital and recurrent expenditures, as a proportion of GDP, were protected while in the health sector, only capital expenditures suffered a decline [see Table 2.6]. During the current year, both capital and recurrent expenditures are projected to decline in both the sectors. Even more tellingly, in real per capita terms, expenditure on education has stagnated since the conflict broke out two years ago,. while expenditure on health has actually declined [see Table 2.7]. The lost opportunities in terms of development foregone could multiply rapidly, the longer the conflict situation persists. Second Round Effects Of The Conflict 2.19 The second round effects of the conflict fall into two categories: first, direct fiscal costs, mainly by way of expenditure on relief and rehabilitation to the displaced and returnees, and second, the adverse impact of the conflict on the macroeconomic situation. Assistance To Displaced And Returnee Population 2.20 Discussions with the Disaster Prevention and Preparedness Commission (DPPC) indicated that 405,245 people (displaced from the war zones and returnees from Eritrea) needed assistance in FY98/99 while the corresponding estimated figure for FY99/00 is 350,000. In FY98/99, some 47,335 metric tons of food aid (about 58 percent of the total requirement) was distributed to the affected population together with some non- quantified amounts (in monetary terms) of non-food aid in the form of shelter, clothing and drugs. PER 1999 estimated the annual cost of assistance to the displaced population at about Birr 550-950 million (US$ 70-120 million), depending on the severity and duration of the border conflict. Analysis for FY99/00, based on the estimated 350,000 displaced population, shows a lower projected assistance figure of Birr 281 million (US$ 35 million). The analysis was based on projected food aid requirement of 62,131 MT (estimated at Birr 106 million), food transportation cost amounting to Birr 28 million and 15 non-food aid amounting to Birr 147 million.4 The estimated assistance represents about 0.5 percent of GDP for FY99/00. Cost Of Re-Routing Road Freight Via Djibouti 2.21 According to the Ministry of Transport and Communication, the cost of re-routing road freight to Djibouti has resulted in a 31 percent increase in the cost of road transport via the Djibouti corridor compared with the Assab corridor. This increase is due to such factors as longer distance of 43 km, rougher road on the Djibouti corridor resulting in higher vehicle operating costs, lower vehicle turnaround per month (3 as against 3.5 for Assab), and demand pressure relative to supply of trucks suitable for long haul traffic.5 The additional estimated costs of road freight on the Djibouti corridor for both FY98/99 and FY99/00 represent 0.8 percent of GDP respectively (Table - 2.2).6 Table 2.2: Estimated Additional Cost Of Re-Routing Road Freight Via Djibouti FY98/99 FY99/00 Actual Projection* Distance (Krn) 910 910 Total freight (million MT) 3.26 3.46 Total ton/km (million)** 2,967 3,149 Unit cost (Birr/ton/km)*** 0.43 0.44 Total cost (Birr million) 1,276 1,386 Additional cost compared with Assab corridor (31%) 396 430 (Birr million) % of GDP 0.8 0.8 Notes: * PER projection based on freight data for six months of the fiscal year. * Total of dry and wet cargo. * Average of peak and low season rates. Source: Ministry of Transport and Communication and Road Transport Authority 2.22 Other additional costs at Djibouti Port compared to Assab Port, which are difficult to quantify without detailed freight data on individual shipments, include: lower port handling charges for bulk cargo, containers and spare parts; higher port handling charges for iron bars and motor vehicles; higher cost of pilotage while moving general cargo to containers; and higher cost of temporarily stocking empty containers. On the cost side, at the Djibouti port, there is a user fee of US$ 1 per ton. There is also an additional cost to the Ethiopian Shipping Line in terms of lost earnings from the Eritrean freight traffic due 4 The bulk of the food aid is usually grains (mostly wheat and maize). The estimated cost of one metric ton of wheat according to the World Food Organization is US$ 210 or Birr 1,709 at the current exchange rate. Estimates for the cost of food transportation and non-food aid are based on DPPC estimates in their document: 'Assistance requirements in 2000 for internally displaced people', Volume 2 of 3, January 2000. 5 The Ministry of Transport and Communication provided data on road freight, while the Road Authority provided the unit cost per ton km. 6 It should be recognized though that part of the additional cost in FY99/00 is on account of increased tonnage. 16 to the conflict. While some of the port handling costs may neutralize one another, on balance, the analysis shows increased transportation costs borne by shippers which are ultimately passed on to consumers by way of higher prices. Slow Down In Private Investment 2.23 PER 1999 concluded that the border conflict has had a dampening effect on investment owing to the confidence factor. To further investigate this issue, annual data on the volume of private investment (domestic and foreign) since 1992 provided by the Ethiopian Investment Authority were analyzed (Table - 2.3). The analysis reveals that, after an encouraging upward trend up to FY97/98, there was a substantial drop in the volume of approved investment from Birr 9.9 billion to Birr 5.1 billion in FY98/99. There has also been a slow down in translating approved investment proposals to operational stage, in spite of an increasingly liberal investment code, as potential investors seem to be adopting a wait-and-see attitude. Data for the first three quarters of FY99/00 also indicate a similar declining trend, thus underscoring the dampening effect of the conflict on domestic investment. Table 2.3: Private Investment - Approvals And Implementation - Approved investment Commenced operation Fiscal year No. Bir Million No. Birr Million 92/93 545 3,982 25 148 93/94 525 3,364 78 493 94/95 691 5,299 227 1,643 95/96 907 6,485 304 842 96/97 794 6,715 233 2,401 97/98 897 9,925 290 2,871 98/99 704 5,145 313 1,120 99/00 * 454 6,233 60 454 Total 5,517 47,147 1,530 9,971 Source: Ethiopian Investment Authority * 1999/00 data is for the first three quarters of the fiscal year. SUSTAINABILITY OF CURRENT FIscAL STANCE 2.24 The current fiscal stance of the government is unsustainable for a number of reasons. The government's domestic borrowing is rising rapidly and the foreign exchange reserves are declining, thereby jeopardizing its commendable record of macroeconomic management in the mid-90s. Domestic financing of the fiscal deficit is projected to increase by fill 2 percentage points and rise from 2.2 percent of GDP in FY98/99 to 4.2 percent this year. This contrasts sharply with the discipline of the past when the government had very small or even negative domestic financing. If the monthly defense spending in the remaining 8 months of the year is not cut by half as 7 It is however recognized that factors other than the border conflict, for example the difficulty of land acquisition in many regions notwithstanding the new land lease policy, and poor economic and social infrastructure, may also have contributed to the poor investment response. 17 projected, the fiscal strains will intensify rapidly and result in even further rollback of capital expenditures and the more crucial recurrent (mostly non-wage O&M) expenditures. (Refer to earlier para 2.13 for a prognosis of the fiscal situation.). 2.25 The burden of the conflict is exacting a progressively heavier toll outside of the fiscal sector. As explained above, the government's huge borrowing can potentially crowd out private investment although it should be noted that the banking sector in Ethiopia has traditionally had excess liquidity and the government's borrowing may not have any crowding out effect yet. The conflict related economic uncertainty has eroded investor confidence in general. During July-October 1999, the excess reserves of the commercial banking system declined substantially with the excess reserves of the Commercial Bank of Ethiopia (CBE), by far the dominant bank, declining by as much as half. The import cover provided by the foreign exchange reserves is projected to decline from 3.9 months last year to 3.2 months this year. During the first five months of FY99/00, the level of net foreign assets of the banking system has declined by over US$ 110 million. While this decline in net foreign assets has dampened the effects of the government's borrowing from the banking system on the monetary aggregates, it clearly points to the unsustainability of the current policy stance. Moreover, it also illustrates the unpromising prospects, in terms of rates of inflation and depreciation of the exchange rate and, ultimately, the level of investment and economic growth. ECONOMIC AND FUNCTIONAL CLASSIFICATION OF EXPENDITURE Economic Classification 2.26 The economic classification of expenditure changed substantially during the two year period 1998 - 2000, largely as a result of the sharp increase in defense expenditures (Table - 2.4). The recurrent-capital expenditure balance shifted significantly against capital expenditure. The share of recurrent expenditure in total expenditure increased from 63 percent in FY97/98 to 67 percent in FY98/99 and is projected to further increase to over 74 percent during FY99/00. The share of wages and salaries declined from 23.6 percent in FY97/98 to 22.6 percent in FY98/99, but is projected to rise to 25.2 percent in FY99/00. However, defense salaries and wages registered a secular increase in share, going up from 4.6 percent in FY97/98 to 6.2 percent in FY 98/99 and are projected to rise further to 7.4 percent in FY99/00. The proportion of materials and supplies, again driven by increased defense expenditures, rose steeply from 25.4 percent in FY97/98 to 31.0 percent in FY98/99 and is expected to further rise to 33.2 percent in FY99/00. A particularly worrisome trend is the rising share of interest and charges. 18 Table 2.4: General Government Expenditures - Economic Classification Share of total expenditure (in percent) 1997/98 1998/99 1999/00 Pre. Actual Pre. Actual. PER Proj. Tota expenditure 100.0 100.0 100.0 Recurrent expenditure 63.1 66.9 74.4 Wages and salaries 23.6 22.6 25.2 o/w Defense 4.6 6.2 7.4 Materials and supplies 25.4 31.0 33.2 Grants, contr. trans. & other cur. transfers 2.5 2.6 2.9 Pensions 2.7 1.8 2.3 Subsidies 0.0 0.0 0.0 Interest and charges 7.4 7.4 8.2 External assistance 1.4 1.5 2.6 Capital expenditure 36.9 33.1 25.6 Capital expenditure / total (non-defense) expenditure 45.9 44.9 37.5 Source: Ministry of Finance, World Bank Database and Projection for 1999/00 by the PER Nission (Febnuary 2000). 2.27 Capital expenditures have inevitably yielded to recurrent expenditures, their share dropping from 36.9 percent in FY97/98 to 33.1 percent in FY98/99 and projected to drop even further to 25.6 percent in FY99/00, the lowest in recent period. Excluding defense expenditure, the share of capital expenditures remained relatively stable at around 45 percent in FY97/98 and FY98/99, but is expected to decline to a proportion as low as 37 percent in FY99/00. These trends lead to two conclusions: first, defense expenditures have crowded out public investment and second, even in the non-defense segment, the ratio is shifting away from capital and toward recurrent expenditures. Functional Classification 2.28 Increased defense expenditures have also altered the functional classification of public expenditures quite dramatically during the two year conflict period 1998-2000. From a proportion of 20 percent of total expenditures of the general government during FY97/98, defense-related expenditures increased to over 26 percent during FY98199 and are expected to further increase to close to one-third (Table - 2.5). The increase in the share of defense-related expenditures is even more dramatic with respect to recurrent expenditures. In this case, defense-related expenditures increased from 31 percent of recurrent expenditures during FY97/98 to 39 percent during FY98/99 and are expected to rise to 43 percent during FY99/00. 19 Table 2.5: General Government Expenditures - Functional Classification In Percent 1997/98 1998/99 1999/00 Pre. Act. Pre. Act. PER Proj. Share In Total Expenditure General Administration 29.6 36.8 42.1 o/wDefense 19.5 26.2 31.6 Economic Infrastructure 12.7 12.3 9.8 o/w Road Construction i/ 10.5 10.4 7.2 Economic Services 17.2 16.0 13.4 o/w Agriculture & Natural Resources 11.7 11.9 10.3 Social Services 25.4 22.5 19.9 o/w Education 14.1 13.2 12.5 Health 6.6 5.7 5.2 Others 15.1 12.3 14.9 o/w interest & charges 7.4 7.4 8.2 external assistance 1.4 1.5 2.6 Share In Recurrent Expenditure General Administration 46.9 55.0 56.6 o/w Defense 30.9 39.2 42.5 Economic Infrastructure 1.5 1.7 1.6 o/w Road Construction 1/ 1.3 0.9 0.8 Economic Services 7.8 7.3 6.3 o/w Agriculture & Natural Resources 6.7 6.1 5.2 Social Services 24.1 21.8 17.9 o/w Education 15.7 14.3 12.6 Health 5.6 5.2 4.1 Others 19.7 14.2 17.6 o/w interest & charges 11.8 11.0 11.0 external assistance 2.3 2.3 3.5 Share In Capital Expenditure Economic Infrastructure 31.8 33.8 33.5 o/w Road Construction 26.2 29.6 25.8 Economic Services 33.3 33.7 33.9 o/w Agriculture & Natural Resources 20.2 23.5 25.1 Social Development 27.5 24.0 25.7 o/w Education 11.4 11.1 12.2 Health 8.2 6.5 8.3 Others 7.4 8.5 6.9 Note: General Government is Central + Regional Govermments. 1/ Road construction under recurrent expenditure includes urban developments. Under capital expenditure, external assistance has been distributed across the sectors. Source: Ministry of Finance, World Bank Database and Projection for 1999/00 by the PER Mission (February 2000). Trends In Social Sector Expenditures 2.29 During FY98/99, the government was able to protect critical expenditures on social services (education and health), and particularly their recurrent portion (Table 2.6). In the education sector, the shares of recurrent and capital expenditures were maintained at 2.5 and 1.0 percent of GDP respectively, the same proportion as in FY97/98. In the health sector, while the share of recurrent expenditure was protected at 0.9 percent of GDP, there was a decline in the proportion of capital expenditure from 0.8 percent of 20 GDP in FY97/98 to 0.6 percent of GDP in FY98/99. During FY99/00, both recurrent and capital expenditures in both sectors are projected to suffer a decline. Overall, as a share of GDP, education expenditure is expected to decline from 3.5 percent to 3.2 percent and health expenditure from an already low level of 1.5 percent to 1.3 percent. Table 2.6: Recent Trends In Social Expenditures Education Health Social sector expenditures FY 97/98 FY98/99 FY 99/00 FY 97/98 FY98/99 FY 99/00 ----Pre-Actual---- Projection ---Pre-Actual---- Projection As % of GDP 3.5 3.5 3.2 1.7 1.5 1.3 Of which: Current expenditures 2.5 2.5 2.4 0.9 0.9 0.8 Capital expenditures 1.0 1.0 0.8 0.8 0.6 0.5 As % of Total expenditures 14.1 13.2 12.5 6.5 5.7 5.2 As %of Currentexpenditures 15.7 14.3 12.6 5.6 5.2 4.1 As % of Capital expenditures 11.4 11.1 12.2 8.2 6.5 8.3 Source: Ministiy of Finance, World Bank Database and Projection for 1999/00 by the PER Mission (February 2000) 2.30 As a proportion of total government expenditures, however, the shares of both education and health have fallen during FY98/99 and are projected to fall further during FY99/00. The share of the education sector in total expenditures declined from 14.1 percent during FY97/98 to 13.2 percent during FY98/99, and is projected to further decline to 12.5 percent in FY99/00. The share of the health sector dropped from 6.5 percent in FY97/98 to 5.7 percent in FY98/99, and is projected to drop further to 5.2 percent during FY99/00. The projected decline in these sectors, however, is particularly pronounced in the case of recurrent expenditures, as their shares in capital expenditures are expected to increase. This is a particularly worrisome prospect in light of persistent complaints about inadequate recurrent expenditures to complement sunk capital investments and the consequent erosion of the development impact of public expenditures. 2.31 Over the past three years, the government has put considerable emphasis on the social sectors through the donor-supported Education and Health Sector Development Programs. This emphasis, however, is not borne out by figures of real per capita expenditures (Table - 2.7). In the education sector, real per capita expenditure increased from Birr 8.6 in FY97/98 to Birr 8.8 in FY98/99 before declining to Birr 8.4 in FY99/00. The health sector witnessed a secular decline in real per capita expenditures, falling from Birr 4.0 in FY97/98 to Birr 3.8 in FY98/99 and further down to Birr 3.5 in FY99/00. Given Ethiopia's already low expenditure levels in these sectors, reversing these trends is particularly critical for the realization of the poverty reduction objectives of the government. 21 Table 2.7: Real Per Capita Social Sector Expenditures 1997/98 1998/99 1999/00 Pre. Actual Pre. Actual PER Proj. Education expenditure (million birr) 1,586.3 1,719.2 1,778.3 Health expenditure (million birr) 736.0 736.5 735.6 GDP Deflator 3.08 3.17 3.33 Population (million) 59.9 61.7 63.5 Education expenditure-real (million bir) 515.03 542.33 534.02 Health expenditure-real (million binf) 238.96 232.33 220.90 Education expenditure per capita (biff) 8.60 8.80 8.42 Health expenditure per capita (birr) 3.99 3.77 3.48 Source: Ministry of Finance, World Bank Database, Projection for 1999/00 by the PER Mission (February 2000) and MEDAC. 2.32 Social sector expenditures in Ethiopia compare unfavorably with those of other African countries in the same per capita income bracket.8 Available comparative figures (Figure - 2.2) show that the ratio of education expenditure to GNP for Ethiopia was below the average for Sub-Saharan Africa and, was lower than those of all other countries with the exception of Zambia and Burkina Faso. The relative performance in the health sector is considerably worse, with Ethiopia having the lowest health expenditure GNP ratio in the sample group of countries. 8 Poor countries with per capita GNP lower than the average for the Sub-Saharan Africa for which data are available in the World Development Report for both education and health expenditures. 22 Figure 2.2: Share Of Social Sector Expenditure To GNP Comparative Perspective Education E_enditure as a share of GN, 1990-97 Zambia(1990-97) Togo (1990-97) &Sb-Saharan Africa (1990-97) Kenya (1990-97) Gambia (1990-97) _ _ - - Burkina Faso (1990-97) Ethiopia (FY 99/00)* Ethiopia (1996) 1 2 3 4 5 6 7 Percent Health Expenditure as a share of GNP,1990-97 Zambia (1990-91) - - Togo (1990-97) Sib-Saharan Africa (1990-97) Kenya (1990-97) Gambia (1990-97) Burkina Faso (1990-97) - - Ethiopia (FY 99/00)* Ethiopia (1996) 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 In Percent Note: * Government projection for FY99 - 00. Source: World Development Indicators - World Bank 1999 23 3. INFORMATION SYSTEMS FOR STRATEGIC EXPENDITURE MANAGEMENT INTRODUCrION 3.1 This PER's focus on information management issues was requested by the government at the 1999 Brussels workshop. Practical problems mentioned included: * Delays in receiving information needed for budget preparation (e.g. indicative planning figures, information about aid flow, notification of the regional subsidy). * Difficulties in reporting expenditures (e.g. delay in closing accounts, problems with the channel 1 reporting, limited capture of donor disbursements through channels 2 and 3). * Limited capacity to satisfy competing and overlapping demands for information. * Weaknesses in horizontal and vertical communications across government agencies, and between aid agencies and government. 3.2 There are theoretical as well as practical reasons for analyzing information flows in public expenditure management: - Budget outcomes are affected not only by incentives but also by the information policy-makers and managers have in spending public money. Information is a constrained resource, not only because it costs money to produce and distribute, but because the amount of information that can be generated and considered in the compressed budget schedule is severely limited. Ignorance and information asymmetry are widespread behavioural conditions in contemporary budgeting, even in countries that have state-of-the-art expenditure management systems. These conditions are due to two related factors: the costs of generating and disseminating relevant information; and the advantages that information producers (agents) have over information users (principals).9 3.3 The potential scope of this topic is huge. This report focused on a subset of critical issues, whose choice was guided by: * The need to look at both the demand and the supply side. This involves: 9 Allen Schick, A Contemporary Approach to Public Expenditure Management, World Bank,1998. - Considering the scope for reducing demands on information providers as well as enhancing their capacity to provide essential information. - Considering also the constraints on information users (analysts and decision-makers), and hence their ability to use available information effectively. * Awareness of the costs of generating and disseminating information, and hence the need to focus on a core of most relevant information that is practical to collect and assimilate, and to minimize overlaps and duplication in information management. * Considering the information implications of decentralization. * Taking stock of the existing situation and reforms already under way as a basis for any recommendations, and trying to identify practical ways in which ongoing efforts can be reinforced, improved or complemented. Donor information requirements in particular should be seen in the context of basic government systems. EMCP, a comprehensive program of reforms in public expenditure management is under way, as one major component of a broader Civil Service Reform Program. These reforms, which have a bearing on almost every issue discussed in this chapter, are summarized in Appendix 2 (Volume II). 3.4 There are distinct information requirements at each stage of the planning and budgeting cycle. Although financial management and expenditure reporting are core concerns, they must be set in the context of strategic planning of expenditures and the need for complementary non-financial information. Box 3.1 (based on Schickl0) summarizes the information requirements for each element of public expenditure management. These requirements are an ideal that no country fully achieves, but they provide a useful standpoint from which to assess the strengths and weaknesses of information flows for resource management in Ethiopia. 10 From Schick, ibid, Tables 1.1 - 1.4. 26 r Box 3.1 Information Requirements For Public Expenditure Manageent- EBB. ~OumcertvE . - osMaoNXSQtnalBtndRhNTS Aggregate fiscal Budget totals should be the result of explicit, Tic median-tecm expenditure fratewolt puviides a di=ipline enfotcd deci5ions; they should riot merely baseline for accommodiate ape odemais. These totals . should be set before the indsridteal speeding ftleasriwg.te budgebty impats ofpolicy tisatg. - dcisions arc made. and should be sustainble lTroughout fomlatiorn ofthe budget. information is over the mnediumi-terM and beyond, provided on changes to the basliine. During inweretati nof tlXb bwgect spaedigisa nmtored to ensure campliace witlsvthe fiscal aggregates Allocative: Expenditures shouId be based on goverYnmet Mirijts and nanagzrs geneate or rive infortion.on etticiescy . priorities and on effectivenass of ptic . the nalt -cexpect effleirotivrs of pr s, asvwi progransm The budgt system sbosiud spur As th1a social oeuto enigtontongoingproenss, realltocation fmrris lesseeto higher priorities And budget actions, and policy iititives. They also recqive from lesato more elfeative progrMs. istfortnation ortt th expendiure impacts (reltive totlh - ndiso-ton framewawi) of astcrised. and proposed budget actions. Operational Agencies should produce goods and services at Budgeted oups re specified in 4dvasse ap actual ! effi~ scy . - , coat tht abhieves ongotg etfitienty gains, Comuts arenspared to ttargelin Cot are loreled -and (to the extent appropriate) is ompetitive (ideally on an accal basis) thtle activiies responsible with market price, for them. Infoematien on finncial and rannialtional perfornianee is pablised in annal repot and other L . _ . .. . ... . .,_..docurnt 3.5 The next section addresses three main issues in the strategic planning of expenditures: government's capacity for medium-term expenditure planning; the impact of donor requirements for strategy dialogue and documents; and decentralized planning of expenditures. The subsequent section focuses on financial management issues within the annual budget cycle. A final section summarizes the team's main findings and recommendations. STRATEGIC PLANNING OF EXPENDiTURES Context 3.6 The government has a strong sense of strategy, bottom up procedures for defining priorities and discipline in pursuing objectives. An exercise to prepare the next round of federal and regional Five Year Plans is currently under way, and the government has also prepared a variety of other strategic planning documents (including those associated with the various SDPs). As Box 3,2 describes, Ethiopia has avoided the extremes of underfunding on the recurrent budget and over-commitment on the capital budget. However, expenditure allocation is dominated by the task of fitting expenditures to available resources within a strictly annual horizon and this is done largely on an incremental basis. Even for the capital budget, the system is geared to the implementation of most projects within a single fiscal year, and medium-term strategic issues are not rigorously addressed. 3.7 These weaknesses are recognized in the government's own analysis which underlies the Expenditure Management and Control Program (EMCP), and the 27 strengthening of strategic expenditure planning has been a prominent topic in the last three PERs. Indeed Ethiopia has made considerable headway in adopting a medium-term perspective on the capital budget side. MEDAC is implementing the PIP of EFY91-93 cycle, while the PIP of EFY92-94 cycle is reportedly awaiting the approval of the Council of Ministers. 3.8 The principles of medium-term expenditure planning are now widely accepted, and put into practice in various ways in OECD as well as developing countries. The World Bank's generic model, popularly called 'medium-term expenditure framework' (MTEF), has many possible variants.'" There have, however, been concerns about Ethiopia's model of approaching MTEF through the PIP route. The argument is that-PIPs are rather outdated and inappropriate. Modem planning and budgeting systems emphasize the importance of an integrated approach to recurrent and capital expenditure, whereas PIPs may perpetuate and a bias towards capital expenditures. However, the Ethiopian PIP is intended as a staging post towards a more comprehensive Public Expenditure Program (PEP) covering both capital and recurrent expenditures. 12 Arguably, the MEFF provides an appropriate framework for strategic planning of public expenditures. It introduces a multi-year perspective, it requires an analysis of overall resource availability and provides the opportunity to address tradeoffs between federal/regional, recurrent/capital, and alternative sectoral expenditures. " The World Bank's approach is described in its Public Expenditure Management Handbook (1998). The Asian Development Bank's more recent Managing Public Expenditure (Salvatore Schiavo-Campo and Daniel Tomnmasi, 1999) includes a review of OECD practice and notes (p287): The semantics, however, should not obscure the common purpose of all variants of an MTEF approach: the extension of the budgeting horizon through a systematic approach that includes a strong policy-budget link under a hard constraint derived from a sound macro-economic framework. 12 It is not clear if the proposed PEP will take into account the recurrent cost implications of only future capital investments, or also of past investments as required under the MTEF discipline. 28 Box 3.2: Recurrent Costs,1 Dual Budgeting And Expenditure Strategy Th classic recurrent cost problem is that r budgets are chronically squeezed. Syt include failure to main physical assets publ services tha are compromised for lack of opati s d inability: to bget prperly_for the operation of capital projects as they are cmpleed. Often the recurren ost p em is associated with inabilitY to control public employment and conseqetsio of ivil service salaries. The'underlying cause of the recurrent cost problem is ta govemments commit themselves to levelof acti that.their resources cannot sustainJ stitui o tat ispo.se oftcl incremental anual budget, Itat are not disciplined by hard medium tem e6lings, the separation o recurrent from capital budgets and a poitiical bias towards capital expenditures. However Btopia dos not aly fit the classic patten:- Civil servi=ce siz2:e is noat excessive .(e.g n terms of numbero civil sens ve totl puatio), althuh there are signs of ineent between publ and ivate r ates. - * The sequence: in which budgets arm pare (see Box 3.4) does make g obligations the fist claim on the resources. £- Although there is a dual bgig structure (wih recurrent and capita udgetd separately), there are some orts (tgh he ao of project work -betwen fince and planig bodies, etc) to ensure that e orating costs of n projects are incorporated inf the recurrent budget, that starere ands t - . These m smsgenly work: tere have been some hiccs - eg.- e s in Oromia for which operating fuds had not been reserved- but these are extionaL All in l, the present system displays strong disce in: (am g ependitues toe fx i ea and (b) prioritzing epee.to avoid ihe extremes oapi e mm t an recurrentuder;funding, Wt is stll cng is a genuinely strategic approach to.the alsis o and term expen itreq ets (reret a capital)Th: * In practice, only the salary cost elements of recurrent budgets are fully ted ere r < -shortages of fundsor OM costs (hence, for example, vehicles off th^e road for lac of und for fu or reoairs). The governments budet r program (se Appdix 2wo to devlop noums to estimate the full OM requirements of activities. : Desite systematc coordination between plriu and fiance bodes, reuretad capta planning and budgetigarenot uiny i Common budget code to be iroduc un --der te budget reform progrm will facilitae -more of a= cost-center approach but wilnot, by themseves, enue that reuret and capia expeditures are analyzed in ana inteated i+. pl At regional ievel, t about e fiuture federal subsidy inhibits sda --expendiues. Activities are dominate by -the task of adjsig expenditrs to aloations revealed vey latein thfe budget cycle. - bu Conseqenthiopia still cks a rigorous amework for addressing cal stratgic ises such - possible trade-os betwee capital expendiures ad O&M of exist assets; the balancebetween federal and regional expenditure;- - * the mutual compatibility of present and future secor development rgam. - 29 3.9 The essence of a medium-term strategy on public expenditures is that there should be binding decisions at a high level (Council of Ministers and Parliament) on broad, multi-year allocations of expenditure (Indicative Planning Figures or IPFs in Ethiopia's terminology) before detailed annual budgets are prepared. The government's approach (described in Appendix 2) provides for this - indeed it is required by law that the Council of Ministers and Parliament should approve such allocations. A detailed budget calendar to ensure that budgeting is systematically preceded by strategic expenditure planning has been developed13 and a strong recommendation of PER 1999 was that this should be put into effect. In practice, however, this has not yet occurred and the Macroeconomic and Fiscal Framework (MEFF)'4 and Public Investment Program have remained a shadow exercise. Strategic Expenditure Planning 3.10 The key question is why GoE has been so tentative in its approach to strategic expenditure planning. At the federal level, there has been significant progress in developing the technical basis for genuine medium-term planning of expenditures (especially the preparation of the MEFF), but there have been delays in completing the technical documents and neither the MEFF nor the PIP itself has yet been formally presented to the Council of Ministers or to Parliament. Hence, Cabinet-approved IPFs have not been issued to federal public bodies, and notification of the regional subsidy continues to be delayed until the last minute. As a result, budget preparation at all levels of federal and regional governments still takes place without credible ceilings, does not look further than the next fiscal year, and is dominated by top-down trimming to fit the funds available. MoF and MEDAC continue to devote most of their attention to the preparation of the recurrent and capital budgets in the traditional way. 3.11 GoE interprets the problems in terms of uncertainty and lack of information. The border conflict makes economic forecasting more difficult, makes pre-emptive demands on expenditure, and has disrupted aid flows. Even without the conflict, it was proving difficult to make reliable medium-term projections of aid flows. These problems are serious, but it is perverse to argue that such uncertainties make medium-term planning impractical. A major reason for undertaking systematic medium-term expenditure planning is to address such uncertainty (this year's MEFF, in fact, attempts this by exploring a number of possible scenarios), and the preparation of medium-term forecasts could be an important focus for dialogue with the donor community about aid commitments and their potential impact. 3.12 International experience indicates that medium-term expenditure planning will have no impact if it is allowed to remain a separate technical exercise. This is one of the key lessons from South Africa's MTEF, which is described in Appendix 3 (Volume II). 13 The latest version is reproduced in Appendix 2 of Volume II of this PER. 14 See Macroeconomic and Fiscal Framework for (EFY) 1993-95, dmft, Development Finance and Budget Department, MEDAC, December 29, 1999. 30 A key lesson is that the South African MTEF did not run parallel to an existing budget process, but was made the only budget process (it became "the only game in town"!). South Africa did not wait for the medium-term expenditure planning system to be perfected and then apply it. Rather, it was put into practice and then modified and improved in the light of experience. 3.13 A similar approach is required in Ethiopia. Successful application of medium- term expenditure planning depends much more on political will than on technical improvements to resource forecasts. What is now needed is the political will at the highest level to first give a statutory cover to the budget calendar and then to adhere to it and, as financial legislation already requires, to make the setting of IPFs the basis of resource allocation each year. A clear political determination to apply the planning/budgeting procedure that has already been approved would, in turn, create pressure on MoF and MEDAC to collaborate more systematically and effectively than has occurred so far. 3.14 Effective strategic planning of expenditures will require stronger collaboration between MoF and MEDAC, and between finance and planning bodies at the sub-national levels of government. The proposed allocation of fiscal responsibilities within the financial calendar (see Appendix 2) assigns to MEDAC the responsibility for multi-year planning and programming, and to MoF the responsibility for preparing the annual fiscal plan. In practice, both exercises need to be jointly undertaken by the two ministries, and there should be more effective use of inter-ministerial working groups than has so far been the case. Similar collaboration will be needed at regional level too. As the new cost-center approach develops, it will be necessary to move away from the practice of deciding the recurrent/capital split at the first stage of budgeting each year and to develop a more thorough integration of recurrent and capital budgets. Aid Dialogue And Expenditure Strategy 3.15 Ethiopia has great potential for genuine partnership between the government and aid agencies. The government is disciplined and aspires to take the lead in aid coordination. Ethiopia was the first African country to host its own Consultative Group meeting, and the government took the initiative in developing sector wide programs for health, education and roads. For the ESDP and HSDP, innovative governance arrangements have been adopted to support sector-focused, rather than project-focused, aid. Successive PERs have also provided the basis for a more meaningful collaboration between the government and the donor community as a whole.'5 The border conflict has led to tensions in short-run aid relationships, but these should not be allowed to undermine the progress that has already been made. 3.16 Previous PERs have noted the burden on government (at both federal and regional levels) of simultaneously pursuing three major tracks of organizational reform - 15 See Chapter 6 for a detailed account and analysis of the PER process. 31 decentralization, civil service reform, and the SDPs. In addition to the demands imposed by its own agenda, government faces aid agencies' demands for the production of various strategic planning documents (see Box 3.3)16 There is some potential overlap among the exercises and documents summarized in Box 3.3, but there seems to be no clear mechanism in place to co-ordinate and rationalize requirements at the national level - either among the various international agencies involved or between government and the donor community. For the education and health SDPs, a collaborative governance structure has been established which facilitates sector-specific discussions. This is valuable, but systematic review of macroeconomic issues and cross-sectoral concerns such as poverty requires a broader framework. 3.17 This will be particularly important if there is to be further progress towards more effective forms of aid, including budget support. The border conflict has stalled progress for the time being, but as soon as the conflict ends there will be pressure to mobilize support quickly. However, experience with the SDPs shows that effective implementation arrangements are not easy to devise or to agree upon. The government, in collaboration with the donors, should already be working on: (a) the framework of the PRSP; and (b) the modalities for channeling aid that may flow in soon after the conflict ends - both by way of reconstruction support and also normal aid which may increasingly be channeled through SDPs or as free standing budget support loans. These modalities, incidentally, might also serve as a channel for utilizing the resources freed by debt relief 3.18 As next steps, the main aid agencies should rationalize their demands for dialogue with the government. They should put forward, for high level discussion with the government, consolidated proposals for a streamlined set of documents on which dialogue might focus, and propose a unified calendar for donor-government dialogue. This calendar should be synchronized with government's own planning and budgeting calendar, and would, inter alia, link the SDP, PER, PRSP and CDF processes. 16 Many individual donors also demand goverunent involvement in preparation of their own programming documents, and annual aid reviews etc. 32 B,o.x 3.3: Stra,tegy Documents For Intern,ational Dialogue. There are several major focal points for dialogue between.Ethiopia and its.intenational aid partners,but the: connections between them are not yet well articulated..:. Sector Development Programs fior health and education are based on federal and regional sector shtategic plans which are sumnarised in a Program Action Plan for each sector. A governanie structure as been developed based on Joint Steering Committees at federal and 'regionllvels,annual sector-wide Join,t, Revi,ew Missions (JfM), and an Annual: Review Meeting attended by'all the main paities to the prograts. For each sector there is also a co-ordinating.committee amongst the local offices:of the.doniors involved. The Comprehensive Development Framework:(CDF) has been put forward by the World Bank as a holistic approach to development At its.core is the idea of a batrix which, on the one han, identifies links between macroeconomic and social: factors4in development,i and:, on the other, identfies the comple,mentary roles to:be :played by government, donots, N(O0s, civil society and othe skeholders in cartying forward a nationally owned dev.lopment sthategy.' Ehopia has been identified as one of' the pilot countnies to experiment with th CDF. According to the CDF web site.: The government is interested in usi,n,g the COF to promote a meaningful d ialogue with its: partners regarding its development policiy. Secific areas to be addressed could include decentralization, ination,al capacity buidng, or private sector. development The gover,,nment,would.like thfe CDlF to. help improve donor coordin,ation,particularly in:the :sector. investment programs that require harmonization of procedures and: .signiicaant donor,collaboration. Poverty Reduction Strategy Paper (PRS,P). An immediate priority for Ethiopia is ihe preparation of a Poverty -Reduction Strategy Paper'(RSPO) which has. since become a prerequisite for.accessing concessional assistaice form the World Bank (through the lfiternationalDIevelopment.AssociatioOt) and IME (through fte Poverty. Rteduction and-<3rowih Facility). It is expected. that-the PRSP- will become the basis for development assistance from the-entir'community of multilateral and'bilatera donors,.An effe poety rcti' strategy would be * Be prepared by the country with the active part,ipation of stakeholders. * Focus on fAer economic growth - growthfrom which the poor benefit and to which the poor6contribute. * Reflect a comprehensive understanding of' the comp and rualti-diniensional nature of poverty and its-determinants. * Assist inichoosing public actions tht have the highest poverty impact. * Establish output/outcome indicators which are set and monitored using a paticipatory process. Annual Public- Expenditure Reviews are increasingly seen as having a role ingivnmentldonor dialogue, wich may, forexample, undeipin the budgetayprojections on which the SDP targts arebased. The government: indepe,ndently prepares-Five Year Plais (with new federal and regional p now in. preparation for the five year perod commencing. with EFY L993), and has embarked.on the annmal preparation of a three-year rolling Macroeconomic and Fiscal Framework and federal Public Investment Progrm. ' There is some potential overlap in these documen ts. There is also appreh,ensio about the volume of work involved in the preparation of CDF and PRSP as these are relatively recent initiatives, * Poverty Reduction Strategy Papers-Status and Next Steps. Staffs of the IMF and the.:World ak, November 19,. 1999. 33 Expenditure Strategy And Decentralization 3.19 Apart from its political-economy rationale, there are legitimate arguments for decentralization as a means of more efficient information management. Decentralization should allow local preferences more influence and so yield a better allocation of resources. Local managers who are closer to the ground should also be better able to deploy resources effectively and make adjustments at the budget implementation stage. If decentralization is accompanied by a real delegation of responsibilities, it should economize on information flows. However, such delegation requires an increased downward flow of information on resources and strategy, while a reduced upward flow of financial detail needs to be matched by the development of better information systems on performance.17 Decentralization without delegation may increase information demands by simply introducing additional layers of reporting and instruction. 3.20 Ethiopia's model of decentralization is quite radical and accords considerable autonomy to the regions.18 Federal transfers to regions take the form of block grants, and a number of regions have further decentralized budget preparation to zonal or even woreda level (see Box 3.4). There are encouraging signs that this can lead to more efficient use of resources. Bottom-up identification of priorities is an established tradition, and flexibility in budget implementation allows resources to be shifted between projects so as to improve the overall implementation rate. However, effective planning and use of resources at sub-federal levels is compromised: * by the federal government's failure to provide indicative medium-term projections of the regional subsidy, * by donor practices that inhibit the multi-year programming of aid flows (see discussion and recommendations in Chapter 4); and * by management arrangements for the education and health SDPs that demand centralized information and decisions (see further discussion below and Box 3.5). 3.21 The decentralized system is dominated by the annual budgeting cycle, and as indicated in Box 3.2, the links between recurrent and capital budgets are unsatisfactory. Regions have come to expect that notification of the federal subsidy will be delayed, and have various strategies for coping (by making conservative assumptions about funds that will be available, and by last-minute pruning of projects according to priority ranking.) In effect, regional planning and budgeting is concerned with fitting a pre-existing shopping list into an annually revealed expenditure envelope, but this does not amount to strategic planning of expenditures. 7 See Chapter 5. IB See Ethiopia: Regionalization Study, World Bank, Report No. 18898, February 3, 1999, and Implementing Sector Development Programs in Ethiopia, Stephen Lister, December 1998. 34 3.22 The exercise of ESDP and HSDP preparation did involve the coordinated regional-level planning of recurrent and capital expenditures within a constrained expenditure ceiling. Regions have also committed themselves to the preparation of PIPs, along the same lines as the federal government. The federal PIP team has undertaken a review of the scope for a regional PIP in Oromia, and similar reviews are under way for other Regions. PER 2000 reiterates the PER 1999 recommendations that: * With the enforcement of the budget calendar, regions should be provided with more timely projections of the federal subsidy. In addition, there should be a longer horizon on federal subsidy projections as well as notification each year at an earlier point in time. * When the PIP is extended to the regions, sector planners should be required from the outset to prepare consistent medium-term plans for both recurrent and capital expenditure, and regional and zonal councils should be required to undertake joint reviews of recurrent and capital budgets for each sector. 35 set oveall exenditue enveope (fr the otal offederazplus egiona expenitures determine reqirements for ederal recurrnt expenditur divie reidul fuds aailble etwen fdera captalexpeditues nd te reionl susid * allocate 4aAnI~~~~~~~~~~~~o eaitt xenstre mog eerl ubi bdis FINANCIAL MANAGEMENT INFORMATION Context 3.23 This section focuses on information management issues that arise within the context of annual budget preparation and implementation. The Ethiopian financial management system has a number of remarkable strengths. There is a widely understood line-item budgeting system, there is strong budget discipline, budgeted funds are made available in good time, controls are effective and there is a financial information system potentially capable of reporting for monitoring purposes. At the same time, a number of challenges are apparent with respect to financial management and accountability. These challenges have been well described in previous PERs19 and center around five issues. First, financial management in Ethiopia emphasizes control, rather than management and planning. Second, the budget preparation process is not governed by law, the budget calendar is still in a testing stage and the budgeting is incremental rather than strategic. Third, the capital budget has shown chronic underspending, particularly on donor funds, while there are weak links between the capital and the recurrent budgets. Fourth, decentralization has introduced ambiguities and complications in federal-regional relations. Fifth, the introduction of Sector Development Programs, particularly in health and education, has created new demands on financial management and reporting which the system has not yet managed to cope with. In all these areas, there is currently work in progress and commendable efforts are being made to face the challenges (see Appendix 2). 3.24 This section will address these and related issues in financial management information under four headings: budget planning and preparation; budget implementation; financial reporting and financial management capacity. Deficiencies in information provision on donor financing also affect financial management and are touched upon here, but are described in further detail in Chapter 4. Budget Planning And Preparation 3.25 Previous PERs have stressed the importance of an authoritative financial calendar as a means for effective budgeting. Allowing sufficient time for both planning and budgeting would, for instance, imply that public bodies at federal and sub-federal level are aware of budget ceilings in a timely manner, thereby making budget submissions more realistic and avoiding difficult 'fitting' procedures. 3.26 In 1999, the budget ceilings for public bodies and the subsidies for regions were communicated as late as May/June, although, according to the budget calendar under consideration, reliable estimates ought to be provided in the second half of February. 19 See Social Sector Review/PER III, 1997, Chapter 9 - Generic Issues in Financial Management for the Central Government; PER, 1998, Volume I, 'Institutional Issues in Public Expenditure Management', pp. 20-34; and PER, 1999, Chapter 5 - Capacity Building in Public Finance Management 37 According to M:EDAC, the delay in the communication was attributable to the experimental stage of the budget calendar and uncertainty of donor financing. Different regions have adapted differently to the delayed notification. In most cases budgets were prepared with a slight increment or assuming a ceiling similar to the previous year. In other cases 'blue sky budgets' were prepared and later had to be sharply cut back. The most serious problems occurred where block grants are transferred to zonal (and sometimes woreda) level. It was often well into the first quarter of the new financial year before the exact levels of such block grants were known, creating much uncertainty on investments and operations. 3.27 Work is progressing under the Expenditure Management and Control Program (EMCP) of the Civil Service Reform in reforming the budget preparation process (see Appendix 2). The calendar maintains the distinction between planning in the first half of the fiscal year and budgeting in the second. The Fiscal Plan, which will be prepared during the first half of the annual cycle, will determine estimates of the overall ceiling of federal recurrent and capital expenditure, as well as the ceiling for subsidies to the regions and administrative councils. The estimates of subsidy would be released to regions and administrative councils at the completion of the Fiscal Plan (by January 8th). Implementation of this recommendation of the Budget Reform team will provide accurate estimates of subsidies to regions and administrative councils early enough for meaningful budget preparation at sub-national levels. The PER team fully endorses this recommendation. 3.28 The budget for FY99/00 was not only prepared late, it also included significant reductions in subsidies to regions, which were not matched by concomitant increases in regional revenues. All regions visited by the PER mission have adopted a similar approach to coping with the reduced funding, by first budgeting for non-discretionary recurrent expenditure (mainly wages), and then for the implications of ongoing activities (recurrent) and projects (capital obligations). This has meant that in a number of cases discretionary recurrent expenditures have experienced further reductions, adversely affecting the effective running of services as well as the return on past investments. 3.29 Although the underfunding of O&M expenditure may have been more severe than normal in FY99/00, it is not a new problem. While many financial managers acknowledge the potential efficiency gains from adequate budgetary provision for O&M, the practice has been different. This is partly a structural problem, related to the absence of clear links between the capital and the recurrent budgets (see Box 3.2). The introduction of cost-center budgeting, as proposed in the Budget Reform Design Manual, will address this weakness in the existing budget structure. It will allow for total costs for sub-agencies and projects to be clearly identified and linked in the budgets. 38 Budget Implementation 3.30 Budget implementation is generally satisfactory with respect to recurrent expenditure. Performance on the capital budget has differed over the years and over the regions but shows better results for treasury-funded than donor-funded capital projects.20 For treasury funded projects, most regions have a well established monitoring system that assists in assessing physical progress and in reallocating resources from slow to quick disbursing projects. Such decisions may require Regional Council approval, but that is 21 relatively easy. 3.31 Flexibility in budget implementation is much more limited in the case of donor funds, and the utilization rate for donor funds is much lower than for treasury funds. As well as the constraints on the donor funds themselves, inaccurate estimation and low disbursement rates in donor funded projects have additional negative effects as they often tie in government contributions (matching funds). It only becomes apparent late in the budget year that donor funds will not reach the anticipated disbursement rate and it is only at that stage that the matching funds may be freed. In some cases, donor contributions have been earmarked for priority purchases (e.g. drugs) and when they do not materialize in time, a reappropriation in government budget is necessary to make up for the shortfall. 3.32 The problems associated with donor disbursement are extensively discussed in Chapter 4. However, there is one characteristic of the Ethiopian capital budgeting system that negatively affects disbursement of both treasury and aid funds: the customary cycle of project preparation is geared to completing the implementation cycle for most projects within a single fiscal year. Thus tender preparations take place during the first quarter of the FY (tenders cannot be called until availability of budgeted funds is confirmed) and contract activity increases during the third and particularly the fourth quarter of a financial year. This leads to a striking concentration of capital expenditures in the fourth quarter. With better enshrined strategic planning and more reliable information on capital budgets for the years to come, there is no a priori reason why the disbursement pattern could not be smoother over the year. Smoother flows would assist in monitoring and lead to improved utilization rates of capital budgets. The one-year horizon is especially problematic for donor funds, since it is exceedingly difficult to fit in the extra stages of programming donor funds and securing donor approvals for tenders etc. Financial Reporting 3.33 Financial reporting on treasury funds follows well established mechanisms and normally involves monthly reporting on both recurrent and capital budgets to the next higher level by woreda, zonal and regional finance bodies. In some cases capacity 20 This last point is illustrated by the experience of SNNPR [Box 4.3, Chapter 4]. 21 Within the capital budget, construction items cannot be vired to operating items and vice versa. Also capital expenditures can not be vired to recurrent expenditures. 39 problems (notably shortage of professional staff) have hampered the timely compilation and dissemination of financial reports.22 Capacity problems at zonal and woreda levels have at times also affected the accuracy and reliability of financial reports. Nevertheless the system reasonably allows for monitoring and feedback on the use of treasury resources and is used as such at zonal and regional levels. However, this reporting is geared more to the control than the management function, and reports are too aggregated to allow any reasonably meaningful analysis of expenditure trends. 3.34 Financial reporting on donor funds is much less satisfactory. Much information concerning disbursements on donor funded projects is not captured in the reporting systems. This deficiency can be broken down into two components. First, there are many externally financed activities that are not captured in the budget (including much of channel 3 financing; NGO support, and food security related activities) and are therefore not captured by the regular financial monitoring and reporting system. Second, for those activities that are incorporated in the budget, flows of disbursement information are usually not systematized at regional levels and availability is dependent on ad hoc arrangements with donors and/or information obtained through MEDAC. 3.35 Information requirements for donor financed projects are much more demanding than for projects financed from treasury funds. This not only refers to additional financial reporting but also to information for the various procurement procedures. The information demands place a heavy burden on an already scarce capacity. On the other hand, some regions have shown that an initial investment in better understanding donor procedures pays off in a much smoother flow of funds over the long term. Given the expectation that at best it may take several years before the majority of donors will be able to use channel 1 financing,23 pragmatism demands that all regions should continue to secure a good understanding of relevant donor procedures, which again requires an investment in human capacity. There are additional issues relating to the SDPs in health and education, which are described in Box 3.5. 3.36 The Expenditure Management and Control Program of the Civil Service Reform involves various activities that will in due course improve the capacity for financial reporting. First, ongoing work on budget reforms implies a modified way of coding both expenditures and revenues (including donor financed projects) which will enable the establishment of better linkages between related activities as well as between capital and recurrent expenditure. Second, a new computerized Financial Information System will be designed and implemented that is meant to improve both the quality and the timeliness of financial reports (see Box 3.6). However, implementation and application of the new systems will still take several years. 22 Tigray is a case in point, where financial reporting to the center currently takes place at six-monthly rather than monthly intervals. 23 Donors have not clearly spelled out minimum requirements for channel 1 (non-earmarked) financing to become feasible. Moreover, the requirements may differ from donor to donor. Nevertheless, reliable and timely financial reporting, a sound macro-economic framework, and an open debate on main policy lines seem to be minimum preconditions. 40 Financial Management Capacity 3.37 The capacity of the government to manage finances and to submit adequate and timely reports is negatively affected by the staffing situation at all levels of the administration. There are numerous vacancies, both of key personnel at federal and regional levels,24 and of supporting staff at zonal and woreda levels. Salary levels are not high enough to attract staff with the required qualifications in key positions. The high turnover of staff further contributes to the loss of skills and institutional memory in the system and has made past efforts of capacity building through training less effective than anticipated. At zonal and woreda levels the problem is compounded by the often insufficient skills of financial management staff. Existing staff are stretched to the point that it becomes difficult to undertake training for capacity building without falling behind on essential routines. This situation has led to inadequate and incomplete financial reporting, requiring remedial actions at higher levels and leading to delays in both reporting and the closing of accounts. The issue needs a system-wide solution, preferably as part of the Civil Service Reform program, but is the more urgent as other sectors (e.g. health, education) face the same problem. 3.38 Adequate supportive equipment is absent, particularly at regional levels and below. Computerization is spreading rapidly at regional levels but is mostly in its infancy at sub-regional levels. Communication at these levels is also problematic with the number of operating vehicles being far short of the demand and sometimes the lack of telephone connections to more remote woreda offices. At the same time, the federal MoF has made good progress in 1998/99 with a move to an open PC-based system which now includes the modified Government Accounting Package (BDA) as well as the payroll system. A project to computerize the Regional Finance Bureaus and to provide them with appropriate software and training is programmed for implementation in 2000. The aim is to facilitate a smooth information flow (in electronic form) between the Regions and MoF. 24 In Tigray, one department of the Regional Finance Bureau is presently doing the work of three departments, as the other departments are not adequately staffed. 41 ~BGX.3.5: I,ssues Arising From Th Social Seor Devlomet Prorm The Setdor Deelpmn Progasin Health anid Educaion have placed additionali-demandsonfwsf inancial nformatio. A secto-wieapprach reurstall(governmnent, donor NO informaio should be apurd ndreguary reported upon.~ New, hrmoniserprtn forat have bee devloe an 'wdel distributed oas par of1the -rga mplemnenttio n Manuals. Hoeer oe etin rblemis ocurdi 1999wit repec totimly ad aequte epotin,wichnegaieyafce h 99Annual Review Meeting for theSPs. Tfhe ARM, i tb was splitinto two:afirsts sssion on physicalpprogress in March; a second, on finiancial achievements and plans in May/Junte. AccoQrding to the& PM, th Rsol normally take place mi Februar.y Moreover,~ the finaniitAl reorts and pasresented in the second 'leg' :of the ARM were not Al Sway cmplt and eliale, hic hameredthepurpse o themeeing.The onslidaed reports fo F 989,submiitted i However, both copaison with ofM reots andeckingTh ofndat durn h reeo AM (ac 20) showe tha thee arestil conideral omssine dcnrdctosi iacaladpors eprignh D reporting fomats remain ubject to re isoandi 1~.hhav nsm ae been simplified. Nevrgthees tein iess Eesides the need o ntertdsectorbal reotn,teSPshv oto ote unqe etue ta hv affected the normal government plann~~~~~~~~ingTv an re7tn systems county.mThispoint was,ti fhct, acnoleged duin dta hghe recent ARM Marh 200) categorization in~~- pl-pohscrae an pblemininaca repotin, wichhasbee addressed in the designoftheitrmrprtnbytmlnoligaraoae dmatci n g. -Pof, _expndtre-. coden, healthand a edcatonith budet nacaoies Italogreass on deofifi f ly n ntgrtingo P prgamnv ihte rmso egulare planng andbugein acivtes o oermndn: ft otno f Mor pundamentaily,lthere is tensinbtentedcnrlztion, O proequsfin th counbtry aioe capaity. Do nornt reurmnstvamr hi udn fteSP eursarespodn .1mids-the centralizationed-o fncIa inomatio fosp (epcall inepomtiomtio expndiure i a ondtio o futhe d Dbusements).ta nq#1atts~ft a af AhSuc earmarking,ewhen plip nkredpo etaiespornnnyomoo udsslorn counter. To he n fgvn regon ad onesi more responibiityforplanin an Theous havedht enaneiiWd th roe f theSD centr~alominisries(MHndWt MOe), notg onlyse In the area of shewtorplices o andstandard, but aloing liaisona wihdnre n progranuingofdoorfuds Tisw cadn conus zoe'andtegrions: usals lateso comuiato thouhPh plganuning anitnac bodies.ad Mid-term eviews of. .SPad...ae.u.o.aepac-uin 00 Rve o mlmetto arrangements sould be a com Ise exeriseorboth SDPsi ad6 sihudpaprtclrtenino th es sus 42~~~a p rcs iRccu Box 3.6: FIS, Interim FIS And Future FIS. One of the components of the Expenditure Management and Control sub-program under Civil Service Refo,rm,is the Financial Iformation System (FIS) Prject. This project aims to.improve the quali accuracy and timeiness of the financial ,infbrmation provided at operational and management levels for quick a.nd reliable decision making, effective control and effidient utilization of resources. In the long run. the project will lead to an integrated-financial information systm ,among the Federal Ministry, the Regional Finance Bureaus, Public Bodies and Banks, with linked automation having spread to zone and woreda level. The project started with the change from an antiquated mainfame system to an.open PC-based system in the Ministry of Finance in 1998/99, involving an adaptation of systems software and.the transfer of data for budget, accounts, disbursement, treasury, and payroll. A complicating factor has been the parallel ir,oduction of a reporting;. system meeting:the demands from the multilateral donors in the Sector Develop,ment Programs. This has led to the development of an Interim System with a specific donor reporting module micorporated in the modified version of the governent accounting system, also tailored to the needs of .Regioil Finance Bureaus. For FY99/00, the widespread introduction of the' existing system mith the' interim reporting arrngements was foreseen at regional leveL- However, the. delayed pur.chase of PCs-for the RFBshas slowed down this process somewhat. In the near. future a completely new FIS will replace the existing' system and .bring :together all the modifications under the-various components of the EMCP. in one integrated&systemn A company has been. selected for the design study, which shouild have started in early 2000. But as some components of the EMCP (notably the cash management project). have not :progressed sufficiently, the start of the design work has also been somewhat delayed 3.39 Another aspect related to financial management capacity relates to inter and intra- institutional co-operation. Adequate financial management far transcends the organizational boundaries and competencies of the MoF, the RFBs and financial offices at zonal and woreda levels. It requires good cooperation with other government institutions, particularly Planning and Economic Development, but also with Agriculture, Health, Education etc. Such cooperation could be improved, e.g. where capacity constraints affect one organization with a bearing on overall financial management and reporting (like drawing up financial reports on sector activities at woreda and zonal level). Concluding Remarks 3.40 The main findings and recommendations of this chapter are summarised in Box 3.7. Many of the issues raised in this chapter are further explored in the next two chapters, on aid management and assessing the impact of public expenditures. A number of issues may require further analysis in subsequent PERs or other forums. Possible topics would include the information requirements of wider stakeholders (including federal and regional legislatures, service users and the public), and more detailed assessment of progress in strengthening expenditure reporting systems. It should also be recognised that strengthening of information management systems, especially when it involves a re-design of those systems, is a long-term process, which requires persistence and stamina on government's part with concomitant support and understanding from its aid partners. 43 Box ,3.7: Summary .OfConclusions And Recommendations i"EA FNIG IRECoMMETiATmN RESPONSIBILITY ANDY Srtgc: plnig (oeuen taeyt ntoue Mi euieeti oiial will. Duin EY 1993 cycle., make these exercises central~gy; tto th osdrdb Cou l o reinlusdyoe theveya expen&otsffcenl codiaed ane tuchd blge oldfou, an eeopigpooalsd decisionmakes for dialosg_ue, likdt to Io s own overnMen. (sor cause 6f. ~ fiscalop calnd,ai4ir. o.WF s i: tem)m adM Decetralzatin Prnmipe o de~ci"entraliztion h is a) Aplybude Calendar, a)fMDC and expenditure ound, and, Withi larger Regions incldis ng tifmelykuti-year MoF(asaboe strategy ~ ~ ~ ~ ~ ~ Y moesoudb ncuapdvrjctoso rei nabeingwthF domuitedb shot-tem adustmnt b Enoua emlt-er b) Donors an 0 DIMmarcurentln c)d Ru~iii''~ii eview mtanageme mnt andfri c) Donorstan t o t Decentralizatio undermined by oordination struture ag oreeT, shoartrl 1~~~AWQ & failur toproviE subsdy dungiri td -te rev Siesof d en projections, by lank of multi-yeapr- ESDPh dan ud HSD cetrlzn trends. ig al Budget planning Late comuiato gof fisa dp n der obde&)a bv approah to bdgetin.C ) cans houh. 1MCP partlof ongin There s disiplin 'n'rting budgets prposalt eeo os-bde n dcsorinaioxnk betee reurentan as fonaionkfr ettGEr rga rnia capital budgets isiunited; dele 41 ibugato Of repniblt esosbliywt DeO P~~~dkispeinrecretcssar nknertion of recuw n eea n compressecL and capital udgets. Regional financ #1 OF,(~~~~~~~~~~~~an lann afidtXpeftilture sowo di I i4yeki M a~~~~~~~~~odis.bv).~~ an4 xW; giW a N9944 Y :~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C i 9~~~~~~~ = |KGDIS 1 3 £wV ;3 £ §£ -~~~~~~~~~~~~~~~~~O 1 .. X k. i S . . ~~~~t.A CD et.§S..## ...~~~~~~~~~~~~~~~~~~~~~~~~~~ :. -1 n :.:. . .n CD 0Q (~~~~~~~~~~~tI 0~~~~~~~~~~~~~~~~~~~C a ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~r ~~~~rio 0 r~~~~~~~~~yj~c CD 99: 4. AID MANAGEMENT AND INTEGRATION OF AID INTO THE PLANNING AND BUDGETING PROCESS INTRODUCTION 4.1 External loans and grants in the current FY are expected to account for nearly two-thirds of the federal capital budget plus subsidy appropriations to the regions.25 This alone makes aid an important aspect of Ethiopia's public finance management. Co- ordination of aid is a major challenge, given the large number of active bilateral and multilateral agencies (to say nothing of numerous international NGOs). Adding to the complexity is Ethiopia's federal structure since aid agencies may interact with many different regional administrations as well as the federal government. 4,2 Primary responsibility for aid management on the government side rests with MEDAC, which is also responsible for capital budgeting, for medium-term planning and for preparation of subsidy appropriations to the regions. However, there are important roles also for the Ministry of Finance (responsible for expenditure control, and accounting for both capital and recurrent funds, as well as for managing all debts - including aid loans and credits - contracted by the government); for a number of line ministries (especially those which serve as secretariats for SDPs); for the Disaster Prevention and Preparedness Commission (DPPC - which coordinates food aid and licenses the operations of NGOs); and for the Prime Minister's Office. 4.3 Two key principles of sound budgeting and financial management are comprehensiveness and discipline.26 GoE has been making a determined effort to apply both to its management of aid. Financial Regulations adopted in 1997 make it a statutory obligation to account for all aid as part of the consolidated federal budget.2" Discipline is reflected in the federal government's efforts to ensure that aid is directed to national priorities (hence the offset system, discussed later, to ensure that aid does not distort the regional distribution of resources), and in efforts to ensure that donors, as much as possible, use the same channels28 of disbursement as the government itself. Ongoing reforms under the Expenditure Management and Control Program (EMCP), are intended, inter alia, to improve the management of aid. Directly relevant initiatives include the introduction of Macroeconomic and Fiscal Framework (MEFF) and a Public Investment Program, and the strengthening of accounting systems and budget preparation procedures. A manual on aid procedures, as part of the budget reform project, and an aid database are on EMCP's near term agenda. (See Appendix 2 for a summary of these and other EMCP activities.) 25 Source: EFY 1992 budget proclamation. 26 Listed as the first prerequisites in the World Bank's Public Expenditure Management Handbook. (pl) 27 Financial Regulation, part IV, article 13. No 17/1997; 1 July 1997 2s See the note on terminology [up front with the acronyms] for an explanation of aid disbursement channels 1, 2 and 3 and other common terms. 4.4 Ethiopia has also been a focus for innovation in aid relationships. The SDPs represent a major and continuing innovation in GoE-donor relations. Among other initiatives, Ethiopia is one of the pilot countries for the CDF, and for enhanced cooperation between the EU and the World Bank, as well as between the World Bank and the IMF. Nevertheless, many problems and imperfections persist and both sides of the aid relationship experience frustrations. Aid management was a topic of previous PERs, but at government's request, it was decided to revisit a number of issues from a more problem-solving perspective, while donors requested consideration of two new topics - fungibility and food aid. 4.5 The rest of the chapter is thus organized into topical sections which cover: (a) integration of aid into planning and budgeting, including the forecasting and reporting of aid flows; (b) aid management at regional level, including implementation problems experienced and the offset system; (c) the implications of fungibility; (d) aid management and food security; and (e) government-donor partnership. 4.6 This report has adopted the perspective that effective aid management is a joint responsibility of government and donors. The accent is on identifying the most significant bottlenecks and putting forward practical ways to address them. Even where solutions are not easy, it is hoped that careful description of problems as seen from different perspectives will increase mutual understanding and move the discussion forward. Integration Of Aid Flows Into The Planning And Budgeting Process 4.7 For strategic expenditure planning purposes (see Chapter 3) it is important that donor aid is integrated into the planning and budgeting processes of a country. Planning of aid flows is also important for implementing agencies who wish to have an early indication of sources of funding available so that they can prepare for fulfilling the special conditions, if any. Inclusion of aid flows into planning and budgeting has become a key principle of sound public finance management, especially in those developing countries whose (capital) budget is largely funded through external grants and loans, as is the case in Ethiopia. 4.8 External assistance supplements the government's own revenues and thus increases the available resource envelope. At the same time, however, aid flows are mainly directed towards capital expenditures and entail important consequences for the government's recurrent budget, which is typically almost entirely funded by domestic resources. As such, decisions on donor resources clearly pre-empt decisions on the use of domestic resources.29 In order to be able to assess the consequences of the inflow of external assistance for the allocation of resources among different priority development 29 Hence the recommendation in the Budget Reform Design Manual, version 1.0, [paragraph 2.22, page 18] that "all capital projects must have a statement of the non-wage and wage recurrent costs at completion", Ministry of Finance, January 1998. 48 sectors, it is of strategic importance for countries to produce multi-annual rolling MTEFs and to incorporate as much information as possible on aid flows into this planning document. The inclusion of aid flows into the MTEF will facilitate strategic expenditure management and improve the quality of the annual budget which should be based on MTEF.30 Furthermore donors, on their part, should allow the government to be fully involved in the programming of aid, so as to ensure that aid is aligned with the government's strategic priorities. 4.9 All levels of public expenditure management (macro control, strategic allocation, operational efficiency) require that donor aid is integrated into the planning and budgeting processes of a country. "Integration" may be a matter of degree: * Complete integration occurs when aid resources are treated in just the same way as the government's internal revenues at each stage of strategic planning, budget preparation, budget implementation and reporting. * When complete integration is not feasible, it is still important (a) to take aid fully into account at the planning and programming stages, so as to ensure that aid and other resources are fully complementary and equally focused on the same national priorities; (b) to record and report actual aid flows and disbursements, so that all resources are effectively monitored, and the implementation picture is comprehensive. 4.10 As already noted, GoE regulations require that all aid flows be brought to account. This means that even flows which are not disbursed through the budget should be reported to the Parliament alongside regular consolidated fund expenditure. Actual practice falls well short of this objective: the PER team found that a large volume of external aid flows is still not captured in the budget. For example, of the forecast amount of Birr 3.6 billion in external grants, only about Birr 1.6 billion is captured in the FY99/00 federal budget. The situation could be improved in two main ways: (a) by more systematic efforts to include projects in the budget documents; and (b) by joint GoE/donor efforts to feed more complete and more timely information into the planning and budget processes. 4.11 Concerning the first point: there are no clear guidelines as to which projects and funds should be included when budgets are prepared. External loans are, as a general rule, included in the budget, because of the legal obligation to repay them and also because GoE has to budget the matching counterpart funds. For grants, however, no clear pattern can be detected. MEDAC's own summary notes that ".... .there are a number of externally funded projects that are simply not included in the budget. This is particularly the case where the externally funded project does not require a contribution 30 For the importance of including aid flows into planning and budgeting, see Mick Foster, "Integrating aid into budget management', Centre for Aid and Public Expenditure, Overseas Development Institute and Allister Moon, "Aid, MTEFs and Budget Process", World Bank, November 1998 (Draft). 49 from the Treasury."31 In practice this means that many of the projects funded by bilateral and multilateral donors remain outside the budget. Some channel 2 projects are included, others are not. Channel 3 projects are generally not included, though there are exceptions. Food security projects, some specific bilateral arrangements, and more logically, foreign aid channelled through NGOs,32 are also not included in the federal budget. Towards remedying this situation, this report has two main recommendations: * MEDAC should prepare systematic guidelines as to which projects should be shown in the budget and how. This would be an important topic for the forthcoming aid management manual under the EMCP. * The World Bank (chosen because of its international perspective) should try to identify other countries which are significantly more successful than Ethiopia in capturing aid, and assist in transferring relevant lessons from their experience. 4.12 As regards advance information on aid flows, this is needed in different forms and for different purposes at different levels of expenditure management (Box 4.1). There has been vocal concern about the difficulty of making aggregate budget forecasts when there is so much uncertainty about likely aid receipts. However, even if the total of aid flows were known for certain, it would not be possible to allocate domestic funds optimally or properly prepare for implementation of aid projects without much more specific information about the allocation of the aid funds and any conditions attached to their use. Moreover, such information is really needed much earlier in the budget preparation cycle than it is normally available. Box4.1: Aid informationi RequiremFents or Public Expenditure Management QObjective: "Aid information needed. Wy ..Macro......ii0ty.0 O)verall levels of aidfows WIplications forf balance of. ;ipaymerkts and Strategic Total resources.available throug aid perd. To eable the: governmentto ensure that alloaon priontized sector agggate expenditure tarets are sustainble; to i enstrethat thecombined distribution of'donor an raury.resources% is opiaan int lwewt Operational Advance notice of which projects will Neessary to allow irnely fulfillmient of pror efficiency be2: funded by which source And' vath conditions tiatin of 0gprocuretallocationg; what cmnitions attachedthis inludes&i of cmplementayreasry.fimas etc. pr ojects*which possibl stay outside the budget,flike those su rted *bN;s.i: 3 Forecast of External Funding for EFY 1993-95, MEDAC, December 23, 1999. page i. 32 It is almost impossible to capture NGO-projects in planning and budgeting since NGOs tend to work outside of and independently from government structures. However it is clear that NGOs - more than hundred of them are active in Ethiopia - contribute considerably to the country's development effort. In some regions, the NGOs, between them, may well bring in more resources than official bilateral and multilateral aid programs. 50 4.13 Work on the Macroeconomic and Fiscal Framework (MEFF) has, for the first time, led MEDAC to prepare comprehensive aid disbursement forecasts. The Forecast of External Funding for EFY 1993-95 represents the second such annual exercise. In Ethiopia, as in most other countries, there is no tradition of donor agencies providing multi-annual forecasts on a periodic and timely basis. In the past, MEDAC tried to get standard formats for quarterly donor reporting and forecasting accepted, but this method never really worked out. Instead, for drafting the annual report on forecasting of external funding, MEDAC relies on non-systematic methods for collecting data on external assistance. The bilateral cooperation department often uses information acquired during annual consultations with a development partner for the preparation of three-year disbursement forecasts. Data on expected multilateral donor disbursements are provided by MEDAC's multilateral cooperation department using the "managers' club" meetings as the most important source of information. At these meetings, implementing agencies discuss progress in project implementation with MEDAC and the donor. The project managers also provide forecasts on future disbursements, which are cross-checked by MEDAC departments. Usually, however, these data are not checked by donors themselves, resulting in inconsistencies. In recent years the variations between forecast and actual disbursements exceeded 50 percent. Such unreliable forecasting is one of the factors responsible for the unsatisfactory budget execution rates.33 4.14 Apart from the problems on aid forecasting, MEDAC also does not have sufficient data on actual aid disbursements - indeed the general lack of good quality consolidated data on aid flows is alarming. Actual disbursements should be promptly reported so that all the parties concerned have a full picture, at least ex-post, of where resources have gone. This is vital in order to be able to check that resources have been properly allocated and that actual allocations of aggregate resources are in line with declared sectoral priorities. Moreover, timely and reliable data on actual disbursements would provide a much better foundation for projections of future aid flows. 4.15 MEDAC's work on the collection of reliable data on actual disbursements is complicated by several factors: (a) many donor programs are very segmented and different disbursement channels and procedures are being used; (b) many projects are implemented at a regional or zonal level and these administrative entities do not always 34 have the capacity to report adequately on disbursements; (c) MEDAC depends to a large extent on supplementary information provided by line ministries (notably for the SDPs) which are not necessarily the most efficient in gathering and communicating the relevant data; and (d) unwillingness, according to MIEDAC, of some donors to provide information. 3 Of course, the sum of individual forecasts prepared donor-by-donor or project-by-project is likely to have an optimistic bias. Some projects always experience implementation difficulties, and so forecasts of aggregate disbursement usually need to be adjusted downward for this factor. 34 In some cases they may even have a vested interest in withholding information on actual donor disbursements (see discussion of the offset system below). 51 4.16 Under the PIP component of the EMCP (see Appendix 2) it is planned to develop a comprehensive database on aid to Ethiopia. This will be an important focus for improvements to the tracking of aid from commitment through to disbursement, but improvement of aid flow forecasts should be seen as a joint responsibility of GoE and donors. Donors can assist in a number of ways: (a) by increasing the time horizon of their aid commitments (it is inconsistent to avoid long term commitments concerning aid and yet criticize the government for failing to plan ahead); (b) by providing more accurate and timely data on actual disbursements; and (c) by providing feedback to MEDAC on its aid forecasts. PER 1999 recommended the establishment of a joint technical working group on aid forecasts. This has not yet been convened but should now be made operational as a priority - (see Box 4.2). It should also assist the regular collection and publication of actual aid disbursement data, including breakdowns by region (and zone) and sector. 8Box4.20 0 Go-Donor Working Group On Aid'Reporting And Forecasts It was& proposed during thePER-workshop in Debr Zeit inZ September 1999 tha a joint (oE-dono orkin group should be set up to examine wayskof improvingforecasting on aid flows. 'During a preparatqy PER mecting in Decee 1999 underthe. chairsh of the Vice-MinisterI o Fina the following composition wasagreed upon for th proposedf workiggroup: - represeisfrom theO Mistiesf of Finance, Economic Development & Cooperation, Health and Education, th romn yatRegional State, tee multiateral donors '(WorldN Bk, European Co:ommiission UNDP) and three biaterals (NetherlandsIreland and XFrance).g :Ef :S SS: : :f ; i ; f S :0: E; : E; EE; fIt is important and urgent thaiGoEIMEDAC)takes the inititivea to-make.ftis group operatioa and to includ& the impremet of reporting on aid flows into its mandate. Th eforemost ppnority of the grou-pshould be to rawupavork poa- acomawie by a timeale-e which cou cover the folloving elements: i to iden tifyand examine the main bottlenecs inihibitln MEDAC in collecting reliable-aid forecasts at the present tlime;~ ii. to develop foraiizedprocedures, ta new stan&dar format and a sandrd calendar for intimhation of foreass by donoSI fiii. to determine which informationj isexactly required andi for whichoprojects (size,channel, tpe, disaggreated9 by region andIcomponent etc); iv. to propose methods for' theimprovement of the collecion of data on actual aid disbursment; v.a to examine waysJto improve inrmin flows on externtal aid between donors, federal ~7goverhnment andregios,2zones-and woredas;: vi todO .t fo 0 00 an do .di K i &U Ai fl vi. to;identify the requirements for an operaton and comprehensie dataase on.aid flows; vii. toqact as a liaison between (jE anid donors assisting in the ireguar collection and quality conto of aid forecasts and reports. This should be a tasnding techical kworing group, 0prviding practical assistn in the ga g of aid dta, notia forum fr poliy discussions.01The calenda for: aid aita collection should be linked to the timitable of budget preparation and to the PE process. The group shold "ensure that compreliensive data on afctalind forecastaidflows-sare assemblin advanceoffftureP PER missions.i Ding discussions, the govenunent felt thattihe same objective could be tmetb b improving ad redeslgning thejj jexistiWngProjectManagers Club. This.difference of opiionllhas to6be resolved as part of the PER procss. 52 4.17 As a step towards strategic planning of expenditures, making aid forecasting more accurate and timely is necessary, but not sufficient. GoE also needs to improve the dialogue in the planning and budgeting process by sharing the draft PIP/MEFF and their supporting assumptions with the donors. This would give donors the opportunity to check and react to assumptions made by GoE on external assistance. Publication of the consolidated forecast would create peer pressure among donors to provide timely information. It would also facilitate better fine tuning between external aid disbursements and GoE's budget strategy. External Aid Flows And The Regions 4.18 Regional capital budgets are financed by a combination of federal subsidy and external loans and grants. These loans and grants are earmarked to specific purposes and projects, whereas the federal subsidy ("treasury money") is an untied and unconditional block grant. Table 4.1 indicates the shares of regional capital budgets financed through external aid (loans and grants) during the three-year period 96/97-99/00. Because of the fiscal developments described in Chapter 2, this year's transfers to the regions have been significantly cut (see Appendix Table 9 - Volume II). The burden of adjustment has fallen almost entirely on the capital budget, and one effect has been to increase the relative importance of aid funds. Table 4.1: Aid In Regional Capital Budgets - FY 96/97 - 99/00 Percentage Of Aid In The Capital Budget Region 1996/97 1997/98 1998/99 1999/00 Tigray 33 18 32 48 Afar 19 7 10 37 Amhara 20 12 18 60 Oromia 18 20 19 51 Somali 19 15 18 25 Benushangul Gumz 19 10 18 22 SNNP 33 29 24 63 Gambela 21 8 15 23 Harari 7 8 12 23 Addis Ababa 40 42 42 36 Dire Dawa 34 17 24 35 Total 26 21 25 44 Source: Regional Planning Department, MEDAC. 4.19 All regions have a very explicit preference for treasury resources. Unlike, aid funds, they are considered reliable as to their availability and timing. They are also more flexible - e.g. funds may be switched from a slow-disbursing project to one that is proceeding faster than anticipated. Consequently, implementation rates for treasury- funded projects are strikingly higher than for aid projects. All these points are illustrated for SNNPR in Box 4.3. 53 0;i;000Jinx 4.3:f0jiig SNP Regional000 B0000-- Bxii,.00t udges - ilThe lStoy The FgrsTeSll00tj9020Xiig|00000000)Q ;~~~ ~ ~~~~~ ~ ~ d t,&4 T.,EfiiIiSERL 35Othero SnclPds recurvets and catibuts for p9ensiotan 990ds. well as on t l til n of budThe estim'ted reurrent of p treasuay ounth s e of prnan in capiulte ex efdres). -- Tresuy fnd no uedbuheedge Alofhicatio yand Epctned retilipratedbi rgoian.toiie o "capcit buldig" urpses(pjurcaes i f paetequismindcte traicigent.es tot total) rcie a oedbh PER team4;0 .. that visited Amharin Recurr5000 00040 entt C4:ap40i4tal P A0 002lanned utilzatSS); SSiont;))f EY98/99 FY99/ARSD ::2;0 dA;2tCaa 000 Sour -cC)DS;:: R0 of Finmncin22 EY901f0:S/992 FY99/00a):SLR Adm~~ ~ ~~~~~in1 0l29.7 -01:12.6L iQ0|0 Tras- 237.$0X40giik00 4 7w8.572iiiS;Bl iliile::0004(23.9) (22 S.-, (76 s."0 .2)Xd (37.2) OterS 40. 44.4 reurn brde Th remainderg ofl;j th recurrent budget i fiane th00 m rough j)Z re:;leSygionlrvne(198)Thbiggelst (11.8'gE pe20rcet). Th main::22 vic00fl:Ntims of teSSD00 rd cdfnanilfos :torAwae; are the Abwater ;seco an sustinalew 0~~~Atog hettl(aia +4 recrret) budgto SNN hasbenreuedb approximtel Bin-4 130 ;0~~h cptal budget xgtho0;ht hasgone dow byior 0thaEiirs 100millio an thReion w;4B04000;00fBB0000illt; only00000 be;00 0 .Th capital bud0|get's resuy shr 0wAg- cosidered rl0iable as t ailbity an timin 0ig - ha tt been 3Other Reinalfcalsuals resevesaled cothabt,itemofcpalbgthRgionshsbenal to spendo onuvergesve 36 80.0 percma e ntilzto of theafsuproieby Treasur atransfer;36i acuae bfoecpctybidn" *rau 366d penotusdbthen of the fundsl yacnberpprovidaed by external loans;tie and "cpct bu31ng percetofste ns (purcaeov qid mn,tann t..Asmlrpatc aoed by exerahgans PRatesai theauret yiiedArna. 54 4.20 Considerable delays in the disbursement of donor funds included in the annual budget - mainly for reasons related to donor imposed conditionalities and implementation requirements which the different layers of government find difficult to meet - have a particular impact on the availability of resources at the regional level. Donor funds included in the regional budgets are offset against the treasury subventions and aid disbursement delays hamper the regions' ability to execute their budgets in conformity with their work-plans. Donor requirements and the offset system therefore emerge as the two big bottlenecks for the implementation of regional budgets and are discussed in greater detail below. Implementation Of Donor Aided Programs 4.21 Serious difficulties and delays often occur in the implementation phase of externally funded projects. Even after a project agreement is signed, delays occur due to shortcomings on both donor and GoE side. The procedures on the donor side for approval (e.g. on tenders, awarding of contracts, etc.) are often lengthy and bureaucratic. Implementing agencies, especially at regional level, complain that projects are delayed because of donor-imposed conditions which GoE finds difficult to fulfill. This is at least partly a structural problem affecting project design: the donors' concern to ensure that all systems and procedures are technically and financially correct often militates against the efficiency of project execution. (see Box 4.4 for an example). Box 4.4: The ADF Credit F4r ESPD Support The ADF credit for suDport to the, ES)P (worth approximately USD 40m) is a -striing example -f delayed project commencement. The prognra agreement was signed.in July 1998,-and subisant disbursements were included in the federal budget for FY99/00 and, consequently,o9ffset in.the treasury subsidies to the regions. -Disbursement was to be via Chinnel l, but:with fiunds earmarked to specifc projects (in many respects echoing the World Bank's ESDP proceduresre However, GoE has- not yet met all the project's conditions precedent, and consequenily the credit has not yet becoome effective. Under ADB regulations, project agreements would'lapse if they have not become effective within two years. Here the consequences of failing to satisfy the prior conditions are very visible. At one level', this indicates a capacity problem, and capacity constraints -at regioal level. are certainly a factor. 'However, the conditions imposed are very demanding. Althiou,gh thi project is regarded as part of a sector program, all its funds are earm,arked to-specific primary sebools, which bave to be identified down to site level before the credit can become effective. And, although, implementation will be handled at regional and zonal levels,: all stages of implementation (tendering, pocurement, payments) are subject to prior approvals. Thus it is also legitimate to question the appropriateness.of the project design, and, in pDarticu the extnt and the rgidity of the conditionsimposed-. 4.22 Even after a project becomes effective, the complexity of implementation procedures (e.g., on disbursement, procurement, reporting and monitoring), combined with capacity constraints at all levels of government, make it very difficult for implementing agencies to meet specific donor requirements. The ADF and IDA credits for ESDP and HSDP are often cited by regional officials as being particularly fiustrating. 55 The issue of the SDP modalities and their potential inconsistency with the principles of decentralization is reviewed in Chapter 3 on information management.37 4.23 In general, implementation problems related to aid projects will not easily be overcome, but both short and longer term remedies should be pursued, including: * In the long run, more unearmarked funding though sector programs and budget support. Joint GoE-donor efforts are needed to develop modalities for budget support and sector-focused aid that is not tied to specific projects (see the later section on fungibility, and also the discussion in Chapter 3). * Meanwhile, further measures are needed to simplify existing SDP disbursement and reporting arrangements and to align them more closely with GoE's own procedures, which are generally functioning well and compliance with which is within the reach of administrative capacity available at the implementation level. * An urgent effort by donors to move to multi-year programming of their funds, so as to give implementers a longer horizon for organizing project implementation (as well as, at aggregate level, to increase confidence in macroeconomic projections). * Steps to support capacity building and coordination, especially at the regional level. Training of local staff of implementing institutions on specific donor procedures needs particular attention. When talking about capacity constraints, one should not, however, focus only on the different levels of government: resident missions of donor agencies themselves frequently face shortages in human and technical capacity. * Many implementation problems stem from misunderstandings on or different perceptions of conditionalities and procedures. The intensity and efficiency of communication flows, between donors and federal and regional governments, related to conditionalities and disbursement and reporting mechanisms should therefore be improved. The Offset System 4.24 The offset issue has been examined more than once by previous PERs, which have recommended modifying the system, as have some other studies.38 Because of continuing concerns, it has been decided to take up the subject once more in this PER. 37 The 1999 PER mnain report contained in paragraph 118 some recommendations on the improvement of the programming and management of IDA-funds. Thus far, these recommendations have not yet been implemented, but discussions on methods for making the IDA credits more effective are ongoing between World Bank and GoE. 38 The 1998 PER set out how both the treasury subsidy and the offset are determined and applied and discussed its merits and demerits, while the 1999 PER made some recommendations on how to improve the 56 4.25 The budget offset system is intended to ensure inter-regional equity in the availability of total resources for development, as required by Article 94 (2) of the Constitution. Essentially MEDAC reduces the formula based federal subsidy due to each region by a factor reflecting the expected aid flow into the region in the coming year. The underlying principle, that each region should get a fair share of resources, and not be arbitrarily much better or much worse off than other regions because of the way donors happen to distribute their funds, is incontestable.39 Whether the offset indeed achieves an equitable allocation between regions is hard to assess, because good data on aid flows by region are not available.40 The concerns about the offset system are practical: does it discourage regions from taking up available donor funds with adverse consequences for the country as a whole? 4.26 Although its objectives are prima facie rational, there is little doubt that the way the system operates in practice has undesirable side-effects and gives rise to perverse incentives. First, in the regions, there is already a decided preference for treasury funds over external aid, for the reasons described earlier. Applying an offset tends to reinforce this bias against aid flows. Consequently, the goal of maximizing the inflow of total external aid into the country is compromised. Second, the system gives an incentive to the regions to attract donor resources that are less likely to be factored into the offset calculation, like for instance NGO financing and, in general, channel 3 aid. The federal government's preference for channel 1 types of external aid could be frustrated by the regions' attempts to convince donors to apply disbursement procedures which will not be accounted for in the budget. Third, regions may also want to refrain from reporting on extra- budgetary donor support, so as to avoid the penalty of a reduced subsidy. 4.27 MEDAC, responsible for the application of the offset principle, is aware of some of the problems thrown up by the offset and has made regular adjustments in order to try to improve the system. In FY99/00, MEDAC altered its calculation of offset. Previously, the components of donor support were examined and some elements left out e.g., technical assistance. However this year, MEDAC offset a flat 30 percent of external grants and 70 percent of external loans. This reduction of the offset percentages is a substantial improvement in itself.41 The higher percentage of offset for loans, as compared to grants, is also rational since loans carry a repayment obligation. If the principle of offsetting only a portion of aid received becomes established, it will reduce the most serious perverse incentive. offset system in the light of substantially lower donor disbursements than expected. The topic is also explored in, e.g., Ethiopia Regionalization Stuady, World Bank 1999, Section 2B: 39 According to MEDAC, "offset is a means of balancing treasury resources, external loans and external grants in such a way that the overall outcome of the treasury subsidy formula will be maintained to a large extent." 40 A second objective the Federal Government seeks to attain through the offset system is to maintain control over aggregate expenditures, including donor finance. 41 Rational regions now have to consider whether a larger amount of 'difficult' donor funds is worth more than a smaller sum of 'easy' treasury money. When a given amount of 'difficult' money replaced the same amount of 'easy' money, the rational response was much simpler. 57 4.28 Any specific criteria adopted can be questioned (e.g. why such a large difference between the percentages applied for loans and grants? Why exclude some types of aid- like the Ethiopia Social Rehabilitation and Development Fund (ESRDF) and food aid- from offset calculations?). But the most serious issue-the strong disincentive for regions to seek funds-has been addressed at least to some extent. Further improvement of the system is, however, desirable and possible. The calculation of the offset should be based on recorded actual ex-post levels of donor disbursements instead of ex-ante estimates of aid flows. When this exercise is conducted annually, regions will retrospectively be compensated for lower than expected donor flows on a moving average basis. As the external financing share of regional budgets varies significantly, using historical rates will enhance GoE's principle of regional equity. It would also have the additional merit of capturing supplementary budgets, which at present are totally additional. It may even capture off-budget donor assistance retrospectively. More attention should also be paid to the issue of establishing better data on flows of resources (including aid) to the regions, and linking this to analysis of impact (the topic of Chapter 5). 4.29 The following steps should further contribute to improving the offset system: i Maintain the principle that offset should be substantially less than 100 percent of the funds received (in other words, regions which go to the extra trouble of utilizing donor funds, and thereby expanding the total resource envelope for Ethiopia, should get some reward for doing so- this is the same principle as the revenue element of the basic subsidy formula). B Base the offset on actual recorded aid flows in the preceding year minus one. This would be transparent, and would also be seen as fairer: regions are particularly fiustrated when the offset reduces their funds on account of aid funds which were expected but did not, in the end, arrive. * Determine, and publish, clear guidelines on what is and is not included in the offset calculation, and avoid frequent changes. (Although MEDAC informs each region of the amount of its offset by donor and sector, it should make additional efforts to provide complete information to the regions-and the donors -on how the offset is exactly calculated.). * Publish comprehensive annual figures on actual aid flows by region (and if possible by zone).42 This would enable the equity of resource distribution to be monitored in a transparent way. Fungibility 4.30 Fungibility has become a prominent topic in the aid debate; but it is a much misunderstood topic.43 On the one hand, some government officials argue that fungibility 42 Ths is especially relevant for regions which devolve budgeting to the zonal level and divide their federal subsidy among the zones accordingly (see the decentralization section of Chapter 3). 4 Given special prominence byAssessingAid: What Works, What Doesn't and Why? World Bank 1998. 58 should not be raised as a problem because government accounts are transparent and aid funds are not diverted from their agreed uses. Some donors, on the other hand, appeared to think that the tracking of funds to particular projects would address the problem. Both reactions show a misunderstanding of the concept of fungibility. Box 4.5 attempts to clarify the definition and its implications. The crucial point is that fungibility is not about what is done with aid funds directly, but about what they may finance indirectly. Box 4.5: Fungibility - What Diflerence Does Aid Make? Fubility arises beause resources can be substituted for eah otr. When aid is pvided itrease the total resources evailaba to the reipient If ai-d fiacs ary activity tat th recipient would otherwise hae-nne itai the resources thiat thle recipiet woldhave spent on that activity becomeavaial t lancesoetin es.: A simple example:. Donor A povides couty B with whchare a a to buldR3ad . R Xi r most irportat rod int countr and i o of the govenn s p,roect pioity list'. 'f ad had aot e avabk the Government wouldhave financed therad itelf. Bcaue it has aid fo , the n us n fuds for the ne pricr Road Y. Now, whatdid th aid f ? atwo an r, t t The afdfinanced RoadX. he road has beeb, nd thea recds ae to shw that thLe contraor was paid with tends t he'donor pTovid, ad ath tose is w' t used fir a,ny othr purpose. The aid fanced tad . In the 'without aid! situation, the governmen woXld havedfinancedRoadXIC itself Because of aid, the government two ioads instead of oe, ane deree wit idf is -. thatRoad Y has also been built. hat is fangib- and there are two important points to nte: * Fungibiliy dsnodependon te ecien aaid fuds oeag any aid ams Earmarking of aid funds to a specific project doe norpreve fungibihty Some donors may be content to follow h ri of Their' funds, +chec that the fnds wereide spent as speified ini theprojectagreement, an'd ask nofrthejrquestions., te mayonder, 'at d.rencedd ourraid make?? which is no,t an easy question to an,s,we because: (a)-W allr th ipientexdtus hav.e iibe conidre7 id (b), ven if adequxate data are aailable the actual \v ith aidi' situationt needs to be compared Wit a imptetcl: 'withot aid' situation. A rigorous answer= at national level would rqieacareful analysis- of macroeconomic data and yaggregaes, using siutaeus economic moes to captue th fl-ais rne of4 aid' ...hc;s Howmuch does flangibility matter? AsesstegAit' has two reeat conclusions: don bosshud lak it for rnedta ther finacing is- funible, because tat is. th reality, but in ptactcet flinility -is an issue olif: the obetives of donors andrxecipients are ifferent. - - : 44 See Howard Whute, Macroeconomic Effects of Aid: Literature Review and Mehtodological Framework. SASDA, The Hague, Janauary 1994, and The Implications of Foreign Aid Fungibility for Development Assistance, Shanta Sevarajan and Vinaya Swaroop, Development Research Group, World Bank, October 1998. 45 Assessing Aid, ibid, quotes from p80 and p62. 59 4.31 Fungibility has been one of the reasons for donors to consider supporting sector programs rather than individual projects. If the earmarking of funds to individual projects is, essentially, an illusion, it is hard to justify the heavy additional costs that are involved - the burdens discussed in the implementation section above. A framework in which government and donors conduct a dialogue about an entire sector may be a much more effective way of ensuring the best use of aid - particularly when government and donors share the same objectives. At the same time, fungibility is at the root of many donors' reluctance to provide aid to countries in conflict. It is feared that aid will indirectly finance the increase in defense expenditures, or will be seen as doing so (in this case the perception may be as important as the reality). 4.32 As Box 4.5 shows, earmarking of funds does not get to the root of the issue (although it may provide psychological and political comfort in some cases). A better approach would be to make the underlying concerns explicit and encourage donors and govemment to discuss practical ways to address them. A first step is to clarify the nature of the problem, so as to avoid sterile debates based on misunderstandings. Secondly, transparency on public expenditures as a whole provides the factual basis for discussion of donors' concerns (in this connection, the Government's openness to the PER exercise is commendable). Thirdly, donors need to consider more precisely the conditions that would satisfy their concerns (and the concerns of their domestic constituencies) about fungibility. Would their concerns be allayed, for example, if the government could guarantee certain minimum levels of treasury funding for the sectors donors are supporting? Or if government undertook to match donor funds in an agreed proportion? And what independent verification of actual government expenditures would such an arrangement require? The fungibility concern reinforces the case for making strategic expenditure plans, the national budget and its implementation the subject of regular and institutionalized dialogue with donors (See the last section on government-donor dialogue). 4.33 At present, even among donors - let alone between donors and the government - there is no shared understanding of how the fungibility concept will impact on future decisions on donor disbursements. It is therefore proposed that a donor-GoE workshop - eventually followed up by a specific study - be held with the objectives of having an open exchange of views on the issue of enhancing GoE's ability to respond to donor concerns related to the fungibility of aid flows. If it is accepted that fungibility of aid should be a matter of concern only if the objectives of donors and recipient governments are different, the aim should be to ensure greater clarity about the respective objectives of government and donors, and to devise aid modalities that align and focus on common objectives - a theme which recurs in this and other chapters of the PER. Food Security 4.34 Ethiopia suffers from both chronic and transitory food insecurity. Chronic food insecurity is the condition wherein households have insufficient food production or income to meet their basic everyday needs even under normal circumstances. Transitory 60 food insecurity is mainly the result of periodic droughts that occur with tragically surprising frequency in certain parts of the country. The number of people facing chronic food insecurity is estimated to be between 4 to 5 million, and the majority of the affected population also live in the same drought prone areas as the transitory food insecure. 4.35 One indication of the extent of food insecurity in Ethiopia is the level of emergency relief food distributed annually. The average annual food aid dispensed during the period 1985-98 was 703,000 metric tons, which is equivalent to 9.6 percent of domestic food grain production in the same period (Appendix 4). The Annual Emergency Appeal, launched in January 2000 by the Disaster Prevention and Preparedness Commission (DPPC), estimated that 8 million Ethiopians would be in need of food aid in 2000 and that the total food requirement would be 821,835 metric tons. 4.36 Food security, as defined by GoE, has three main components: supply, entitlement and emergency response. The supply component focuses on agricultural research, extension and input supply and has been the focus of GoE expenditure. GoE has also funded the recurrent and much of the capital costs of the DPPC and the Emergency Food Security Reserve (EFSR), both of which address emergency relief. However, food commodities for relief and development interventions have almost entirely been financed by donors, although GoE recently committed itself to the local purchase of 100,000 metric tons of food as its contribution to the above mentioned 2000-Appeal. The third component, entitlement, is to be addressed by the National Food Security Program (NFSP). Until now, the chronically food insecure population's immediate needs have been met through the annual emergency appeal. 4.37 GoE presented its Food Security Strategy to the Consultative Group meeting held in Addis Ababa in December 1996. The strategy discussed the three components mentioned above and identified possible responses to the problems in the respective areas. In mid-1998, MEDAC circulated an NFSP document for the period 1998-2002, which was based on the 1996 Strategy, albeit with a narrowed focus. The approach was chosen to concentrate efforts on addressing the food entitlement (access) issue, the argument being that GoE already had in place programs and policies to address agricultural development and production, in addition to the emergency response mechanism of the DPPC. The NFSP is designed to provide additional resources to specific targeted woredas in the four main Regions: Amhara, Oromiya, Tigray and SNNP, based on regionally developed food security programs. The estimated resource requirements for the initial five-year program were Birr 6.08 billion. 4.38 Admittedly, progress in pursuing the dialogue towards a comprehensive NFSP has been slow. The government has, understandably, been quite wary of initiating discussion with donors on basic policy issues in the strained atmosphere engendered by the conflict. Meanwhile, some regions have started implementing their own food security programs and have established Food Security Coordination Units to channel donor 61 46 resources and to coordinate ongoing programs. A few donors have made commitments to fund components of specific regional programs, but the lack of an operational national program discourages the donor community from making major commitments as they recognize the need for a national coordination fonnat for resource allocations and policy dialogue. Absence of donor initiative could also be largely attributed to their concern about the border conflict. 4.39 In the absence of significant investments in the targeted food security program, chronic food insecurity continues to be addressed through DPPC's Annual Appeals to mitigate disasters. Arguably, this is an insufficient response to the problem of chronic food insecurity. However, as indicated earlier, donors are becoming increasingly reluctant to respond positively to the Annual Appeals for emergency assistance unless, as seems to be the case in 2000, inadequate or delayed donor response could lead to a famine. In addition, donors are also insisting, increasingly vociferously, on at least a minimal government contribution to its own food appeals. At the same time, donors recognize that addressing the root causes of chronic food insecurity is a long term effort and programmed food and cash transfers will have to continue in parallel with targeted development interventions. 4.40 Many donors are sensitive to the fact that GoE can at best make only minimal contributions to the NFSP in the near future. Formulation and implementation of NFSP need not, however, be held back on that account. Donors recognize and accept that community participation in food security related development projects, in particular through labor contributions, could be envisaged. At the same time, community participation in the planning, implementation, and monitoring of projects could address ownership concerns. There is, however, an expectation that GoE will open a dialogue with donors on the more critical policy issues relating to the root causes of chronic food insecurity. 4.41 In view of the above, and in order to ensure more timely and appropriate donor response to emergency appeals, the following recommendations are made: (a) GoE should continue, as it is doing this year, to make adequate and timely contributions to its own relief appeals; (b) Chronic food insecurity and development related projects should be addressed, not through emergency measures, but through programs included in the annual budget, irrespective of the sources of financing (external or GoE's own resources). The fact that, at present, food security allocations to the regions may not be based on the budget subsidy formula and, therefore, not subject to the offset-system, should not prevent GoE from including resources planned for tackling chronic food insecurity into the MEFFIPIP and the federal budget. Figures about the average number of people being food insecure on an annual basis are well known and can be factored into the planning exercise well ahead of the actual disbursement. (c) The NFSP - albeit perhaps in a modified version taking present economic and financial realities into account - should 46 Food Security progrms of Ororniya and SNNP are not yet well defined in comparison to those of Amihara and Tigray. 62 become operational as soon as possible; (d) The public expenditure implications of food security should be studied in greater detail as part of the PER process. Government-Donor Dialogue 4.42 Aid relationships in Ethiopia are in a transitional phase. GoE has taken commendable steps in seeking to ensure that aid is properly integrated with other resources and in developing sector-wide approaches that can avoid the weaknesses of traditional, fragmented project aid. Adopting such a broader perspective requires a continuous dialogue between GoE and donors. Some of the mechanisms for dialogue have been put in place, especially with respect to the SDPs. Besides, the PER itself has evolved into an important forum for reviewing, and helping to resolve, issues that are more systemic or macroeconomic than those which the SDP mechanisms focus upon. 4.43 However, and apart from the specific fungibility and food security issues, there is insufficient understanding and appreciation between GoE and the donor community on mutual concerns related to aid management. At present there is no institutionalized basis for dialogue on aid management subjects between GoE and the development partners as a group, except for those issues related to the implementation of the SDPs. Establishing such a system will go a long way in improving aid relationship.s. 4.44 The meetings of such a forum should preferably be linked with important stages in the planning and budgeting cycle (e.g. before the presentation of PIP/MEFF and the Federal Budget). The sharing of documents like PIP/MEFF and budget with the donors, and allowing the latter to check assumptions on external aid flows, before they are adopted, would take the partnership to a higher level and would enable donors to strategically link their program into GoE's fiscal framework. 4.45 Some other African countries, notably Uganda and Tanzania, have started experimenting with such a partnership dialogue on a comprehensive budget framework. The MTEF, initiated by the Ugandan Government in 1992, provided a reference frame for such a dialogue. The scope and dimensions of such a dialogue were considerably expanded with the Ugandan Government presenting its entire budget framework at a workshop under the PER process in June 1999. The workshop, bringing together virtually all ministries of the government, as well as the donor community, provided an unusual opportunity to discuss the sector frameworks and the entire budget strategy. Such transparency has several advantages. First, it builds donor confidence in the government and encourage them to align their allocations to the MTEF strategy. Second, such an exercise increasingly obliges donors to shift their assistance from project aid to sector support or even free standing budget support. Third, it challenges donors to make long term commitments and improve the predictability of their financing.47 Finally, consultation with the donor community on the budget framework would also entail more 47 See Aid, MTEFs and Budget Process by Allister Moon, World Bank, November 1998 for a description of the Ugandan experience 63 explicit consultation with the legislature and publication of the framework on a more explicit and timely fashion. By raising the political cost of any subsequent breach in the framework, the process ensures grater stability in the framework. 4.46 GoE should start moving towards such an enhanced coordination process with the donors. Apart from the advantages mentioned in the Ugandan example above, such a dialogue will also fit in with the coordination structures that will be required for the formulation and implementation of new initiatives like CDF and PRSP. ~Box4 46: SummayOf Conclusi'ons AnidRecommendations -AREA FIDI~NGS REO2rmMENDATIONS RESPONSIELITv AND TIMING~ ointeraton of *Lag share of exera i To establish clea i. NWME C short: aid low naidremins ff udgt gideine asto which external and mediumtr i audfeting guidrelie on th budgtetin proesses wok .w mnulcn inlsofaidflwei database) Em pi atd:: the bugtid. to sekadsaegoii. Wrdfak&hr arming -V9 ~~~~~outr eaplsocatrn te. bii gI*line& 6:41 the bu~aid nPthe budet. k diihalf forecasting of -cusonsoidfated1dtaongedtehia woringgrup.mmdiate, wit aid flows act~uaai dibusmet (TWG) inclde reporin ad co bprld Bono other MEDA colectsdataon frecats i itsTORgr pmebes f~~~~ orecastino o s onia dtaono datarfor MEF and 0 fornet upor ofTWGn Aidfl anjadidbursemensertendto-i be W 10 PERrmssio donoprs.n0,,C uxenrlableadin Avi. TG oernmentan sdoenor i. hr emIe * GoE, ~at peet osnt sol disus ar6id foeast laeV etino i anssumtions oan extra inlzd aidewith mdonors bdor vii NEstbls axi vii.io SorflMGed Hm- adisbudgetnet:e dM b~i EDAC suppn-d(nortsfE poet Implementation unreimple L Gvrmenta fad vi. Cniu effort rs to VIii Meiumtem of donor fund~GEd,activseitioesnoften mv thowardiscunearid rked at donors in dalou it programs d~syselhate. ocapacity isS7 secto EF and budget arppor,wt o ias omtime evient dieveo oaiis donoprioceurs and ixeoxteosit oad i, Sot/ei codtonos ma re muI-ya rormigIftr; oos inappr oprae coEFFPlex toprh,eirfndsiv bs.t ti DC t 41P n 1an Chetoo tigid. ai VI Coni effortsito x. Sotadmeim db4'9]r fd&d ktiviies bfsimpl tifyandhroie em o annakd dnp di uonois, procedures,une,di g o To. 0 capacit ~~~is Support Su caacty i xi. CotnosaG~. a factq~~ tOObuiling GEand tefipoeet (fdfladrei)pu conple: teiof cmmnicds, fosdoos betweend don niors eea and,Gdwdnos; Redona (loerment 64 14~~~& tF m 4. -U : - - g;- %-t i4 §-- :-, 4 $t $;0 X Zz rz .4 9 - .. . s4 c,...fi - .:-s::- s::-5> . , _K7iTgig*fiSi ,fB E3Yltti- ! 0 ~~81c- t4 v~ :s Cfi t5lt 5 X ;Bsg y a,i . .. to - -: . : . ': . ..... .. ' :: , : ' -: .. , :! j: . ' , - ' : ,, . - . . .. '.: '' :, : '. - - 0 . . * . .: X - . - T 4 : -- -. . -. : - I 5. PUBLIC EXPENDITURE - SHIFTING ATTENTION FROM INPUTS TO RESULTS INTRODUCTION 5.1 Increasingly Public Expenditure Reviews are moving from an analysis of spending patterns focused mainly on issues such as the sectoral composition of spending, or the balance between capital and recurrent spending, to a more precise assessment of the results of these programs. This reflects a trend in both developing and the OECD countries toward evaluating the benefits of public spending relative to its cost, as a means of improving both the overall cost-effectiveness of spending and the accountability of the public sector. Program evaluation is not new, but the conscious attempt by governments to assess and compare the returns on different programs by some common yardstick such as effectiveness in achieving key results is now gaining momentum. 5.2 Two interesting developments are giving impetus to this trend in developing countries. First, decentralization, which gives more power to local communities to make spending decisions in accordance with local priorities, is increasingly viewed as an essential element of an effective poverty alleviation strategy.48 But when decentralization is supported by large budgetary transfers, good governance requires that the results be clear for all to see. Second, donors are trying to secure stronger public support for debt relief and more generally development assistance by committing to global development targets and channeling their financial support towards countries that establish a credible record in good governance and poverty alleviation. 5.3 Given the priority Ethiopia accords to reducing regional disparities in development and improving the welfare of the most disadvantaged groups in society, a focus on results promises to be an area for fruitful dialogue. This chapter briefly reviews Ethiopia's experience to date with tracking results and the reasons why an expanded focus on this may be beneficial. The chapter then suggests steps to improve current practice as well as actions to move beyond current practice so as to achieve greater gains in transparency, efficiency and accountability. Ethiopia's Experience In Tracking Results 5.4 As discussed in Chapter 3, Ethiopia, in stark contrast to many countries, has a rigorous financial management and budget coding system that makes it possible to track the public funds being spent on sectoral programs by the center and the regions. Furthermore, the government has made a big and relatively successful effort to track donor expenditures and incorporate them into the budget. In recent years, Ethiopia has 48 It should be noted though that decentralization has not always been a totally benign option. Its success would depend upon effectively guarding against its pitfalls. made considerable progress in looking at results as a way to monitoring public expenditures, and indicators have been identified and used esyecially in the context of the Sector Development Programs, in collaboration with donors. 9 * The Education, Health and Roads Sector Development Programs specify performance indicators to be monitored and targets to be achieved for these indicators for the year 2001/02. Both the Education and the Health SDPs use gender-disaggregated indicators (the indicators used are discussed at greater length in Appendix 5). - The Program Action Plan for the Education Sector Development Program (ESDP) includes a set of indicators that aim to capture both financial inputs and outputs/outcomes (access, quality and equity) of the education system in Ethiopia. The indicators have performance targets for the year 2001/02 (the first year of implementation being 1996/97). There were no annual targets set but these are gradually being developed. The regions also have similar indicators included in the Regional Sector Programs and have 5 year targets for each. - The Health Sector Development Program (HSDP) uses considerably more perforrnance indicators than the ESDP (but the last Joint Review Mission was unable to obtain, in consultation with regional health authorities, reliable figures for many of these performance indicators). - For the Road Sector Development Program, sixteen national indicators have been identified and baseline data for these indicators collected for 1996/97. * In sectors not covered by SDPs, monitoring of results is less advanced. The mission was not in a position to make a comprehensive evaluation of the use of indicators, but learned of some encouraging examples. - Ethiopia has been rapidly extending a package of modern agricultural inputs and credit to traditional farming areas. Progress has been carefully monitored and the results and regional experiences have been shared in national workshops. Similarly, some output indicators in the agriculture sector are monitored in the context of ongoing donor- supported projects. For example output/outcome indicators such as changes in (i) production and yields (kg/ha) in crop and livestock; (ii) number of small farmers using new technology; and (iii) accelerated 49 Tlis report adopts the following terminology for indicators: input indicators (expenditure flows, means), output indicators (physical results, products/services delivered), outcome indicators (access, use and satisfaction of available outputs, e.g. enrollment rate) and impact indicators (final result of use of services, eg. literacy level). See: Result-based management in CIDA: Introductory guide to concepts and principles, CIDA, January 1999 and Sourcebook for Poverty Reduction Strategies, World Bank, February 2000 (). This report focuses on output and outcomes indicators as their link with public expenditure can be established more directly than in the case of impact indicators. 68 transfer of new technologies to extension services, which are being used, appear to be suitable for result-oriented monitoring. - Some output/outcome indicators were selected in a Food Security Program (FSP) prepared in 1998. Given the comprehensive nature of food security, these indicators could also serve the requirement of public expenditure in the agricultural sector. * In addition to the indicators used in the SDPs, indicators are also included in the development index of the government's budget subsidy formula used for distributing the budget subsidy horizontally across the regions.50 The development index is based on output measures in six sectors.5' The indicators used measure existing assets (outputs) for health, education, roads, water, electricity and telephone coverage. Outcomes (use of outputs) are considered only in the case of school enrolment and under-5 mortality rates. * Ethiopia has also established a Welfare Monitoring System (WMS) to monitor the social consequences of its economic reform programs, and the impact of the programs on the poor. The WMS generates monitoring indicators and is built around an annual review process whereby the data generated by the system is reviewed annually and the policy implications considered. The main sources of data for the system are the annual household welfare monitoring surveys undertaken by the Central Statistical Authority (CSA). A Welfare Monitoring Unit has been established in MEDAC to coordinate the production of periodic analytical reports on welfare. Why Shift From Inputs To Results 5.5 This emerging emphasis on results of public expenditure follows a trend observed in other countries, New Zealand famously among them, that have successfully used performance indicators and other budget management tools to improve service delivery and enhance the efficiency of public policy. A useful parallel that helps think in terms of results is one which compares ministries and government agencies to companies. In this perspective, governments buy goods and services and delivery systems from line ministries and other institutions; budget allocations are the price that they pay; the transactions mimic contractual obligations; and buying is guided by the overall needs of the economy as defined in the government's public policy. 50 These block subventions to the regions currently account for about a quarter of the total federal budget. 5' The index contains indicators of: education (pupil section ratio, pupil teacher ratio, and primary aad secondary enrollment rate); health (ratio of clinics, doctors, nurses, health assistants, hospital beds per thousand people, and the under five mortality rate per thousand births); roads (length of road per thousand square kilometers); electricity (kwh per thousand population); telecommunications (telephone lines per thousand population); and safe drinlking water (proportion of population with access to safe drirning water). 69 5.6 Arguably developing countries face greater difficulties than developed countries in effecting this shift from inputs to results for a variety of reasons. Typically, in developing countries, institutions for enforcing transparency and accountability are weak; capacities for designing and measuring output and outcome indicators as well as for enforcing the hard-edged contractualism characteristic of the developed country model are limited; and markets are often not mature enough to give the right signals of the willingness and ability to pay for public services. Nevertheless, with the encouragement of the donor community, a number of the countries participating in the Highly Indebted Poor Countries (HIPC) Initiative, Burkina Faso, Mozambique and Uganda among them, have also begun to track results. 5.7 Shifting from tracking expenditures (an input into the development process) to tracking what expenditures produce (outputs) and their impact (outcomes and impacts) is advisable for a number of reasons:52 * In the context of budget fornulation, output and outcome indicators could become the basis for spending ministries to justify their budget proposals and for the ministries of finance and planning to scrutinize them more objectively. This would not only facilitate more efficient budgetary outcomes, but could also become the foundation for building contractual obligations between the government and the sector ministries. • Emphasis on output and outcome indicators will improve the technical efficiency of public expenditure management. Defining and agreeing upon results in qualitative and quantitative terms would make it possible to give managers of sectors/zones and programs a higher degree of managerial responsibility and flexibility regarding the means used to achieve them. Indicators of results would then become a tool for management to detect off-target situations at an early stage, to enhance financial accountability and transparency and to assess the managers' performance. * A shift in emphasis from inputs to results could also enhance the allocative efficiency of public expenditure across and within sectors. Budgets are traditionally made on an incremental basis, based on spending patterns and established costs. The accent of line item budgeting is more on control than on results. On the other hand, efficiency demands that resources be deployed where their marginal impact is the largest. This requires an understanding of the inter-sectoral and intra-sectoral trade-offs and complementarities. The problem in Ethiopia is three-dimensional: allocating resources between federal and regional governments, at each level between capital and recurrent expenditures, and across sectors. These allocative decisions require indicators that measure the contribution of 52 See footnote 49 for definitions. 70 sectoral outputs to broader development goals like food security, reduction in disparities across regions, health and education status etc. Defining such indicators, measuring them and feeding that back into public expenditure decisions would help in aligning intra and inter-sectoral as well as inter-regional allocation of resources with the broader development objectives. Attention to output and outcome indicators can help in aligning external aid with the government's development strategy. Donors are typically under pressure from their home constituencies to show results for the money put in. Their usual response has been to resort to enclave projects, which nominally render accountability to home constituencies, but may not necessarily be compatible with the development priorities of the host government. The most common malady is the failure or inability of the government to support these enclaves with recurrent cost financing. A possible solution to this problem, as discussed in the previous chapter on aid management, is to encourage donors to increasingly channel their aid as sector support or even more flexibly as free standing budget support. Donors are wary of doing this because of concern that they will not be able to 'track' the results of their assistance. But if there are well defined and objectively measured indicators of outputs and outcomes, donors may feel more comfortable about providing sector and budget support, a move that would greatly enhance public expenditure management. The Comprehensive Development Framework provides a forum for the government to discuss with the donors a set of priority results and actions to achieve them that the donors could support, and the Poverty Reduction Strategy Papers which need to be presented to the Bank and Fund to access debt relief and concessional lending will also need to contain targets for key outcome and impact indicators.53 Attention to results indicators in public expenditure can provide a strong basis for measuring the 'track record' under the HIPC Initiative and for preparing and implementing the poverty reduction strategy. * Because of statutory and legislative requirements, budgets are usually so complex that even voters or their elected representatives may find it difficult to understand where public money is going and for what purpose. In a classic case of information asymmetry, this weakens accountability. Explicitly linking inputs to outputs and outcomes would also have the advantage of bringing public expenditure strategy within the grasp of the larger public and of giving civil society an effective 'voice' in improving the quality of both outcomes and the processes of budget management. It 53 CDF and PRSP are both in an evolving stage. Conceptually, CDF describes an approach to policy making based on greater balance between the economic, human and structural elements of development, and on new partnerships that support countries in achieving this balance. PRSP transforms the principles underlying the CDF into a plan of action for poor countries. 71 would also aid ownership and assumption of responsibility by the executive. 5.8 Obviously, these improvements will be achieved not by tacking results alone, but by making use of the information collected: in public debates; in the developments of strategies; in the design of polices and programs. Information that is not used to inform decisions is wasted. Indicators - Key Characteristics 5.9 In order to be useful to monitor the impact of public expenditures, indicators of outputs and outcomes should have some key characteristics: * The indicators should be relevant - be related to the results; to the extent possible the same key indicators should be used in the five-year plan, medium-term expenditure plans, annual budgets and sector development programs. * The indicators should be direct and unambiguous measure of progress and improvement. In other words, whether a change in an indicator is beneficial or not should be unambiguously obvious. * The indicators should be uniform over territorial levels so that partial results by woreda or region can be aggregated up to the national level. * The indicators should exhibit some variability in their levels across areas, groups, gender, overtime, etc. and be sensitive to changes in policies, programs, institutions etc. * The indicators should be robust, statistically reliable, not easily blown off course by unrelated developments, and not be amenable to manipulation to show achievement where none exists. * The indicators should show the combined effect of recurrent and capital spending as well as of different components of the sector expenditure where appropriate [for example, school achievement levels (graduation rates) are a result of the combined effect of school construction, teacher training and motivation, appropriate curricula and availability of materials]. * The indicators should be available fairly frequently and not be too costly to track. 5.10 While tracking outputs and outcomes is desirable, caution is warranted in linking expenditures to results in a mechanistic way. A comparison of expenditures and results achieved should not entail automatic lower allocations to programs or areas with weak performance; rather it can be regarded as a diagnostic tool helpful in identifying the reasons underlying weaknesses in program execution. In fact, the use of result indicators would allow a wider range of responses to fostering inter-regional equity, including modifying intra and inter-sector allocations; changing the balance between current and 72 capital allocations; modifying the planning process; giving targeted capacity building support, and so on. Thus, linking expenditures with results may reinforce the goal of pursuing regional equity objectives through the allocation of public expenditures. Experience Of Other Countries 5.11 A number of developing countries, particularly those eligible for benefits under the HIPC initiative, have started collecting result-based indicators with a view to feeding them back into the planning/budgeting process. The results so far have been modest, but the prospects are promising. Two cases are particularly relevant as a basis for the way forward in Ethiopia: Burkina Faso, for the mix of indicators used to orient aid disbursements, and Mozambique, for the link between planning and budgeting. 5.12 The experience of Burkina Faso is of interest for three reasons: (i) aid allocations and disbursements are linked to outcome indicators (not outputs measures) in HRD sectors as well as the poverty sensitivity and efficiency of public finance management; (ii) outcome indicators from administrative sources are backed by information from surveys on user satisfaction; and (iii) productive and social sector indicators are coupled with measures of efficiency/effectiveness of the budgetary process. Box 5.1: Burkinta Faso - P ilot Case. On Conditionaity Reformn:: In 1997, a two-year pilot was launched by the SPA group under th ordtion ofteE with e obei es of Improving and strengtheng govfo i Enhancing aid effectiveness through monitoring of ponce ind rs. * Finding ways of linking aid levels to performance. A few selected outcome indicators are being moitored (Wit c io to cha w in ds rater than to absou values) as possible reference for ai disburseme::nts: H Health: Attendance rates at heat centers; numbe of cesareans number of 2, prenatalconsui on tia vacination rates, user satisfaction and costof cae (survy).- * Education: Gross enrolment rates, first year- of pjimary school erolmentriate, Scssat at end of school, number of boo per pupil, user satisfation and cot::(surey) * Private sector development: public4private investnixt levels, nm er of. newfai ses averag before payment of suppliers, entepneu satisfaction (survey.) * Ntbfc finance management: level of expenditure in relaom sector o ives, uget uo rtes, sdare Of the budget reaching p,eripheral stures, it costs ilic delays i p s governmnt agencies. -The prncia itimconclusions of tis pilot case baemi forointdoomissin ae: Government ownersip has to start froin the pro desin sge; l nid publicity oe aid pram's rs enhances ownership;- a government/donor conractual retio p based oreults uld letitdet goverment in the choice of polices. * The focus on outcome indicators - wich measureresuts - helps io ep ioig the g efectveness of its policies, and making it acountable for te. It so hlps donor coordintin. * The shift to result-based lending must-lb both gradal (given he large chang tha it invol) and raduated: ony a vatiable, additional, pttion of aid should be linked to resuits, de to equity: on ios an the aim o avoiding abrupt disruption in aid flows,- * More work is required to identif appropriate indicators in other se ( to iove dta aity and to interpret outcomes with care. - Sonireto SPA WorikinsrCn'oun on F.cnnomic ManapemenKt Cond,t, tv Reformthle- T*in nPilot Case. Iiintr t .n-;e 197 73 Box5.2 MSozambique - The PEMidlThel Budgetw For manyq years, _te so-called Economic and Social Plan F(Po Econ mico e Social - e:PES iacop iedt budget prposal t ovehat Gmont snds to Parliament. At is conceived as-an annual slihe of the 5-year Government Program and specifies udiat is to be achieved withthe expeniture psed in the bidgeL Parliamientdiscusses.the P,ES$ beftrpovn to tte budget. The PES maikes use of standardisedindicators. Since the budget only has broad headings (like Healt, Education),Athe i:dicators presented in the PE$ actually specify the purpsef the enditue in greaadetail, sine, 0for in e T:~~~~~~~~~6d b 16e d6ikf th EiT:iiR:\ lschool enrolment figues are broken down ebyelevel of educaion, wheeas the budget does nEotprovide that breakdown. The PES is monitored systematically. Quty reportsie presented to th resonsile phlamewt ommit and discussed there, and a more dbtailed ahnnualTeport on-the implementaiton of the PESiss pepaed, discussed and The mergeri betweenVlthe foly separated Ministrl oPlanningr andaFinace at the opeo l levelhas m madethe Whole process possible. the futurethe PES indicators will also be used to monitor compiance of the budget with a fPoverty Action ]Plan under pEreSparaton.:000AA 000;; 0 0ASFL L ;= g0LX 40000P 0 5.13 The Mozambique experience is useful in the context of (i) allowing Parliament to monitor the results of budget execution in a way that is more appropriate for a political body; (ii) the integration of planning and budgeting; (iii) the merging of formerly separate institutions; (iv) the standardization of monitoring procedures; and (v) the method used to specify what allocated funds are meant for in areas that cannot possibly be captured through budget headers, like girls' education or the rural/urban split of some social services. Improving Current Practice In Ethiopia 5.14 While performance indicators are used in monitoring public expenditures in Ethiopia, their use could be widened and strengthened. There are several problems with current practice: • The planning and budgeting processes are quite separate. Planning appears to be correctly based on needs assessment and target setting; yearly or mid-term evaluations of the five-year plan take place with a variable degree of stakeholder participation (intensive evaluations were under-way in some regions during the PER Mission's field visits). However the link between the two processes seems weak, implying that the use of output/outcomes indicators that may take place in the planning process does not inform the annual budgeting exercise. * At the regional and zonal levels, budgets are prepared on an incremental (input) basis, where the priorities are: (i) covering 74 recurrent expenditure; (ii) ensuring capital budgeting for the ongoing projects; and (iii) allocating the remaining resources for any new investments. This is a lot better than the lack of discipline, common in many countries, where new programs are continually added without regard for the need to complete, operate and maintain existing programs. Nevertheless, the system could be improved if results achieved in the past systematically informed the budgeting process. Moreover, sector vertical budgeting is predominant, with sector budgets being aggregated to form the overall budget. In this context, there is a risk that inter-sectoral linkages and tradeoffs may not be sufficiently appreciated or factored into the budget process. Outcome and welfare impact indicators might be of help in assessing both intra and inter-sectoral linkages and allocations. 5.15 Several actions could be taken to improve current practice. It is important to coordinate the introduction of result indicators closely with the on-going Expenditure Management and Control Program, so that models and procedures that the reform project will translate into manuals and computer programs do allow for the complementation of financial figures with result-related indicators. The ongoing EMCP may have promising implications for linking inputs and outputs in public expenditure management. The Budget Reform Design Manual54 outlines in detail the format in which the budget should be presented to decision-makers; several items are important in this context: - Cost center budgeting will define a cost center as an administrative unit that is responsible for a set of activities with a similar objective. - The classification of recurrent and capital expenditure will be harmonized, which would make it possible to map recurrent and capital expenditure to an identified set of outputs/outcomes. This should make it possible to calculate unit costs (for example, the cost of providing one year of elementary education). - A unique budget calendar for capital and recurrent budgets will be introduced. The EMCP was conceived well before the emergence of the need for linking public expenditure to output/outcome indicators. In this regard, there is a clear need to revisit all the activities under the EMCP to ensure that they meet this requirement. This should be done at the earliest, and certainly before the manuals and the computer software are finalized. Trying to adapt them afterwards will result in avoidable waste of time and effort. 54Civil Service Reforn Budget Design Team: Budget Reform Design Manual: Chart of Accounts, Budget Preparation and Presentation for the Federal Govermment, Version 2.1, January 24, 2000. 75 * Beyond these reforms, indications of expected results should be provided together with the resource envelope in the following stages of the planning and budgeting process: - before fixing annual ceiling, so that needs and targets are factored into the process from an early stage; - when budgets are prepared within the limits of the fixed ceilings, where political choices are made; - when budgets are presented to elected bodies (House of People's Representatives, Regional Councils) and to donors; - when reporting on actual expenditure, preferably in a time series format. * The use of output and outcome indicators should be extended to the sectors where this is not yet the case. 5.16 Admittedly, the introduction of these mechanisms will necessarily be a gradual process. Possibly, this could start with sectors already using quantitative results and/or some of the regions with more capacity. Perhaps a useful beginning could be made by (i) taking this matter up during the next series of SDP reviews; and (ii) taking up user satisfaction surveys to get quick feedback on whether the services provided through public expenditures reach their intended beneficiaries and satisfy their needs. 5.17 In general, capacity constraints at the federal, regional, and sub-regional level, in terms of both insufficient skilled human resources and computer equipment, are significant. Thus, any effort to integrate planning, budgeting and result-based monitoring at regional level would require a detailed review and prioritization of capacity building actions (this is an area which the mission was not able to explore in detail). Next Steps In Strengthening The Focus On Results 5.18 Beyond improving current practice, monitoring of results could be strengthened in several ways so as to increase the effectiveness of public expenditure in promoting welfare and alleviating poverty. A first step in strengthening the focus on results would be to define through a participatory process the overall results that the government wants to achieve in terms of improving the welfare of the people and the indicators chosen to monitor these results." The five-year plan 1995/96-1999/00 went some way in this direction by indicating the overall goals of the development strategy proposed for the country: (i) 55 In Ethiopia, the term "welfare" is used to capture the multi-dimensional nature of wellbeing and is generally preferred to the term "poverty", so it is better to talk about the impact of public expenditures on welfare. 76 ensuring accelerated and sustainable economic growth; (ii) guaranteeing peace and stability; (iii) and promoting the democratization process. However, generally indicators were not selected to monitor progress towards these goals nor were targets set. The upcoming five-year plan, currently under preparation, presents an opportunity to move forward in this direction by specifying indicators and targets for the key goals. Note that this will also be necessary for the Poverty Reduction Strategy Paper summarizing the government's strategy which will need to be presented to the World Bank and the IMF to access debt relief and concessional assistance. The International Development Goals (see Box 5.3) represent a point of departure to begin discussing country-specific goals and indicators; this discussion should also be based on what welfare indicators are currently available in Ethiopia (see Table - 5.1 for a list based on information collected by the mission). Second, it may be worth examining whether the current WMS provides sufficient information and whether there are cost-effective ways of learning more about results on the ground. While some indicators of welfare are available on an annual basis and respond quickly to policy/program changes -for example, the literacy level of school-age children or malnutrition -and can thus be used to provide timely feedback, others are not available - for example, indicators of learning achievement. Other indicators still - most notably income poverty and income inequality - are available only infrequently, and usually with a time lag, because of their slow changing nature and of the time and resources involved in measuring them. Thus the existing WMS could be improved in several ways. First, user satisfaction surveys could be taken up on the effectiveness of public expenditure and the quality of delivery systems. Second, it may be worth exploring the use of "proxy" or "lead" indicators to track changes in income poverty and inequality (see Box 5.4). Third, it may be worth thinking about innovative indicators such as those being developed in Uganda (see Box 5.5). * A third step would consist of studying in some depth the linkages between public expenditures, outputs and the outcomes/impact indicators selected to track results. Government activities, most of which are reflected in the budget, have a major impact on welfare. But many other factors also have an impact on welfare: uncontrollable factors such as the weather; exogenous factors beyond a country's control such as the international prices of commodities; factors largely under the control of actors other than the government such as foreign investment or the amount of foreign aid flowing in; and factors which change only slowly over time, such as culture, norms, beliefs. Thus, assessing these linkages is not easy: various methodological problems arise in mapping inputs into outcomes and impacts. First, there are inter-sectoral and inter-temporal spillovers: interventions in one sector 77 will have an impact in other sectors, and the impact in the original and/or other sectors may appear after a gap of time, sometimes even a generation. Second, as mentioned earlier, public expenditures are only one of several variables influencing outcomes and impacts. Nevertheless, some techniques exist to isolate the impact of public expenditures on welfare, and there are data in Ethiopia to begin studying the linkages. For example, good program evaluations could be conducted; regressions could be used to estimate the impact of public expenditure on selected indicators using time series or cross- sectional data; and computable general e uilibrium models and social accounting matrices could be developed.5 Finally, it would be possible to use existing and new information as well as the results of these studies to assess whether the distribution of expenditures not just within one sector but across sectors is the most effective given the results the government wants to accomplish. Table 5.1: Some Welfare Indicators Available in Ethiopia Dimensions of Welfare Indicators Source Frequency Welfare Economic Income or expenditure per capita HICES 1995/96, opportunities Income poverty incidence, depth and severity (CSA) 1999/00 Income inequality measures Capabilities Literacy, Basic numeracy WMS 1996, 1997 (CSA); 1996/97/98/99 ESDP Malnutrition measures (weight-for-afe, height-for- WMS 1996, 1997, age, weight-for height for children under five) (CSA) 1998, 1999 Infant, child, under-five mortality rates HSDP 1998/99 and annual Matemal mortality rate HSDP 56 These options were not fully explored during the PER mission. 78 oz&3: .dThe Intk g a ti-D iD 1 men, Goats ec g nd indiatos h a, a d t -t 1, r variou UF conlliferencemXX s:-i thReducing the leincdec of extem poverty byf haltt .. .. : . ' .. ...' + i ;} * Reducig infa t, hld aternfal mortal ;cty by two-thirds * Ac hivn nivesa en rollen ino pdiar edoatiox * liintig ene disaity a' eduatio (by 2005). 0 * Providi u s a es ctv healte lack of inom bu loilieay,poealth, ack o aces to.¢ sevie a - eniomn: contant,a < ~~~~~~~~~-n t ins. +° dicatorswere lectedidi .s ra. . '.*nd Ptiverqy gap ratio. _y We- ~~~-tem fWheWM-! i WU:- : Ineqult: pors fifWs sh*are lof aionlw consumto;5 u:n ior smne:e * (141 mlnuriton prvalnc ofundeweighn ncide ne ...... . Bo 5.4: Pry And LoVd Idic rs " What can beoe ltoas tng n l dts c iMncoe povery,hi ch cannot be p c e an c mt l t forv xa eriusille wages Mother p y ioidey s ite poor sehds con ewhen t e ustd enouh icom an tht wul.be he irs thy 'oul 0estop: cosuin wenthirinoumefll(od .~~~~~~~~~~o e.: :: :- ::t <: , .,, .: ,,e ::, + e+::U+ d.,.9 with 4h e ei lastict oWt lincm levlsSch indics calle led i-end*:ictorls" becus they ivea ring signal hofachaneeryon Aobndotio ist idets indic$ors X