Report No. 41840-SD Sudan Public Expenditure Review Synthesis Report December 2007 Poverty Reduction and Economic Management Unit Africa Region Document of the World Bank CONTENTS ABBREVIATIONS AND ACRONYMS .................................................................................. i ACKNOWLEDGEMENTS....................................................................................................... EXECUTIVE SUMMARY..................................................................................................... ...i 111 1. 1 A. Background................................................................................................................... 1 B. Complex And Uncertain PoliticalEconomy Text ..................................................... Macroeconomic Stability but BuildingPressures ...................................................... 2 C 4 An unsustainablefiscal expansion............................................................................. Daunting Development Challenges........................................................................... D.. 6 2. 9 A. Managing and MobilizingNational Revenues............................................................. 9 B. Non-Oil Revenue Ranagement................................................................................ Oil Revenue Management.......................................................................................... 13 C. ChallengesArising from Decentralization.,............................................................. 14 3. Key Issues inGNUExpenditure Management........................................................... 19 A. 20 B. Trends inGNU SpendingComposition................................................................... 22 23 Budget Process and Credibility................................................................................ Federaldevelopment expenditures .......................................................................... FederalCurrent Expenditures.................................................................................. D. C. 32 E. Financial Management............................................................................................. 33 4. Toward a Pro-Poor Public Budget .............................................................................. 37 A. A Proxy for Poverty Reducing Expenditures........................................................... 37 B C.. Applying the Definition........................................................................................... 38 Increasingthe Level and Quality of Pro-Poor Spending......................................... 41 5. 45 A. Intergovernmental Transfers and Fiscal Decentralization.......................................... Motivation................................................................................................................ 45 B. 46 C. Current State of the Intergovernmental Transfers inthe North............................... Going Forward-Challenges and Policy Options.................................................... 51 6. 59 A. LimitedProgress on Fiscal Decentralization ........................................................... Budgets and Fiscal ManagementinNorthern States .................................................. 59 B. 62 C. Weak Fiscal ManagementCapacity......................................................................... Sub-national Pro-poor Expenditures Rising ............................................................ 64 D. Key Next Steps ........................................................................................................ 65 7. A. Expenditure and RevenuePolicy inSouthern Sudan: Progress and Challenges........67 BuildingBudget Institutions for Good Governance................................................ 67 68 C. B. An Analysis of 2007 GOSS Budget......................................................................... Performanceofthe 2006 Budget ............................................................................. 74 D. Key challenges......................................................................................................... 75 ANNEXES.............................................................................................................................. 77 Annex A: Fiscal Decentralization Responsibilities According to the InterimNational 78 Annex B:GNUBudget Tables ......................................................................................... Constitution....................................................................................................................... Annex C: Northern States Expenditures........................................................................... 82 89 ListofTables 3 Table 1.2: Large Regional Disparities inKey MDG Indicators ............................................................. Table 1.1: Trends inKey Macro and Fiscal Indicators. 2002-2006 ...................................................... 5 Table 2.2: Relatively weak tax efforts: Sudan and selectedcomparators 1/........................................ Table 2.1: Growth inResourcesfrom Oil Revenues ............................................................................. 9 13 Table 3.2: Sector shares ofnational development projects (percent) ................................................... Table 3.1: Defense Spending Rises in2006 ......................................................................................... 23 24 Table 3.4: Excessive Movements inthe Portfolio ofNational Development Projects......................... Table 3.3: Budgeted and Actual Expenditures on National Development Projects ............................. 24 25 29 Table 3.6: Federal and Northern States Education Spending, 2000-2005 ........................................... Table 3.5: Federal and Northern States Health Spending, 2000-2005.................................................. Table 3.7: Main Steps and Timetable for Budget Preparation.............................................................. 31 32 Table 5.1: State Revenue Sources inNorthern Sudan.......................................................................... Table 3.8: GNU Budget Execution(actual expendituresas percent of budget) ................................... 33 47 Table 5.2: FiscalTransfers to Northern States (2006/2007)................................................................. Table 6.1: Per Capita Pro-Poor Spending............................................................................................ 48 -63 69 Table 7.2: Expenditures Relative to Plans, 2006 USDmillion and percent ......................................... Table 7.1: GOSS Overall Balance During 2005-2006, millionUSD................................................... Table 7.3: Execution of Wages, Operations and Capital, percent of plan, 2006 ................................. 70 72 Annex Table B.1: GNU Revenue and Expenditure: Economic Classification. with older chapter structure for compatibility across time (inbillion SD. 2000 -2006) ................................................... 82 Annex Table B.2: GNURevenue and Expenditure: FirstQuarter 2007 Performance. inEconomic Classification (SDDbillions) ................................................................................................................ 84 2006) ..................................................................................................................................................... Annex Table B.3: GNU Expenditure: Preliminary FunctionalClassification(SDDbillions, 2000 - Annex Table B.4: GNU Expenditures: 2007 Budget inFunctionalClassification(SDD billions) ....88 86 ListofFigures Figure 1.1:Agriculture. Mining and Services dominate the economy (2006 GDP by sector. at factor cost)......................................................................................................................................................... Figure 1.3: Recovery since sharp fiscal adjustment inthe 1990s.......................................................... Figure 1.2: SudanI s a Country Outlier on Education (Net Primary Enrolment) Relative to Income ....35 6 Figure 2.1: Oil RevenuePerformanceDeteriorates............................................................................. Figure 3.1: Rapid increase inGNU.expenditure................................................................................... 10 20 Figure 3.2: Growing obligations to sub-national governments, inpercent of total GNU expenditures21 Figure 3.4: FunctionalComposition of GNU Spending(percent shares)............................................. Figure 3.3: Federal expendituresby economic classification, 2000 -2006 ......................................... 22 Figure 5.1: Intergovernmental Relations inSudan ............................................................................... 27 46 Figure 5.2: Transfer to the Northern States (% GDP, 2000-2006, and budgeted 2007) ....................... Figure 5.3: State Transfer Levels (Per Capita, 2006) ........................................................................... 49 50 Figure 5.4: State Transfer Realization Ratios (2005/2006) .................................................................. 51 Figure 5.6: FFAMC ProposedFederal Transfers ................................................................................. Figure 5.5: Simulations ofAlternative Transfer Formulas to States (2007) ......................................... 54 55 60 Figure 6.2: Share of Federal Transfers That Are Earmarked................................................................ Figure 6.1: Share of Federal Transfers in State Revenues.................................................................... Figure 6.3: State Budget Credibility - PlannedVs.Actual Expenditures............................................ 60 Figure 6.4: Planned and Actual Development Spending...................................................................... 61 Figure 7.1: Deviations o f Actual Pro-poor Spendingto Plans, 2006, inpercent.................................. 63 73 ListofBoxes Box 2.1: Medium-Term Gains from Tax Reform................................................................................ Box 1.1: A Vision for Equitable Use ofNational Wealth inthe 2005 InterimNational Constitution ..2 Box 3.1: Agriculture Subsidies and Support under the Green Mobilization Program......................... 16 Box 3.2: Insights on Sub-national Health Spending............................................................................ 29 26 ACKNOWLEDGEMENTS This Public Expenditure Review (PER) report is the result o f collaboration between the World Bank and the Government o f Sudan, with support from development partners including financial assistance from the British, Dutch, and Germangovernments. The counterpart team inthe Government o fNationalUnity (GNU) was ledby the Ministry o f Finance and National Economy (MOFNE), under the guidance o f H.E. Minister Elzubeir Ahmed Elhassan, State Minster Dr. Ahmed El Majzoub, State Minster Dr. Lual Deng, and Undersecretary Sheik El Mak. They provided enormous support and encouragement to the PER team throughout concept and implementation stages o f the report. Other members o f MOFNE senior management that were particularly helpful include Director General, International Cooperation Omer Abdel Salam, and Director General, Development Directorate Hafith Atta Elmanan. The PER team worked in close collaboration with a dedicated Steering Committee established by the MOFNE, which provided excellent support to the overall work and many hours o f discussions during the exercise. The PER team would like to express its sincere gratitude and appreciation for the Steering Committee's cooperation and contributions throughout the process. We are particularly grateful to the Chair o f the Committee, Mustafa Layla Omer Bashir (MOFNE - Director General, Macro Policies and Programs Directorate). Yousif Houli (MOFNE - Director General, Budget Directorate), and to the Deputy Chair The team is also very grateful for the inputs o f the other members o f the Committee: Muna ElsaidAbu Haraz (MOFNE Director General, Revenues), Rabaa Ahmed Elkhalifa (BOS - - DG, Research & Policies Directorate), Ikhlas Foad Eltoam (MOFNE - Director General, Expenditures), Ibtisam Hassan Jadaah (MOFNE - Development Directorate), Abdallah Ibrahim (MOFNE - Head, Budget Performance Unit), Asim Hamza Hussain (MOFNE - Development Directorate), Nadia Shukr Alla (MOFNE - National Accounts), Hussein Huliman Koya (MOFNE (MOFNE International Cooperation Directorate), Wisal Awad Mohamed Khair (MOFNE - - Head, States' Affairs Unit), Yousif Mohamed b e d Bashir - International Cooperation Directorate), Ala Eldin Elmajzoob (MOFNE - Expenditures), Jamal Awad (FFAMC Representative), and Badr Eldin Osman (MOFNE -Macro Policies and Programs Directorate). Special thank also goes to the PER Steering Committee Secretariat memberswho provided substantial support and smoothed the PER'S logistics and meeting arrangements. We are particularly grateful to Musa Ali Mohammed (MOFNE - Budget Performance Unit), Shaza Abd Allah (MOFNE - Budget Performance Unit),Khalda Abd Allah (MOFNE - BudgetPerformance Unit), Osman Mohammed (MOFNE - Revenue), andHussainOmer Awad Allah (MOFNE - States' Affairs Unit).Lastly, the team would also like to acknowledge the role of the Steering Committee's former Chairperson Dr. MohammadEl Hassan Makkawi, Advisor to the Minister, MOFNE, who helped guide early stages ofthe exercise. i Onthe World Bank side, the PER was managedby Bill Battaile (Senior Economist, AFTP2). Team contributors included Jeni Klugman (Lead Economist, AFTP2), Mosllem Alamir (Economist, AFTP2), Michael Levitsky (Lead Energy Economist, COCPO), Vivek Srivastava (Senior Public Sector Specialist, AFTPR), Kai Kaiser (Senior Economist, PRMPS), RahiAbdula (Consultant, AFTP2), Jack Van Holst Pellekaan (Consultant, AFTP2) and Patrick Mullen (Health Specialist, AFTH3). The chapter on fiscal issues in Southern Sudan was led by Allister Moon (Lead Economist, AFTP2), with a team including Sanjeev Ahluwalia (Senior Public Sector Specialist, AFTPR), Michael Levitsky (Lead Energy Economist, COCPO), Parminder Brar (Lead Financial Management Specialist, AFTFM) Kenneth Schwartz (USAID), Damoni Kitabire (AFDB), Bob Searle (Consultant, AFTP2), Andrew Hall (DFID), and Rahi Abdula (Consultant AFTP2) and extensive support from Hari Nayer (Consultant, AFTPR). Excellent contributions by the IMF on the revenue management chapter are also gratefully acknowledged, particularly the inputs from Todd Schneider (Lead Economist, IMF). Dr. Yagoub Ali Gangi (Consultant, AFTP2) facilitated much of the day-to-day work of the PER by working directly in MOFNE and coordinating closely with the Undersecretary MOFNE and the chair and members of the Steering Committee. The PER was undertaken with the overall guidance of Kathie Krumm (Sector Manager, AFTP2). Ishac Diwan (previous Country Director, AFCE3) reviewed an earlier draft, provided comments, and supported the whole process. Ken Ohashi (current Country Director, AFCE3) supportedthe final roundof revisions. The state case studies provided invaluable insight into fiscal decentralization issues. Special thanks are given to the consultant teams that carried out this work, including Professor Madani M Ahmed, Dr. Ahmed A Badawi, Yousif M A Bashir, Abdalatif Taha Ali, and Tarig A Ismaeil. The work would not have been possible without the excellent cooperation of the state Ministries of Finance in Red Sea, Gazira, North Kordofan, Khartoum, South Kordofan, North Darfur, South Darfur, and West Darfur states. The peer reviewers were Adrian Fozzard (Country Manager, LCCHN) and Bill Doritinsky (PRMPS, concept stage). Many thanks to their valuable comments and suggestions, as well as from various other Bank staff includingAsif Faiz (Country Manager Sudan, AFMSD) and Shamima Khan (Manager, Juba Office). Reactions and feedback from multiple sessions of the Donor Economic Group in Khartoum were useful, as well as input from various development partners, particularly UNDP (Dirk Hansohm) and DFID(Andrew Hall). Dora Harris, Marjorie Kingston and Arlette Sourou provided excellent support in finalizing the report. Support of the Sudan field offices i s also gratefully acknowledged. InKhartoum, particular thanks goes to Tarig Osman (RM), Hassan Gaafar (IT), Yousra Abdelhaman (Executive Assistant), Azza Imam (Team Assistant), Abir Abdelrhamn (AdmidFin. Assistant), and Frdos Akasha (previous Team Assistant). In Juba, special thanks goes to Milicent Ndolo (Team Assistant) and Juliette Guantai (Team Assistant). .. 11 EXECUTIVESUMMARY A. Introduction 1. The signing of the ComprehensivePeace Agreement (CPA) on January 9,2005 marked the end of nearly four decades of civil war in Sudan and an optimistic beginningof reintegration into the internationalcommunity. This peace, together with macroeconomic stability and significant natural resources, has offered a tremendous opportunity to increase broad-based economic growth and poverty reduction in a united Sudan. However, more than two years after the signing o fthe CPA, the political and security context remains fragile and complex, even for a post-conflict situation, with lingering conflicts in several parts o f the country, delays in implementing some important provisions related to wealth sharing inthe peace agreement, and a pending referendum on national unity scheduled for 2011. The record on humandevelopment outcomes also remains weak (Sudan ranks 141 out o f 177 countries in UNDP's 2006 Human Development Index), especially relative to the fact that income per capita has passed $1,000 or roughly 25 percent higher than the SSA average. 2. This Public Expenditure Review (PER) is the first for Sudan, and reflects the initial phase o f engagement with the newly formed Government o f National Unity (GNU) and Government o f Southern Sudan (GOSS) on fiscal management issues. The objective i s to establish a baseline understanding o f key fiscal management and policy challenges, highlight priority reform areas for policymakers, and set the agenda for the next phase o f work. A fiscal dialogue has been rebuilt at the federal and sub-national levels, after more than a decade hiatus o f World Bank engagement in Sudan. During this re-engagement, process has been as important as deliverables, with close and frequent discussions with the MOFNE Undersecretary and senior officials via a comprehensive PER Steering Committee, as well as a substantial resident presence inJuba. The scope o f initial PER issues was agreed jointly with the GNU, and this report is essentially a snapshot o f the on-going dialogue and draws largely on five background papers covering: revenuemanagement, pro-poor spending, intergovernmental transfers, fiscal management and pro-poor spending at the state and locality levels, and fiscal management in Southern Sudan. The PER exercise has been an interactive process o f focused and timely deliverables and dissemination through in-country workshops, and has served as an input into the 2007 and 2008 budget processes, the preparation o f pro-poor monitoring for the 2006 and 2007 Sudan Consortia, and fiscal management at the state and locality level. 3. The political and legal context reflected in the 2005 Interim National Constitution (INC) has key implications for expenditure management. First and foremost i s fulfillment o f the wealth sharing obligations o f the peace agreements, which limits the discretion o f GNU spending as significant shares o f oil revenues must be transferred to the GOSS (roughly 15 percent o f total GNU expenditures in 2005-2006) and other oil producing states. Second, the vision of fiscal decentralization is a key aspect of a unifiedandpeaceful Sudan, potentially addressing inequalities and the root causes o f conflict ... 111 by increasing financing for marginalized and underdeveloped areas. This also has significant implications for discretionary spending o f the federal government, as a greater share o f revenues is transferred to the Northern states. And third, accountability at all levels o f government i s required to ensure resources are shared equitably and used inaccordance with each level o f government's constitutional responsibilities. 4. The agenda for addressing challenges in public expenditure management is huge, and a programmatic approach has been adopted for the PER program that is pragmatic and with reasonable expectations o f what can be accomplished inthis post-conflict environment. Accordingly, this first PER underscores the need to address basic elements o f effective public expenditure management including comprehensive and transparent functional budget reporting, monitoring o f actual spending and its outcomes to bolster accountability, costing sector and cross-cutting policies to make budget prioritization more transparent, improved budget credibility, linking development planning to the annual budget preparation process, and generally increasing the use o f expenditure analysis in decision- making. Many o f these issues are recognized as priorities by the authorities and reforms are either planned or underway. There are other key issues that this PER does not tackle in depth, including defense allocations and improving the transparency and accountability o f oil sector governance. With the solid foundation o f this first PER and on-going dialogue with the Sudanese authorities, these issues will hopefully be addressed as the PER program moves into its next phase. B. Understanding Public Finance inSudan 5. Fiscal management since the CPA has been dominated by the attempt to balance pressures for continued increases in federal expenditures and the need to finance new obligations to sub-national governments, on one side, with oil revenue volatility and shortfalls inexpected revenue growth. 6. While the discovery and exploitation of oil resources has facilitated an increase in nationalwealth, it has also brought a myriad of problems. Sudan is one of the newest significant oil producing countries in the world. Significant oil production started in 1999, and the country i s now the third largest oil producer in Sub-Saharan Africa, behind Nigeria and Angola. Real wealth has subsequently increased, with economic growth currently over 10 percent and among the highest on the continent. Oil revenues have expanded budget resources potentially available to support pro-poor development. The political economy of oil has also brought with it additional pressures for expenditure (some o f which may be conditioned by capacity and other constraints), rent-seeking behavior, and reduced pressure to undertake key fiscal reforms that would help ensure medium- and long-term stability irrespective o f the course o f oil production and prices. Oil revenue in Sudan has also proven volatile. Oil production has fallen short o f expectations inrecent years. Moreover, Sudanhas thus far not benefitedfully from the recent rise in global oil prices, since much of its crude production i s now o f a lower quality and had been fetching unexpectedly low prices. Taken in tandem with a relatively low tax effort, these factors have caused severe difficulties for budgetforecasting, formulation and implementation. iv 7. Public expenditureshave grown sharply-from a low base o f 7 percent of GDP in 1998 following the fiscal stabilization program to 22 percent in 2006-initially using the fiscal space from new oil revenues and more recently outpacing the growth in revenues with rapid increases in most categories o f spending, including for some CPA-related obligations. A new feature is the fiscal deficits that have emerged (roughly 2 percent o f GDP in2005 and 4 percent of GDP in 2006), after nearly a decade o f prudence. CPA-related spending pressures are partly responsible. The dominant effect o f the CPA on the composition o f GNU expenditures has been large increases in new obligations to sub-national governments-both CPA obligations to the GOSS and fiscal decentralization transfers to the Northern states-with the federal share dropping from 92 percent of total GNU spending in 2000 to 64 percent in2006. Responsibility for basic service delivery has been devolved, and increases in sub-national resources implies higher levels o f pro-poor spending (as discussed below) but has also underlined the importance o f addressing deficiencies in effective expenditure management at lower levels o f government. The effectiveness o f decentralization and resource use at the sub-national levels i s subject to improvements in public financial management. 8. Fiscal expansion at the current pace appears to be unsustainable. The 2007 budget called for a 13 percent increase intotal expenditure over last year's budget to a total o f $11.8 billion. Large increases are planned for both the federal wage bill and transfers to Northern states (roughly 25 percent nominal increases for each). National development was also budgetedto increase significantly. The likelihood o f meetingthese expenditure plans i s low given oil revenue shortfalls in the first half o f 2007, at least not without high levels o f domestic or external deficit financing which would be problematic, as noted by the IMF. The pressure for expenditure cuts and reallocation is high, raising the relevance o f this PER and the need for subsequent government action to address the current trajectory o f fiscal imbalances and avoid weakeningthe country's hard-won macroeconomic credibility. 9. The fiscal pressures in the South have proven even more acute. The overall balance in2006 moved sharply into cash deficit due to limited spending discipline as well as oil revenue shortfalls in the second half o f the year. Aggregate spending was driven by outlays on wages and operations that were roughly double planned amounts, while capital expenditures were cut sharply. In the first half o f 2007, oil revenue shortfalls continued, GOSS significantly over-estimated non-oil revenue, and in the face o f huge development needs, expenditure plans in the 2007 budget were significantly expanded relative to 2006. The cash reserves that could have provided cushions were largely run-downlast year and are no longer available. One risk is that plannedinvestments inroads, schools, clinics and so on will be squeezedby the burgeoningpayroll. 10. At the national level, budget credibility is hampered by limited costing and prioritizationof sector and thematic policiesduring the budgetpreparationprocess, as well as execution volatility. The macroeconomic framework, started early in the budget preparation cycle, focuses on deriving macro targets, but deriving sector envelopes through consideration o f costed policies and a transparent prioritization process remains a challenge. Introduction o f a functional budget classificationpromises to improve budget coordination o f sectoral policies. Aggregate annual expenditure deviations from approved budget have improved since 2000, though 2006 experienced a low 87 percent execution rate due to oil V revenue shortfalls. Monthly variation o f MOFNE releases i s a significant issue. Without a predictable flow o f resources to spendingunits, execution o f spending plans i s hampered, and o fparticular concern are the effects on financing o f capital expenditure. 11. Developmentfunding is a key component for CPA implementationand poverty reduction, but has been volatile and heavily concentrated in a handful of large endeavors. The largest five national projects absorbed more than 60 percent o f total GNU investments in 2006, causing weak financing performance in the remaining national portfolio. Spending above budgeted allocations i s common for the large projects. New projects make up more than a quarter o f the recent annual portfolios. State-level development in poor states and rural areas is generally underfunded. There are serious deficiencies in development planning, at the program planning and project analysis levels, and at both the national and state levels. Linkages across levels o f government, especially with respect to supporting service delivery on the ground, are weak. Evidence of project codbenefit analysis i s limitedfor the majority o f investments inthe annual portfolio. Deficit financing via government promissory notes i s also particularly opaque. 12. Notwithstanding on-going efforts by the authorities, public financial management systems in Sudan are generallyweak. Recent GNUmeasures, supported by the IMF, should be recognized. These includerestructuringthe budget to reflect sub-national transfers, computerizing the budget performance reporting system, training on functional budget classification, and establishment o f a high level committee to help ensure allocation and monitoring o f federal government resources in line with budget priorities. Nonetheless, significant deficiencies remain. The commitment control system is not effective, and expenditures are committed on the basis o f appropriations rather than availability o f cash resources. Cash management i s weak, and for other than salary payments, there i s a large amount o f unpredictability regarding resource flows at the federal, state and locality level. Due to this, there is a regular build up of arrears although information regarding commitments or arrears i s difficult to collect. Also difficult to define but widely acknowledged i s extra-budgetary revenue collection and expenditure, including at the federal line ministry level. The Chamber o f Accounts in MOFNE compiles monthly financial reports for the government as a whole, and monthly government finance data i s produced on time, with a six week lag. However, the utility of this data is undermined due to shortcomings inthe classification system and comprehensiveness. 13. In the South, the GOSS is establishinga system of public financialmanagement virtually from scratch. The former SPLM Secretariat of Finance, which managed resources o f around $100,000, has transformed itself into a Ministryresponsible for managing over one and a halfbillion dollars annually, including significant external financing. The government has committed itself to establishing sound andtransparent financial management systems and to combat corruption. While some progress has been made, including the outsourcing o f accounting and auditing functions and actions to address corruption, government finances are still characterized by weak management and lack o f accountability. vi C. Toward a SustainablePro-Poor Expenditure Framework 14. With national income approachinglower-middle income status, there is a strong imperative to ensure adequate resources are channeled to address the weak record on human development outcomes and wide regional disparities, and relatedly, the underlying structural causes of conflict. Promoting a shift to a more pro-poor budget has been a key focus of the PER exercise, and there is increased attention to defining pro-poor allocations in MOFNE planning/execution analysis. But there are serious limitations to measuring and monitoring pro-poor spending in Sudan, due to the quality o f data on spending and outcomes and the complexities added by fiscal decentralization. Moreover, Sudandoes not yet have a full PRSP, so it is not possible to point to a coherent set ofpolicies that would anchor a definition o f poverty-reducing allocations. 15. Working within these limitations but also spurred by the rapid rise in public expenditures, the World Bank has been working with the GNU to define and monitor the share of public expenditures targeted to the poor. According to the PER working definition (see Section 4), poverty-reducing expenditures have increased from a low base o f $16 per capita in the 2000-2004 pre-CPA period to $68 per capita in 2006. But the amount realized-5.5 percent o f GDP in2006-was below budget plans (6.6 percent of GDP), J A M commitments (5.9 percent o f GDP), and the average for HIPC countries (7.0 percent o f GDP). 16. The available information suggests a welcome increase in pro-poor allocations. Much higher levels o f MOFNE releases are designated for activities expected to benefit the poor than before the CPA. However, we don't have information directly from the spending units on the use o f funds for the intended purposes and the quality and efficiency o f spending. This i s especially true at the state and locality levels. The growth in pro-poor spending has occurred in the context of rapidly expanding total expenditures, and did not involve reallocation in level terms from other spending categories given all areas experienced real growth. It will be important to protect and strengthen pro-poor spending as a policy reflected inthe budget, given the current fiscal environment. The formulation o f a national poverty reduction strategy should help inthis regard. 17. The increasing pro-poor efforts over the last two years were driven by higher transfers to Northernstates, which accounted for more than half o f total pro-poor spending in2006. The JAMcalled for a large majority of GNUpro-poor budget efforts inthe form of transfers to the Northern states, to support decentralized delivery o f basic services. The increase has been more federal than expected, with resources transferred to the states heavily skewed toward current expenditure over development, raising concerns given improving service delivery to the poor in the longer term requires development investment in schools, clinics, roads, etc. As more resources have come to the states, it i s critical to improve the transfer system to do so with equity, transparency and predictability. Northern states rely heavily on transfers and problems with predictability are a key factor behind poor budget credibility. 18. Progressinincreasing financial support to the Three Areas has been slower than expected. The Three Areas-South Kordofan state, Blue Nile state and Abyei-have been vii accorded special status inthe CPA, inrecognition o f their war affected status and challenges. As documented inthe JAM, the bulk of the needs are inbasic services delivery (education, health, water and sanitation), rural development, and new construction-in particular, of roads and a connection to the northern power grid. Financial support to the Three Areas in 2005-2006 was around 35-40 percent below what had been programmed inthe JAM. Inthe 2007 budget, the GNU plans an increase to $249 million, compared to $183 million in 2006 (which i s still 32 percent below the amount envisioned underthe JAM). 19. The PER exercise identifies three areas for increasing the level and quality of pro-poor spending: improving planning and the analytical underpinning of budget preparation, enhancing budget credibility, and reallocating from non-pro-poor spending. Specific actions include addressing data constraints mentioned earlier (including a sorely-needed updated poverty analysis and household budget survey), requiring explicit poverty-motivation of budget plans at the sector level and within a medium term timeframe, eliminating the gaps in budget execution rates between pro-poor and non-pro-poor expenditures, much better data at the state and locality level on the use o f funds to inform decisions and allocations, and dedicating resources to development expenditures at the state level where it will directly benefitthe poor and marginalized areas. D.ImprovingAccountabilityandEfficiencyinPublic Spending 20. Accountability and effective public expenditure planning and management in the North are hampered by limited information and monitoring. Inthis area, there are four significant deficiencies: (i) current budget system does not provide a classificationo f the expenditures according to their function and purpose; (ii) country lacks a consolidated the budget covering the various levels of government; (iii) GNU expenditure data currently only captures MOFNE releases to spending units, rather than actual use by spending units; and (iv) there is limited attention budget execution and less to the outcomes o f expenditures. Overcoming these challenges and achieving the basic information requirements should be at the forefront o f expenditure management reforms. 21. Addressing budget planningdeficiencies is a longterm and multi-faceted effort, with many aspects dependent on immediate efforts to filling minimal information requirements. The PER process has encouraged on-going reforms for adoption of the GFS budget classification and a consolidated GNU budget including coverage o f the spending in the Northern states, but progress remains slow. A concerted effort by MOFNE and state ministries o f finance is needed to accelerate progress, with buy-in at top levels o f these government units and belief that the cost o f adjusting systems and practices will deliver significant efficiency gains in expenditure management. The PER attempted to map the MOFNE releases by spending unit, combinedwith new information from the Northern states, to produce a consolidated functional dataset. While caveats regarding the source information remain, analysis in this PER o f sector funding and trends demonstrates potential benefits o f pursuingthese reforms inexpenditure management. 22. From a more strategic standpoint, efforts to improve upstreambudgetplanning activities,particularly the macroeconomic framework, may provide the best short term ... Vlll gains. The following areas are starting points for improvements: (i) linking sector better planning to current and capital expenditures; (ii)ensuring that allocations reflect policy anchors, such as the on-going Five Year National Strategic Plan; (iii) concrete estimates o f the fiscal costs of policy actions are required; and (iv) predictability in execution is a necessary incentive for investment inthe planningprocess. 23. Other key areas that can deliver expenditure efficiency gains in the near term are improvementsin development spending and core fiduciary dimensions. Addressing development planning and implementation deficiencies will require a sector focus and significant capacity building at the federal and sub-national levels, with some efforts currently underway through World Bank Institute training workshops. Key aspects of financial management are being tackled by the on-going Country Integrated Fiduciary Assessment (CIFA). This assessment, in reference to international good practice (PEFA indicators) and appropriately accounting for the decentralized environment, will identify fiduciary strengths and risks and provide the basis for a policy agenda. Recent work by the IMF's Fiscal Affairs Departmentshould also informthe agenda. 24. Specific to the conditions in the South, budget and financial management concerns are acute, and the MOFEP has made a series of commitments in order to ensure sound public resource management and to guard against corruption and diversion. Some of these have been fulfilled, but there is a significant way to go, and the following steps remain priorities: Adopt and implement the public finance bill that i s currently under consideration by the Southern SudanLegislativeAssembly (SSLA); Put in place a transparent procurement system, by applying interim procurement regulations, enacting the new procurement law and proper use o f the recently appointed Procurement Agent as well as urgently appointing an External Audit Agent; Work to effectively operationalize the Treasury Single Account where all revenues and expenditures are transparently consolidated and accounted for; Appropriately utilize ex ante controls, including the recently established Cash Management Committee; Build on the SSLA debate around the 2007 budget, and establish appropriate systems and structures for regular reviews o f public expenditure and pro-poor spending by Cabinet, the Assembly and the public, with minimal lags inreporting; and Establish mechanisms for executive accountability, as per the Southern Sudan Interim Constitution, including the Southern Sudan Audit Chamber and the Anti-Corruption Commission. ix E. A SevenPointAgenda for Action 25. Key recommendations for the near term, especially inthe context o f the 2008 budget formulation process, respond to the diagnostic work in this PER. Actions inthese areas will address key deficiencies in basic building blocks o f expenditure management, and many are mutually reinforcing. These recommendations are consistent with the CPA and the JAM, but remain relevant as progress has been slow (e.g., intergovernmental transfer reform, budget classification). Some have been elaborated upon and deepened through further analysis and field work o f the PER. This is not a complete list o f needed reforms, given important areas were not covered in this report, and additional reform areas may be tackled in the second phaseo fthe PER as other priorities emerge. 1. Improve oil revenue volatility management, as a basis for budget credibility and better expenditure management. This issue i s particularly relevant for the GOSS given its nearly complete dependence on oil revenues. Oil sector transparency and riskmanagement are immediatepriorities, includingwith respect to activities of state and quasi-state companies. The current oil savings fund has failed to provide a sustained buffer from the inevitable volatility faced by an oil economy such as Sudan. Reserve accumulation and credible management o f the oil savings account through a transparent governance structure i s needed, along with accelerated progress on non-oil revenue reforms. 2. Restore fiscal prudence at all levels o f government by controlling expenditures, in line with abovementioned revenue measures, within a medium term outlook that focuses on the non-oil deficit as a key fiscal indicator. The 2008 budget preparation should promote leveraging existing levels o f expenditure more effectively, though perhaps with changes to composition. A pro-poor expenditure framework is requiredto ensure adequate resources get to responsible spendingunits and levels o f government, consistent with the vision o f the CPA. Budget links to the Five Year National Strategic Plan and the upcoming PRSP should help in this regard, particularly indeepening the analysis o f sectoral priorities. 3. Focus on fulfilling the vision o f fiscal decentralization in the Interim National Constitution, namely monitoring the adequacy and efficiency o f state and locality expenditures, and improving the equity, predictability and transparency o f the intergovernmental transfer system. Successful fiscal decentralization will also require clarifying the legal framework o f federal, state and local government responsibilities and powers, and institutional arrangements for grant allocation and monitoring. In addition, basic state economic data must be collected to bolster the quality o f own-revenue estimation at the state and locality level. 4. Address the planning and implementation deficiencies o f development proiect s u ~ ~ o rincluding at the sector program andproject level, and at both the federal and t , sub-national levels. The dearth o f development spending for poor and marginalized areas at the state and locality level must be addressed, given the expenditure responsibility for basic service delivery. However, assessing the appropriate balance X betweenrecurrent and capital expenditureswill require more detailed sector-specific analysis than i s currently available. In addition, capacity should be built for evaluating investment projects, including modernizing policies and institutions for program evaluation. 5. Strengthenaccountabilitv for the use o f public funds, financial management, and the overall results culture. On-going reforms inthese areas include GFS application and consolidated budget reporting, but there i s a strong need to bring expenditure analysis into decision-making in new areas such as the growth in the wage bill, sector allocations, rationalizing defense spending, and arrears management. These efforts will improve budget credibility and provide direct impact to the efficiency o f spending. Demand-side accountability can also play a critical role and should be supported, for example through Parliament, the Auditor General, and NGOs. The financial management agenda will be informed by the on-going CIFA work. The agenda i s particularly daunting in Southern Sudan. 6. Devote sustained and dedicated efforts to address the very difficult fiscal situation now facing the GOSS. A four point agenda has been identified in the background work-cash management, civil service, revenue management, and debt-and a high level fiscal mission in June 2007 should provide key entry points to these issues. The momentum o fthis on-going support to the GOSS must be continued. 7. Build capacity for more effective resource allocation and use, including in the states where responsibility for front line service delivery lies. While this is the last agenda point, it i s the most fundamental and cross-cutting, as i s necessary to address the abovementioned agenda items. The path to successful implementation o f the CPA mustinclude buildingcapacities at the national and sub-national levels to fulfill roles and responsibilitieso f fiscal decentralization inSudan. x i 1. BACKGROUND 1.1 Emerging from nearly four decades o f civil conflict, the newly formed Government o f National Unity agreed in late 2005 to engage with the World Bank and other external partners on analysis and dialogue on the public expenditure agenda. This agenda supports reforms envisioned in the Comprehensive Peace Agreement (CPA) and Interim National Constitution (INC) which are seen as vital to demonstrate equitable and transparent resource allocations, and thereby consolidate peace and reduce the risk o f future conflict. Given the breadth o f the agenda, the approach i s programmatic. This Public Expenditure Review (PER), the first for Sudan, reflects the initial phase o f engagement, with the objective o f identifying and supporting improvements in budget planning, monitoring and institutional arrangements so as to contribute to improved resource use and outcomes. The emphasis i s on establishing a baseline understanding o f key fiscal management and policy challenges, highlighting reform measures for policymakers, and setting the agenda for the next phase o f work. 1.2 This synthesis provides a baselineassessment of two fundamentaldimensions of publicfinance in post-CPA Sudan-fiscal managementand the alignmentof resources to broad-basedgrowth and reducedpoverty and inequality.' As established inthe INC, Sudan's vision of fiscal decentralization has important implications for both o f these dimensions, and the PER pays special attention to the issues of intergovernmental transfers and sub-national fiscal management, inthe Northern states as well as the new Government o f Southern Sudan (GOSS). A. COMPLEXANDUNCERTAIN POLITICAL ECONOMY TEXT 1.3 The current politicaleconomy context in Sudan is complex and uncertain, even for a post-conflict country. In January 2005, the Government o f Sudan and the Southern Sudan People's Liberation Movement (SPLM) signed the CPA and entered a six-year interim period after which a decision on national unity will be made by referendum. Thus the inherent fragility of a post-conflict situation has an extra-ordinary element o f political uncertainty. The peace deal brokered between the North and South contains extensive provisions for sharing o f power and wealth, including most notably a new autonomous regional government inthe South, the creation o f the Government o f National Unity (GNU), and equal sharing o f oil revenues generatedinthe South. 1.4 The politicaland legal context now reflectedin the INC has key implicationsfor expenditure management(Box 1.1). First and foremost is fulfillment of the wealth sharing obligations o f the peace agreements, which limits the discretion o f GNU spending as half o f oil revenues generated in the South must be transferred to the GOSS (roughly 15 percent of A backgroundvolume available on the World Bank's Sudan externalwebsite contains a compendiumof supportingbackgroundwork. Please visit www.worldbank.org/sd. total GNU expenditures in2005-2006). Second, the vision o f fiscal decentralization i s a key aspect o f a unified and peaceful Sudan, potentially addressing inequalities, marginalized areas and the root causes of conflict. This also has significant implications for discretionary spending of the federal government, as a greater share o f revenues is transferred to the Northern states. And third, accountability at all levels of government i s required to ensure resources are shared equitably and used in accordance with each level o f government's constitutional responsibilities. Box 1.1: A Visionfor EquitableUse of NationalWealth inthe 2005 Interim National Constitution "Resources and common wealth o f the Sudan shall be shared equitably to enable each level o f government to discharge its legal and constitutional responsibilities and duties and to ensure that the quality o f life, dignity and living conditions of all citizens are promoted without discrimination on grounds of gender, race, religion, political affiliation, ethnicity, language or region." (Excerptfrom INCArticle 185). The InterimNational Constitution provides the vision for equitable andtransparent use of national wealth to sustain peace, implement the CPA and achieve poverty reduction andhuman development. Key dimensions include equitable treatment o f marginalized areas, commitment to fiscal decentralization to allow all levels o f government work toward achieving targeted outcomes especially with regards provision o f services, and the principles o f transparency and accountability. The sharing and allocation o f the resources and common wealth o f the Sudan is premisedthat all parts o f the country are entitled to development, andrevenuesharing shall reflect a commitment to devolution o f powers and decentralization o f decision-making in regard to development, service delivery andgovernance. There is significant detail in the Constitution regarding the distribution o f responsibilities between levels o f Government in Northern Sudan. Most notably, the responsibility for basic service provision is assigned to the sub-national level. The clarity of these expenditure assignments is crucial for expenditure management, andthere remains significant room for improvement regarding state versus locality levels. Annex A details the Interim National Constitution's assignment o f responsibilities between levels o f Government inNorthern Sudan. The South. as an autonomous region. has a setmate interim constitution. Source: InterimNational Constitution. B. MACROECONOMIC BUTBUILDINGPRESSURES STABILITY 1.5 Sudan has sustained macroeconomic stability following the turbulent 1990s, and has performed satisfactorily under a subsequent IMF Staff Monitored Program. Economic growth averaged 6 percent per annum between 2000 and 2004, and at over 10 percent in 2006 i s currently among the highest on the continent. Agriculture remains the largest share o f GDP (roughly one-third), but the oil sector and associated investment and services boom have driven the economy since the turn of the century (Table 1.1 and Figure 1.1). 1.6 Inflation had been largely kept to single-digits, but internal pressures are a growing concern. Through the first half o f 2006, price growth averaged 4 percent year on year, but high growth in monetary aggregates and the August 2006 domestic fuel price increase contributed to inflationary pressures with end-2006 year-on-year inflation growth at 16 percent. Inflation averaged 8-9 percent for the first half o f 2007. 2 Table 1.1:Trends inKey Macro and Fiscal Indicators, 2002-2006 2002 2003 2004 2005 2006 est (annual changes in percent) RealGDP 5.4 7.1 5.1 8.6 12.2 Oil 24.7 13.0 9.9 -0.2 26.5 Non-oil 3.8 6.6 4.7 9.5 10.8 Consumerprices (endof period) 8.3 8.8 7.3 5.6 15.7 (in percent of GDP) GNUrevenue 11.9 16.0 19.7 21.7 19.4 GNU expenditure 8.8 15.3 18.2 23.5 23.7 Overall balance(cash basis) 3.1 0.7 1.5 -1.8 -4.2 Non-oil balance -2.4 -7.6 -8.3 -15.4 -15.9 (in millionsof US dollar, unless otherwiseindicated) Exports f.0.b. 1,949 2,577 3,778 4,859 5,743 of which: oil 1,511 2,082 3,101 4,221 5,174 Imports, f.0.b. 2,153 2,536 3,586 5,946 72 Current account balance(cash basis, % of GDP) -6.1 -4.7 -3.8 -8.3 -12.9 Real effective exchange rate (change inpercent) 2.6 -1.4 2.9 17.3 21.0 External debt (inbillions ofU S dollars, endofperiod) 23.6 25.7 26.0 27.7 28.2 External debt (inpercent of GDP) 157.6 144.6 119.9 99.3 75.4 Net international reserves 84 290 1,144 1,889 1,384 Source: IMF Figure 1.1: Agriculture, Mining and Services dominate the economy (2006 GDP by sector, at factor cost) Construction? EIec & water 2 Manufactr.7 A utner services Trade, restaurants& hotels Source: IMF 1.7 The current account deteriorated in 2006 as import demand far outstripped exports, including lower-than-expected oil exports. The current account deficit is approaching 13 percent of GDP in 2006. Major imports include machinery and equipment which are part o f an ongoing investment boom. Imported fuel costs are also significant, 3 though in-country refinery capacity i s expected to meet domestic consumption demand inthe next few years. Foreign direct investment inflows more than covered the current account deficit, helpingforeign reserves grow to roughly 3 months o f imports by end 2006. 1.8 Inflows of foreign exchange from oil exports and FDI, as well as the recentfiscal expansion, havecontributedto a rapid appreciation of the real exchange rate since 2005 and raisedsignificantcompetitivenessconcerns for non-oil exports and sustainablelong term growth. The dinar experienced double-digit appreciation in2006, reaching the current level o f around 200 dinar per dollar, compared to 250 in early 2005. While the authorities and IMF have supported the high appreciation as necessary for containing inflation, preliminary evidence from the on-going Diagnostic Trade Integration Study suggests profit margins for farmers have been squeezed by the appreciation, in addition to the adverse effects o f supply-sidebottlenecks. 1.9 Sudan's debt arrears prevent access to concessional finance. At end 2006, external debt was $27 billion innominalterms, with over 80 percent inarrears. This includes IDA arrears around $420 million and IMF arrears of around $1.6 billion. Donors have signaled that clearance o f arrears will depend on progress on both the implementation o f the CPA and the peaceful resolution o f the Dark conflict, and direct discussions on debt are stalled. Beyondthis, Sudan i s potentially eligible for debt relief under the HIPC initiative, but only once the government has prepared an interim or full Poverty Reduction Strategy Paper, and cleared its outstanding arrears. Debt relief under the Multilateral Debt Relief Initiative could also be available subsequent to reaching HIPC Completion. Improved external debt indicators reflect higher exports and GDP, but nominal debt keeps rising and it remains unsustainable. C. DAUNTINGDEVELOPMENT CHALLENGES 1.10 The signingof the CPA opened tremendous opportunities to accelerateprogress toward the Millennium DevelopmentGoals and address the wide regionaldisparities in humandevelopmentoutcomes. However,progress so far has been limited. 1.11 Sudan ranks 141 out o f 177 countries in UNDP's 2006 Human Development Index, with 2006 per capita GDP over $1,000 (more than one-third higher than the SSA average). There is no recent representative income or consumption data from which to calculate poverty headcounts, but preliminaryfindings on some non-income poverty outcomes are now available from the nationally-representative Sudan Household Health Survey (SHHS) conducted in 2006. Outcomes for a number o f individual MDG-related indicators show Sudan's low achievements relative to other countries at the same income level and even muchlower income levels (Figure 1.2). The results o f the Sudan Household Health Survey remain under discussion, and the authorities have not endorsed the preliminary findings. The analysis presented in this report should be considered tentative, but given the dearth o f social data in Sudan it is presented as the best currently available. It is the first nationally representative survey in two decades and was commissioned to follow-up on the CPNJAM. The sample size is 25,000 households, an average o f 1,000 households per state, and an average o f 25 households per cluster (40 clusters per state). Earlier data for the North cited here are drawn primarily from surveys including the SMS (1999) and MICS (2000). 4 Figure 1.2: Sudan I s a Country Outlier on Education (Net Primary Enrolment) Relative to Income 120 100 80 h " i? 1 6o 40 20 0 1.12 The SHHS confirms that the major challenge facing Sudan's progress towards the MDGs is the massive inequality in outcomes and access to services (Table 1.2). Overall, there appears to be little evidence o f aggregate improvements in social indicators such as educational enrolment or infant mortality. Moreover, the picture i s mixed across regions and sectors-and particularly in conflict-affected areas, such as Darfur. Most averages also hide large gender disparities, with women suffering disproportionately from limitedaccessto basic social services. Table 1.2: Large RegionalDisparities in Key MDG Indicators Average of Average of National Indicator Highest 5 States Lowest 5 States Average MDG 2 Net primary school attendance rate 84 7 53 MDG 3 Ratio of girls to boys attending 0.99 0.61 0.93 primary education MDG 4 Measlesimmunization coverage 82 31 65 MDG 5 Proportion ofbirths attendedby skilled 91 24 57 healthpersonnel MDG 6 Comprehensiveknowledge aboutHIV 19 2 9 prevention MDG 6 Contraceptiveprevalence 16 1 8 MDG 6 Proportion ofpopulation using 50 10 28 effective malaria prevention measures MDG 7 Use of improved drinking water 80 35 59 sources MDG 7 Use of improved sanitation facilities 66 5 31 Source: SudanHouseholdHealth Survey (2006). 5 1.13 These preliminary findings from the 2006 survey suggest that getting onto a trajectory consistent with meeting the MDGs will involve a significant acceleration of progress. At the same time, data constraints and the lack o f a reliable baseline make it difficult to draw firm conclusions, and further analysis o f the 2006 survey i s needed. A closer look at geographical disparities (intra-state, as well as across states), as well as nual/urban and gender patterns, is needed to better understand trends. Looking forward, it will be important to strengthen capacity for monitoring and benchmarking relevant data to inform policy decisions and guide expenditure allocations. Very few o f the JAM targets on improving monitoring systems have been met, which means that critical gaps remain in the regular information needed to underpin the formulation o f a national poverty reduction strategy and the more comprehensive Five Year National Strategic Plan. D. ANUNSUSTAINABLE FISCALEXPANSION 1.14 The macroeconomic stabilization of the late 1990s was underpinned by a fiscal stabilization program. The center piece was expenditure control through a cash budget system (Figure 1.3). Fiscal imbalances were drastically cut, mainly through reductions in non-wage recurrent expenditures (e.g., transfers to the states, debt service payments, and ' O&M) and development expenditures. Priority o f expenditures ranked wages and salaries at the top (Chapter l), second (Chapter 2), transfers to the states third (Chapter 3), and O&M development last (Chapter 4). Annual fiscal deficits were quickly brought under control. Figure 1.3: Recovery since sharp fiscal adjustment in the 1990s 30 25 20 ? 15 0 10 5 0 Source: MOFNE and staff estimates. 1.15 Revenue collections increased significantly with the advent of oil production in 1999. However, the fiscal picture deteriorated in 2006 driven by oil revenue shortfalls and expansion of spendin some of which relate to the CPA (e.g., transfers to the GOSS and to the Northern states)!' Remedial actions within the context o f the IMF program were taken by the authorities in 2006 to cut expenditures. This included an increase in domestic The DPA in May 2006 and ESPA later in the year both brought additional budget obligations of the GNU. Over halfthe plannedincrease inNorthern transfers inthe 2007 budgetcan be attributed to the DPA and ESPA obligations. However, under the current fiscal environment, it is uncertainifthese plans will be met. 6 fuel prices to trim the subsidies o f domestic fuel consumption by half and large withdrawals from the Oil Revenue StabilizationAccount (ORSA). The latter have exhausted the fund and preclude its ability to provide relief from future revenue shortfalls without significant replenishment. 1.16 The fiscal balance has turned to growing deficits in 2005 and 2006, following three years of surplus. Including the rundown in the ORSA reserves, the projected 2006 budget deficit jumpedto 4.2 percent o f GDP, from 1.8 percent of GDP in2005 and surplus in the precedingthree fiscal years. The budgetdeficit inthe 2007 budgetexceeded 6 percent of GDP. The majority o f deficit financing has been domestic, two-thirds in 2006, including government securities and Ministry o f Finance and National Economy (MOFNE) net borrowing from the Central Bank. Foreign financing has increased significantly, reaching about $450 million in2006, compared to an annual average o f $45 million from 2000-2003. 1.17 Fiscal expansion at this pace appears to be unsustainable. The 2007 budget increased total expenditure by 13 percent in nominal terms over last year's budget to a total o f $11.8 billion. Large increases are plannedfor both federal wagedsalaries and transfers to Northern states (roughly 25 percent nominal increases for each area). National development i s budgetedto increase by roughly half, The likelihood o f meeting these expenditure plans i s low, at least without high levels o f domestic or external financing. The pressure for expenditurecuts and reallocation is high, raising the relevance o f this PER and the need for subsequent government action to address the fiscal imbalances and avoid weakening the country's macroeconomic credibility. 1.18 The fiscal pressures in the South have proven even more acute. The overall balance in2006 moved sharply into cash deficit due to limited spending discipline as well as oil revenue shortfalls in the second half o f the year. Aggregate spending was driven by outlays on wages and operations that were roughly double planned amounts, while capital expenditures were cut sharply. In the first half o f 2007, oil revenue shortfalls continued, GOSS significantly over-estimated non-oil revenue, and in the face o f huge development needs, expenditure plans in the 2007 budget were significantly expanded relative to 2006. The cash reserves that could have provided cushions were largely run-down last year and are no longer available. One risk is that plannedinvestmentsinroads, schools, clinics and so on will be squeezedby the burgeoningpayroll. 1.19 The rest of this synthesis report is structured as follows. Sections 2 and 3 cover GNUrevenue and expenditure management, respectively. Section 4 summarizes key aspects of the PER dialogue on identifying and monitoring pro-poor spending in Sudan. Intergovernmental transfers and fiscal decentralization are the focus o f Section 5, and section 6 summarizes the main findings from the PER state case studies. Section 7 is dedicated to fiscal issues inthe South. 7 8 2. MANAGINGAND MOBILIZINGNATIONALREVENUES 2.1 Sudan's revenue structure mirrors that of a number of other low-income oil producing countries, with a high dependence on oil revenue and relatively weak non-oil revenue collection (Table 2.1). While a number of improvements have beenmade in recent years, at about 6-7 percent of GDP, Sudan's tax effort is low compared with countries at a similar level of development. Direct taxes amount to only about 1.2 percent and indirect taxes for 4.3 percent of GDP. While the oil sector does not dominate the economy (contributing roughly 10 percent of GDP), oil is of critical importance to government revenue and the balance of payments, with its revenues accounting for 56 percent of revenue in2006. Table 2.1: Growth in Resources from Oil Revenues 2001 2002 2003 2004 2005 2006 (as share o f GDP) Total revenues 10.7 11.9 16.0 19.7 21.7 19.4 Tax revenue 5.5 5.4 5.8 7.5 6.9 6.3 Direct taxes 1.2 1.o 1.1 1.3 1.4 1.2 Indirect taxes 4.3 4.4 4.7 6.2 5.6 5.1 Trade Taxes 2.2 2.5 2.4 2.8 3.O 2.6 Excise duties 0.9 0.8 1.1 2.0 1.3 1.3 VAT 1.1 1.1 1.2 1.3 1.3 1.2 Non-tax revenue 5.2 6.5 10.2 12.2 14.7 12.7 ofwhich: Oil 4.3 4.6 8.4 10.3 13.3 10.8 Source: Ministry o f Finance and National Economy, and staff estimates. A. OIL REVENUEMANAGEMENT 2.2 Sudan is one of the newest significant oil producing countries in the world. Significant production started in 1999, andthe country is now the third largest oil producer in Sub-SaharanAfrica, behindNigeriaand Angola, with output at about 364,000 barrels per day (bpd) in 2006. Production forecasts are difficult for Sudan given the lack of detailed information regarding specific major projects, described below, but available data suggests peak production in the next two years near 730,000 bpd, holding for roughly five years and thendeclining. 2.3 Oil revenue in Sudan-from export of crude and domestic sale of refined products-has proven volatile and oil production has fallen short of expectations in recent years. Oil revenue amounted to some 55 percent of total government revenue in 2006. Crude oil exports accounted for roughly 84 percent o f total exports in 2006, at some $5.3 billion or 14 percent of GDP. Lower than expectedproduction and price for Dar Blend 9 have driven significant oil revenue shortfalls relative to budget plans (Figure 2.l), creating fiscal pressures and undermining budget credibility given over 50 percent of total GNU revenuecomes from oil. Fiscal pressures have been particularly strong for the South, where tax and customs revenue are minimal and oil revenue accounts for over 90 percent o f the budget via transfers from the GNUas specified under the CPA.4 . Figure2.1: Oil RevenuePerformanceDeteriorates 1000 Budget I 1 % execution A n 150 u) -5 I- 5c 500 100 5 0 nrl 50 2001 2002 2003 2004 2005 2006 2007:Ql Source: MOFNE and staff estimates. 2.4 The quality of Sudanese crudevaries substantiallybetweenfields,with two main blends currently produced-Nile Blend and Dar Blend. Nile Blendis a relatively good quality, priced at a discount o f $2-3/bbl relative to the IMF's benchmark WE0 price. Dar Blend-which came on stream in October 2006 and was forecast to account for the majority of the production increase in 2007-is a lower quality, high acid crude. Initial sales o f Dar Blend in 2006 were disappointing, and the price of Dar Blend in early 2007 stood at an exceptionally large discount o f about $30-34/bbl relative to the IMF's benchmark WE0 price. These low prices have beenattributed to problems infinding buyers able to refine this type o f crude. By June 2007, the discount o f Dar blend had narrowed to about $20/bbl, reflecting that this crude i s gaining acceptance inthe international market. 2.5 While the discovery and exploitation of oil resources has facilitated an increase in nationalwealth, it has also brought a myriad of problems.Oil revenues have expanded budget resources, potentially available to support pro-poor development. The political economy o f oil has also brought additional pressures for expenditure (some o f which may be conditioned by capacity and other constraints), rent-seeking behavior, and reduced pressure to undertake key fiscal reforms that will ensure medium- and long-term stability irrespective o f the course o f oil production and prices. 4Much o f Sudan's oil production currently takes place in the South, and the CPA specifies that revenue generated in the South is divided equally between the GNU and GOSS after a 2 percent share to the oil producing state government. Revenue from production inthe North goes directly to the GNU, after 2 percent to the specific state where the wells are located. 10 2.6 The Oil Revenue StabilizationAccount (ORSA), which was establishedin 2002, i s now depleted despite rising production and record world prices.' The account accumulated a significant amount of funds over 2002-05, equivalent to 1 percent o f GDP at end-2005. However, there were substantial draw-downs from the account in late 2006 to finance government expenditures, and by end-December 2006 the account was virtually depleted.6 Thus the ORSA, in its current state, is not able to provide any further risk management benefits. A concerted effort i s needed to rebuild the account, along with stronger controls and management to ensure it can play a role in longer term oil revenue management. 2.7 Transparencyin the sector is unusuallyweak, incomparisonto many oil exporting developing countries. Specifically, as described below, the Ministry o f Energy and Mining does not produce detailed statistics or reports about the sector or about project developments, the state oil company provides no public accounts, and there is very little information about business developments from the companies operating in Sudan. In most oil exporting countries one or more o f these sources o f information provides a basis for assessing sector trends and policy. Moreover, it should be noted that nine oil-rich countries in sub-Saharan Africa have joined the Extractive Industries Transparency Initiative (EITI), an international process committed to improving revenue transparency and public accountability in oil, gas and mining. This lack o f transparency appears to be contributing to Sudan's current fiscal problems by making projections o f oil production and revenues relatively inefficient. The Government o f National Unity (GNU) plays a major role at every stage o f production, refining, and marketing through the state-owned Sudan Petroleum Corporation (SPC). Sudapet i s the subsidiary o f SPC that holds the Government's equity share in specific oil exploration and production projects. Although the GNU Ministry o f Energy and Mining i s notionally responsible for the regulation and supervision o f the sector, the GNU'Sregulatory fhctions are not independent o f its commercial activities inthe sector. 2.8 The SPC reportedly prepares an annual report that is sent to the National Petroleum Commission but is not made publically available. It provides accounts to Parliament, though on government terms rather than as a corporate entity. Further, accounts for SPC's subsidiaries are rarely available. For example, Sudapet, which i s becoming a substantial oil producing company, does not publish annual reports or accounts. Improved accounting standards at the SPC and its subsidiaries would contribute to a more transparent oil sector. 2.9 On the positive side, the 2006 IMF program requirement that oil production, export andrefining data be regularly publishedhas been largely implemented. Inaddition, details o f the monthly baseline production and pricing data for the calculation o f the revenue share of the Southern Sudan have been shared and provide an insight into the development o f the petroleum sector. However, these statistics are tailored for the respective requirementso f the 5 The ORSA is a locked sub-accountfor the GNU at the Bank of Sudan, controlled by the MOFNE. At the start of the fiscal year, a benchmark production figure and oil price is agreed by the GNU and GOSS. Any revenues accruing fiom production or price above the benchmark are deposited in the ORSA. Withdrawals are distributedto both GNU and GOSS inproportion to their share oftotal oil revenue. Under the terms of the agreement on the ORSA, when the GNU accesses hnds from the account to finance expenditures, it must transfer an additional amount to the GOSS in accordance with its share of oil revenue- regardlessof whether the GOSS needs the finds at that time. 11 IMF and the CPA, and do not provide a comprehensive base for assessing the petroleum sector. Moreover, the statistics are publishedin summary form and without comment by the Ministryof Finance, and few relevant statistics are publishedby the Ministry of Energy and Miningor the official statistical agencies. 2.10 Transparency i s not facilitated by the fact that petroleumexploration and production inSudan are carried out almost entirely by "new" international companies from China, India and Malaysia, o f which the Chinese National Petroleum Corporation-CNPC-is the largest holder of reserves and production in Sudan. Some o f these companies have less comprehensive accounting and reporting standards than the international norms for the sector. 2.11 Going forward. Oil represents tremendous potential for Sudan, but recent experience has been mixed. The large boost to revenues may have weakened incentives for expenditure management and fiscal reforms. Accountability is underminedby the lack o f transparency that severely hampers the ability to properly manage the oil sector (interms o f natural resource management) and to adequately forecast and manage revenues, which in turn compromises overall fiscal policy. Volatility and over-estimation o f oil revenues documented in Figure 2.1 has undermined budget credibility, forced expenditure rationing, caused complete elimination o f the stabilization fund built up with past savings, and presentlythreatens the macroeconomic stability sustained over the past ten years. 2.12 Several challenges are outstandingwith respect to management of Sudan's oil revenue,first and foremost beingthe need for transparency. World class transparency is entirely consistent with commitments already made inthe CPA and INC. While the sharing o f oil revenue between North and South appears to be functioning broadly as outlined inthe CPA, the lack o f detailed information regarding other aspects o f oil management and accounting i s a cause for concern. Particularly with respect to state and quasi-state companies that deal inthe sector, Sudan would benefit from a greater commitment to transparency and adherence to international accounting standards and best practices. Concrete recommendations could include: full accounting for activities o f the state-owned oil companies and their subsidiaries (SPC, Sudapet, and affiliated operating companies); separating SPC, Sudapet, and the Ministry o f Energy to clearly delineate regulatory versus market roles in line with international standards; reform o f the rules governing the ORSA to replenishmentthe account and make it a sustainable tool for risk management; and clarify further the role and responsibilities of the National Petroleum Commission (NPC) with respect to oil sector policy. It i s also important to finalize division o f oil productionbetween Northand South, which canonly occur once the relevant border hasbeenagreed. 2.13 Improving the transparency and predictability of oil revenues is critical for fiscal management. The shortfall in oil revenue in 2006 forced major deviations from planned GNU expenditures and fiscal pressure on all aspects o f the budget. These pressures were partially offset by exhausting all available resources in the ORSA. However, the underlying production and price problems continued into 2007 and expenditure plans have been impacted more directly, without the benefit o f oil savings. The pressures have been most acute for the GOSS,given its reliance on oil revenues to finance nearly all o f its budget. 12 The abovementioned reforms to improve transparency should help enable a more robust and stable estimate o f future oil revenues, and ultimately more prudent fiscal planning. B. NON-OIL REVENUE UNAGEMENT 2.14 Tax policy and administration in Sudan is relatively weak. At about 6-7 percent of GDP, Sudan's tax effort i s low for its level o f development (Table 2.2). Direct taxes amount to only about 1.2 percent o f GDP and comprise a Business Profit Tax (BPT), a Personal Income Tax (PIT), a tax on Sudanese residents abroadY7and various stamp duties. Indirect taxes (customs duties, VAT, and excise taxes) account for roughly 80 percent o f total tax revenue, with taxes on international transactions dominating. As a whole, the tax system i s fragmented, unevenly applied, and suffers from widespread exemptions and tax holidays that limit the effectiveness o fthe existing regime-a weakness made possible by continued gains fromthe oil sector. Table 2.2: Relativelyweak tax efforts: Sudanand selected comparators 1/ ~ ~~~~~ ~ ~ Average 200 1 2002 2003 2004 2005 2006 2001-06 (inpercentof GDP) Sudan 5.5 5.4 5.8 1.5 6.9 6.3 6.2 Algeria 9.3 10.6 10.0 9.5 8.5 8.2 9.4 Egypt 21 13.4 13.4 13.3 13.8 14.1 15.9 14.0 Kenya21 16.5 15.9 16.6 17.0 18.0 16.6 16.8 Nigeria 17.2 14.2 15.0 16.6 16.8 16.1 16.0 Yemen 7.3 7.3 7.1 7.3 7.3 6.6 7.2 Unweightedaverage 12.7 12.3 12.4 12.8 12.9 12.7 12.7 Sources: IMF country documents; Governmenf Finance Sfatisfics (IMF); andFundstaff estimates liForallcountries,exceptforEgyptandNigeria,thefiscalcoverageisthecentralgovernment.ForEgypt,thefiscalcoverage includes central government, NIB, and social insurancefunds. For Nigeria, the fiscal coverageincludes the federal, state and local governments. 21 For Egypt and Kenya, the fiscal year ends on June 30th. 2.15 The need for improved revenue collection and tax administrationinSudanis clear. A number of significant reforms have been introduced in recent years but Sudan's non-oil revenue base remains thin. Reform of both direct and indirect taxes will be critical to ensuring medium-term fiscal sustainability as well as non-inflationary domestic financing. There is a critical need for a comprehensive tax administration modernizationstrategy, which articulates a vision for tax administration over the long term. 2.16 While some major improvements have beenmade to the system o f direct taxes (such as streamliningthe customs tariff framework), some elements are distortionary, discretionary and inequitable. Corporate tax incentives exacerbate existing distortions and inequities, narrow the tax base, and undermine the overall tax effort. A key example o f this are the unequal tax holidays and import duty exemptions to investors under the Investment Encouragement Act. In particular, reform o f the direct tax structure could include the 'This tax was reducedsignificantlyinthe 2007 budget,andmay be eliminatedinthe future. 13 rationalization o f BPT rates and coverage in conjunction with a larger effort to eliminate exemptions and tax holidays. In addition, all major allowances and exemptions allowed as part ofremunerationcould be brought underthe PIT. 2.17 Indirect taxes account for the lion's share o f total tax revenue in Sudan, with customs duties representing the single largest component-with the VAT a close second. The main challenges facing the indirect tax regime in Sudan are exemptions and the level o f the VAT threshold. While customs duties play a critical role and account for about a third of tax revenue, the regime has an excessive number o f exemptions. In particular, efforts should continue to reduce tariff dispersion, and the top tariff rate. Finally, there i s a lack o f cooperation between GNU and the GOSS regarding the application o f customs duties and revenuecollection-ffectively creating two different trade regimes. As with customs duties and business taxes, reform o f the VAT should focus on rationalizing rates across sectors and eliminating exemptions and holidays. Current exemptions from the VAT reduce the revenue take by roughly 1.2 percent o f GDP. 2.18 Inorder to strengthenthe process ofnon-oil revenue collection, there is the needfor a renewed effort toward reforming the tax administration and developing a comprehensive reform plan. Such a plan must include the development o f an adequate information management system and the further development o f human resources. Additional measures could include the further reform o f the Tax Council (TC) including strengthening o f the large taxpayer office (LTO) and the medium taxpayer's office (MTO), and reform o f taxpayer compliance to avoid further evasion o f taxes and the resulting need for a cumbersome administration. 2.19 The current headquarters structure o f the TC includes both tax-type and function- based departments. Key steps could include a progressive restructuring o f the network o f operational offices, with separate offices for LTO andMTO, and regionalhtate offices for the remaining taxpayer segments. Key steps towards strengthening the LTO and M O i s a rationalization o f processes used in administering tax types which could include: (i)the adoption of unambiguous, transparent and easily understood selection criteria for the selection o f taxpayers; and (ii) moving to a fully integrated approach for all functions across tax types. 2.20 The heavy reliance on summary or arbitrary assessments hampers taxpayer compliance and perpetuates mistrust between the tax authorities and taxpayers. Full self assessment should be introduced progressively-focusing first on LTO followed by MTO. This should be accompanied by a comprehensive training program on self-assessment principles and risk management for all LTO and MTO staff. c. CHALLENGESARISING FROMDECENTRALIZATION 2.21 Fiscal decentralization, while key to the vision o f the CPA/INC, poses a number o f additional challenges to revenue management, Sub-national levels need adequate revenue to conduct expenditure assignments and address regionalAoca1 needs, and at present there are inadequate and unevenly distributed own-revenues at lower levels o f government and 14 transfers have beenunpredictable. A number of dimensions o f fiscal decentralization require attention to ensure fulfillment o fthe vision o f fair and transparent wealth sharing inSudan. 2.22 For the oil sector, greater clarification i s neededas to the role and responsibilities o f the National Petroleum Corporation (NPC), as well as the division of oil production between the North and the South. The CPA provides for broad petroleum sector policy to be set by NPC, with equal representation of the North and South. The NPC also has within its original remit the authority for issuing new exploration and production licenses. The NPC has met infrequently to date, and with limited results. More rapid progress in this area could prove pivotal to resolving emerging problems with respect to new production and associated revenues. Current fiscal difficulties stemming from oil revenue volatility and uncertainty in the South underline particular vulnerability, given their near complete dependence on oil revenue and weak capacity to manage oil sector issues. Many o f the most needy Northern states also are severely impacted, given state transfers are discretionary and have the lowest priority o f GNU spending, especially development transfers. The CPA also provides for commissions to determine the boundary between the North and the South, and for the GNU and the GOSS to establish in which region oil wells are located. The work o f the boundary commissions has not beencompleted, and hence the division o f production appears not fully definedinaccordance with the CPA. 2.23 On tax policy, greater clarity i s needed in terms o f the assignments o f tax authority and revenue collection-both within the federal government and between the federal government, the GOSS and the Northern states. Early action phased-in over a reasonable timeframe will help prevent the need for more drastic adjustments later on (Box 2.1). Harmonization o f the personal income tax base between the GNU and the states i s vital. Preferably, the PIT should be a guide and orhehicle for the states to choose and apply the state tax rate, although, currently, no states have yet approached the GNU to discuss this issue. On excises, while many federal state systems do allow sub-national governments to levy excises, it i s preferable that the base be common across states and that excise rates not differ substantially between one state and the next. 15 Box 2.1: Medium-Term Gains from Tax Reform Sudanhas undertakena number oftax reforms since 2000, includingintroductionofthe VAT, impositionofthe excise on benzene, and improvementof the tariff structure.As a result, tax revenuerose from 5.5 to 7.5 percent o f GDP during 2001-04. However, tax revenue fell back to 6.3 percent of GDP during 2005-2006, and ranks amongthe lowest inthe region. Giventhe challenges ahead, includingoil revenue volatility, trade liberalization, and investing in development and sustained peace, further efforts are needed to increase tax revenue, notably through improvements in tax policy and revenue administration. The followingprovides a basis for a medium-termtax reform-based on a set ofmeasuresrecommendedby the IMF-and anestimate ofthe potential gains over the mediumterm. a Replace the Investment EncouragementAct (IEA) with a single profit tax rate and accelerated depreciation allowance. Unifying the business profit tax at 20 percent and eliminating exemptions underthe IEA could generate about 2 percentofGDP inadditionalrevenue. a Introduce a meaningful personalincome tax. The currentPIT base is extremely narrow, with a large number of allowances. Almost all government employees and around 80 percent of private sector workers are exempt from it. This couldraise PITrevenue fromthe current levelof 0.1 percent of GDP to 1-2 percent of GDP. a Rationalize and/or eliminate VAT exemptions, inparticular on capital goods, water, electricity and other consumer goods. Giventhe erosionofthe VAT baseover time, VAT productivityis amongstthe lowest inthe world, merely 0.12. Eliminatingthe VAT exemptionscould generate additional revenue o f about 1.O-1.5 percent o f GDP. a Rationalize and/or eliminate custom exemptions. Eliminating exemptions not related to international treaties, while reducingthe maximum tariff rate (40 percent), could generate additional revenue o f0.5-1.O percentofGDP, includingthroughreducedsmuggling. Improve tax administration by strengthening the Tax Chamber. Restructuring the Tax Chamber; strengthening the LTO and MTO; and improving compliance programs to support a proper self-assessment systemcouldyieldadditionalrevenueof0.5-1.0 percentof GDP. ource: IMF. 2.24 Weak coordination between North and South. Emerging from several decades o f conflict, South Sudan was a region without a formal tax administration. Since 2005, the GOSS has taken steps to reestablish a structure, but there are emerging issues regarding consistency with the GNU. The 2005 Interim National Constitutions (INC) assigns responsibilities for collection o f specific revenue sources to the GNU, the GOSS, and state levels o f governments. However, recent IMF mission have suggested that the current tax system in South Sudan differs significantly from the structure outlined in the INC, and that the legal framework outlined inthe CPA and INC is not in force. The GOSS collects duties based on a preexisting customs and excise tariff adopted in 2000, which differs from the current national customs law. GOSS also collects export duties on agricultural products (1 percent), a list o f other goods (with four different rates), and re-exports o f finished goods (at 30 percent). It collects personal income tax from Sudanese working with NGOs (at rates o f 5 and 10 percent), and road tolls based on an excise type tariff. State authorities within the South also collect a variety o f taxes, fees, and duties. 2.25 Coordination between the North and the South on a range o f tax issues, such as customs collection and non-oil revenue transfers, i s important. Devolution o f taxes need not entail a decentralization o f tax administration, and there are pros and cons. A careful review 16 o f the current tax system in Southern Sudan could help to ensure that the collection effort i s guided by the provisions o f the CPA and the INC, and that duty and tax collection are supported by valid legislation. Care should be taken to clearly lay out the respective roles o f national, GOSS, and state revenue authorities in administering the tax system. Given the importance o f transparency, it will also be critical to ensure that basic transparent tax assessment and collection procedures are definedand supported by consistent processes. 17 18 3. KEYISSUES INGNUEXPENDITUREMANAGEMENT 3.1 Fiscal management since the CPA has been dominated by the attempt to balance continued increases in federal expenditures and the need to finance new obligations to sub- national governments. Oil revenue has increased but in2006 and the first half o f 2007 have been significantly below budgetedplans. Public expenditures have grown sharply since the adjustment o f the mid 1 9 9 0 ~initially usingthe fiscal space from new oil revenues and more ~ recently outpacing revenue growth with more rapid increases in expenditure across most categories o f spending. Significant fiscal deficits (over 2 percent o f GDP) have emerged after nearly a decade o f prudence, including budget surpluses from 2002-04. At over 20 percent o f GDP, spending has returned to levels that characterized the 1980s when Sudan also ranup large external debts. 3.2 This chapter highlights the main features o f GNU expenditure management. From the outset it is important to recognize that a detailed assessment of Sudan's public expenditure planningand management i s hampered by four significant deficiencies: First, current GNU budgetreporting does not provide a classification o f expenditures according to their function and purpose, which i s required to enable adequate analysis and evaluation o f aggregate fiscal discipline, allocative efficiency and operational efficiency. * Sudan traditionally reports budget data along economic lines, and the data are insufficiently disaggregated to monitor spending geographically, across time (duringthe budgetcycle), or across sectors and programs. Second, the country lacks a consolidated budget covering all four levels o f government-GNU, GOSS,states, and localities. Budget data covered inthis chapter is generally GNU reporting provided by MOFNE that covers federal recurrent and capital spending and transfers to lower levels o f government (GOSS and Northern states). It i s difficult to clearly identify the amount o f resources dedicated to specific purposes at the local, state, and federal levels. This poses management issues, particularly in the case o f development spending-for example, in the health sector, investments in health infrastructure require strong coordination between the federal and state ministries o f health. This coordination can be undermined if expenditure responsibilities are unclear, or the distribution o f resources is inconsistent with expenditure responsibilities-and it weakens accountability o f each entity for ultimate outcomes. Third, GNU expenditure data currently only captures MOFNE releases to spending units, rather than actual use by spendingunits. This effectively precludes analysis of expenditure tracking and efficiency. 8Under successive Staff Monitored Programs and dedicatedtechnical support, the IMF is supporting MOFNE inworking toward implementation ofthe GFSbudgetclassification. Progressto date hasbeenslow. 19 Fourth, there i s limited attention to budget execution, relative to presenting the preparedbudget, and evidence is scarce about the outcomes o f expenditures. 3.3 Aggregate GNU expenditure has exhibited strong growth since 2000, increasing by 23 percent on average per annum in real terms, and nearly doubledto 22 percent of GDP by 2006 (Figure 3.1). Total per-capita expenditure inUS dollars grew from $44 in2000 to $233 in 2006. Fiscal discipline was maintainedbetween 2000-04 with most years having an overall balance surplus (on a cash basis). Deficits o f 2 and 4 percent o f GDP, respectively, were observed in2005 and 2006. Per-capita expenditure inU S dollars Per-capita expenditure as percent o f GDP 250 I 25 I 1 1 20 15 6 10 0 2000 2001 2002 2003 2004 2005 2006 2000 2001 2002 2003 2004 2005 2006 +GNU +Northern * Federal Gov. +GNU +Federal Gov. States --Wt GOSS +Northern States --ItGOSS I Source: MOFNE and staffestimates. A. TRENDS GNUSPENDING COMPOSITION IN 3.4 Since 2006, Sudan's national budget i s presented to the National Assembly in three parts-Part Iincludesfederalrecurrentandcapitalexpenditure,PartI1coverstransferstothe GOSS, and Part I11covers transfers to Northern state governments. Oil revenue transfers to the GOSS, comprising over 95 percent of Part 11, are dictated by the CPA. Before the CPA and establishment o f the GOSS in2005, the budget only included expenditures inthe current Parts Iand 111. 3.5 The dominant effect of the CPA on the compositionof public finance has been largeincreases innew obligationsto sub-nationalgovernments, consistent with the CPA emphasis on greater equity in wealth sharing. The federal share o f GNU expenditure has fallen from 92 percent in 2000 to 64 percent in 2006, in light o f CPA obligations to the GOSS and the sharp growth in transfers to Northern states (Figure 3.2). However, federal expenditure over the period also rose significantly, from 11 percent o f GDP in 2000 to 18 percent in 2004, before falling back to around 15 percent o f GDP. GNU transfers to sub- national levels were modest over 2000 - 2004 (on average around 1 percent o f GDP) and subsequently expanded after the CPA. This increase in transfers was largely driven by earmarked transfers, with the bulk for payment o f civil servants salaries. Part o f this reported 20 increase was due to a reclassification o f previously federal expenditures to the Northern states, specifically the recurrent costs for judiciary, higher education and police which are now reported under Part 111. Figure3.2: Growing obligations to sub-nationalgovernments,in percent oftotal GNU expenditures 2000 2006 - Federal Gov I N.States GOSS 1 N .States Source: MOFNE and staff estimates. 3.6 At the end of 2006, GNU transfers to the South were broadly in line with CPA commitments, notwithstanding the boundary uncertainties and sometimes significant monthly revenue shortfalls. Oil revenues dominate the GOSS budget. GNU per capita transfers to the GOSS totaled $22 in 2005 (2.8 percent o f GDP), which was a partial year given the formation o f the GOSS late inthe year, and $37 in2006 (3.6 percent o f GDP). 3.7 The economic budget classification reported by GNU only covers federal expenditures, or 64 percent of total GNU spending in 2006. There are three economic "chapters". Strict rules establish the priority for release o f funds to these categories-federal wages and salaries have first priority, general operation and maintenance ranks second', and the development budget has the lowest priority. Both current and development expenditures have increased, with relative growth in development spending (Figure 3.3). Chapter Two maintainthe largest share o f GNU federal spending. 9Chapter Two contains: (i) centralized items (Le., obligation towards internal and external debt, contributionto the international and regional organizations, general reserve, capacity buildingplan etc.,) which accounted for 26 percent o f Chapter Two in2006; (ii) and Services; and, (5)social subsidies. Goods 21 Figure3.3: Federalexpenditures by economic classification, 2000 2006 - 500 400 BillionSD 300 200 100 0 2000 2 w 1 2002- 2003 2004 2005 2006 -0-WagesandSalaries Chapter Two ...A ..Development Source: MOFNEand staff estimates. B. FEDERAL CURRENTEXPENDITURES 3.8 Current expenditures are largely for goods and services, which accounts for the largest share of GNU federal spending. Within this chapter, there is a significant "reserve" component that receives inflows during the year (i.e,, beyond budget plans). Wages and salaries jumped in 2006, including a 20 percent wage increase for all public employees in April 2006. However, analysis o f the wage bill to inform overall budget prioritization and possible reallocation is hampered by the lack o f data on the number o f civil servants, pay at different grade scales, and the prevalence o f allowances in civil servants' compensation packages. 3.9 With the lack o f regularly reported functional data, the size o f the military budget is not well understood. Indications are that defense spending declined sharply at the end o f the NortWSouth conflict in 2005, to 1.2 percent o f GDP in 2005 from roughly 2.3 percent of GDP in 2001-2003 (Table 3.1). Subsequently, defense spending has reportedly increased sharply in 2006 to its highest level this decade, nearly 30 percent o f all federal current expenditures or 2.9 percent o f GDP. This increase is consistent with the authorities' reporting o f new spending on demobilization efforts and support of the Joint Forces since the CPA. Defense spendinghas also shown increases inother post-conflict situations, insupport o f DDR programs. The World Bank i s working with MOFNE in order to collect more information to confirm this. 22 Table3.1: Defense SpendingRises in 2006 Item 2000 2001 2002 2003 2004 2005 2006 in billion SD Total Federal Government Expenditure 322.5 384.2 466.8 674.3 1010.7 1028.3 1186.5 Total Current Expenditure 270.3 317.5 344.1 501.0 709.2 730.5 814.7 Defense and National Security 76.4 79.5 92.8 109.2 106.9 80.0 235.5 Public order and safety 23.7 26.6 36.3 52.9 58.8 24.2 90.1 inpercent o ftotal federal government expenditure Defense and National Security 23.7 20.7 19.9 16.2 10.6 7.8 19.8 Public order and safety 7.4 6.9 7.8 7.8 5.8 2.4 7.6 Other federal government expenditure 69.0 72.4 72.4 76.0 83.6 89.9 72.6 inpercent oftotal current expenditure Defense and National Security 28.3 25.0 27.0 21.8 15.1 10.9 28.9 Public order and safety 8.8 8.4 10.5 10.6 8.3 3.3 11.1 Source: MOFNE and staff estimates. c. FEDERAL DEVELOPMENTEXPENDITURES 3.10 Public investmentto facilitate growth and poverty reduction i s paramount to Sudan's reconstruction and development challenge. The JAM highlighted the acute need for rebuilding the country's deteriorated infrastructure and service delivery framework, and the GNU has subsequently devoted an increasing share of its budget to public investment. The 2007 budget outlines over $1.5 billion innational development projects. 3.11 In practice, the increase in spending on national development projects has been volatile and heavily concentrated in several large projects, with spending outside of budget allocations common for these projects. Aggregate spending on development projects rose from 1.3 percent of GDP in 2000 to 4.0 percent in 2004, then slowed-down to 2.9 percent on average over 2005-2006. This increase has not been smooth, with funding actually declining in nominal terms for two years during this period. The investment portfolio i s heavily concentrated in a few large projects, with the largest five projects comprising 60 percent of total investments in 2006. The massive Merowe Dam project has dominated the portfolio with roughly one-third o f project investment in 2006 (see the irrigation share in Table 3.2). Transport and energy make up almost another third o f recent expenditures. The share o f project disbursements inthe social development sector increased sharply in 2006, with two new largely foreign-financed projects - rehiring military retirees (11.2 SDD billion) and the national program for good governance (7.0 SDD billion). 23 Table 3.2: Sector shares of nationaldevelopment projects(percent) -- - 112005-06 Sector 2000 -- -2001 2002 2005 Irrigation 18.3 14.7 7.3 44.9 Agriculture 16.5 5.0 2.2 5.4 ................................... Livestock ............................................................................................................ 0.4 ........................... 0.5 0.3 0.3 Industry 17.9 17.2 12.2 10.4 4.0 Transportation, Roads& Bridges 12.6 7.1 3.3 8.2 9.3 18.7 15.9 Energy 14.9 18.5 16.7 15.9 17.0 17.4 16.0 16.7 Social Development 7.6 12.9 1.2 5.2 5.8 7.5 13.5 10.5 Water 5.1 2.7 0.9 1.2 3.2 0.3 Others* 6.8 -- - 21.3 55.9 22.9 16.0 1.5 1.5 Source: MOFNE and staffestimates. 3.12 The concentration o f the development portfolio in a few large projects is associated with large year-to-year changes and weak financing performance inthe remaining portfolio. Table 3.3 shows the sharp contrast in execution rates o f the portfolio, taking account o f the five largest projects. Spending beyond budget allocations is common for the largest projects. For example, the Merowe Damproject spent more than $125 million more than budgetedin 2006, due to higher than forecast resettlement costs, which crowded out part o f the rest of the portfolio. Many projects slated for funding inthe approved budgettend to receive little or no financing, especially insocial development, water and livestock. Volatility o f the portfolio i s demonstrated inTable 3.4 with a highdegree o f "churning." New projects make up between 25 and 43 percent o f the portfolio, and there are frequent examples o f projects that are not financed, dropped and re-enter the portfolio. Table3.3: Budgetedand Actual Expenditureson NationalDevelopmentProjects Year Budget Total Actual Total Actual Total Executionrate (%) (SDDbill.) (SDDbill.) (% of GDP) Total ITop 5 projects IExcludingtop 5 I I 2000 47.8 36.9 1.3 77.3 160.6 47.4 2001 99.2 56.7 1.6 57.2 80.1 45.7 2002 143.9 115.5 2.9 80.2 131.0 71.4 2003 138.8 107.4 2.3 77.4 166.8 56.4 2004 220.7 221.0 4.0 100.1 125.4 83.5 2005 173.9 185.9 2.6 106.9 135.7 80.0 2006 I( 263.9 I 256.1 3.1 1197.1 I144.5 I65.5 24 Table 3.4: ExcessiveMovements in the Portfolio of National Development Projects Source: MOFNEand staff estimates. *** Projectsnot in the portfolio of the previous fiscal year. Projectsnot in the portfolio of the previous fiscal year, but inan earlier portfolio. *** Projectsnot inthe portfolio ofthe current fiscal year, but inthe previous portfolio. 3.13 There are serious development planning constraints in Sudan that require focused attention by the authorities. Constraints are faced at the program planning and project analysis levels, and both at the national and state levels. A comprehensive public investment program is lacking. There appears to be limited analytical consideration of the appropriate public investment envelope and trade-offs o f interventions among and within sectors. Linkages across levels o f government, especially with respect to supporting service delivery on the ground, are weak. Efforts to improve upstream budget planning activities, particularly the macroeconomic framework and revenue forecasting, may provide the best short term gains. The following areas are starting points for improvements: (i)sector planning needsto look at both current and capital expenditures; (ii) allocations needto linkto policy anchors, such as the on-going Five Year National Strategic Plan; (iii)concrete estimates o f policy actions are required; and (iv) predictability in execution is a necessary incentive for investment inthe planningprocess. 3.14 At the project level, basic information on project objectives and components is hard to find. A recent World Bank mission was unable to obtain any evidence o f minimal cost- benefit analysis and evidence o f adequate project appraisal for a sample o f large public investment projects. Project monitoring during implementation and evaluation o f the achievement o f intended benefits also appears very low, both in financial and economic terms. External assistance, from the World Bank and others, would help to build on this diagnostic and assist in the capacity building o f development planning and project preparation. 3.15 In addition to funding for investmentprojects, development spending also includes several large line items for public capital contributions, agriculture subsidies, and on-lending to development financing institutions. Information on these activities is also limited and hampers an examination the effectiveness o f GNU expenditures and short and long term 25 benefits. Some informationon agriculturesubsidies is emerging (Box 3.l), a consolidated but sector approachis neededthat links bothcurrent and developmentexpendituresto objectives. Box 3.1: Agriculture Subsidies and Support under the Green Mobilization Program In addition to the portfolio of national development projects, GNU development spending has several large line items for public capital contributions, agriculture subsidies, and on-lending to development institutions. Emerging information on the agriculture subsidies is available in the context o f the Green Mobilization Program (GMP), started in 2006 and focusing on improved productivity and livelihoods for small-scale farmers. The GMP i s consideredmostly pro-poor. The Green Mobilization Program(GMP) started in2006 and focuses on improved productivityand livelihoods for small-scale farmers. Most of the activities are considered pro-poor. In 2006 - 2007, agriculture season support budgetedunder the GMP was SD 33.2 billion-r 8.6 percent of the national development budget for agriculture in the years 2006 and 2007 combined-but execution was only 44 percent through May 2007. There can be little doubt that the agricultural sector is currently under severe stress for a number of reasons including export price pressures from recent appreciationof the real exchange rate and low productivity, and temporary assistancecould be welljustified insome circumstances. Financefor Irrigated Cotton Production(4.1 billion SDD, 83 percent executed) The `incentive' payment for Acala medium staple irrigated cotton was intended to encourage farmers to producethis medium staple cotton which accounts for the bulk of the world market and has traditionally been Sudan's major output. Subsidizing the price of Acala was meant to discourage substitution production of Barakat long staple cotton. A shortfall in Acala production would have resultedin a loss of market share for Sudan, and the international market for Barakat is extremely limited. This incentive program for the 2006/2007 crop represented a 28 percent increase in the price paid to farmers in the field comparedwith the typical price that the SCCL had paid in earlier years before the appreciation of the exchange rate and lower pricesinthe international cotton market. At the subsidizedprice the SCCL was able tojust avoid a loss. Other Programsin Irrigated Schemes The GMP program allocated and filly executed SD 1.1 billion in subsidies to compensate for the increased costs of diesel fuel for turbines pumping water for irrigationschemes where water supplies are not gravity fed. Another program providing a SD 7 billion subsidy targeted for electricity tariffs associated with the use of turbines for irrigation has not yet beenused. Ninety percent of allocated funds (SD 2 billion) have beenused under a separate program for capital andrunning costs ofcleaningboreholesand subsidiarycanals inirrigation systems. Ninety percent of a subsidy of SD 5 billion for fmancing horticultural production, typically in irrigation areas, has beenused. Finance for Improved Seeds (3.9 billion SDD, 195 percentexecuted) This program has achieved notable success in achieving its pro-poor goals of raising productivity of small- scale farmers in the traditional rain-fed sector. This is a continuation of an existing government program [initiated in2003) which was adoptedby the GMP to scaleup earlier success. Finance for Traditional Rainfed FarmingAreas 3ther potential `pro-poor' subsidies within the GMP program include SD 3 billion allocated for rain-fed igriculture. It i s unclear what form this subsidy might take but in any case none of the SD 3 billion allocated #as used in2006/2007. In addition, SD 5 billion allocated for the financing of animal resources has not been sed althoughits purposes are not clear. .I6 The lack of functional information and analysis of expenditures is a major shortcoming. Analysis of expenditures by purpose or sector is valuable for planning, monitoring and linking expendituresto outputs and development outcomes-in other words, to determine whether spendingby relevant agencies is generating the expected results. For 26 example, a public expenditure review would typically attempt to correlate changes in the level and distribution o f spending at the sector level, and increases or decreases in relevant outputs and outcomes. It would answer questions such as: 0 Inputs: How much was spent on primary education? What was the intra-sectoral breakdown (how much was spent on wages versus textbooks versus infrastructure, and so on)? Outputs: What were the tangible outputs? (How many teachers were paid, textbooks were distributed, schools were built, etc.?) Outcomes: What was achieved? (How many children completed primary education? What share are girls? Over time, what i s happening to the literacy rate?) 3.17 More sophisticated and in-depthstudies could look more closely at value for money within the sector, since international evidence shows that higher levels o f spending do not necessarily lead to better results. However, sector expenditure analysis in Sudan has been limited by both the types o f classification used and a lack o f a consolidated budget. An attempt was made to overcome the first problem by mapping MOFNE releases by spending unit to a functional template. While not strictly conforming to GFS methodology, the data should be reasonably robust in looking at broad trends in the fimctional use o f federal expenditures. 3.18 The preliminary functional data on GNU expenditures shows the increase intransfers to sub-national levels o f government has been largely financed by a reduced share o f general services (Figure 3.4). The shares o f economic affairs, social affairs and reserves have also decreased, though the latter remains large by international standards. Besides the increase in transfers, social protection also increased. This can be traced to hugejumps inthe wage and salary expenditure o f the Ministry o f Social Care and Development, from less than 0.1 percent o f total federal wages and salaries between 2000 and 2004 to 30 percent in2005 and 40 percent in2006. The PER did not examine the destination o f this spending. The current- development mix shows that two-thirds o f current expenditure for general public services and social affairs, while nearly three-quarters of economic affairs expenditure is development. Figure3.4: FunctionalCompositionof GNUSpending (percent shares) 1 40 ~ % 25 20 "6 r) 5 0 G Pub Econ Social Social Others not Loans Rep Conti Transfers Services Affairs Affairs Prot class Reserve 0 Average2000 2004 - IAverage2005 2006 - Source: MOFNE and staff estimates. 27 3.19 Working together with MOFNE staff, an attempt was made to supplement the functional data on federal transfers with new data on state expenditure covering GNU transfers as well as expenditures financed by own-revenues. The result is a more comprehensive measure of expendituresthat allows more complete sectoral analysis, given the divided responsibilities across government levels. But it is important to emphasize the data limitations. The states' annual total expenditure was estimated by using the individual final accounts annual report for all the northern states for the period 2000-2005. These reports were reviewed by the Chamber of Accounts federal representatives in the state, but given the overall weak state o f PFM systems in Sudan, especially at lower levels of government, the findings can only be taken as indicative. 3.20 By way o f examples, Table 3.5 summarizes a comprehensive picture o f federal and Northern states' s ending on health, which has grown sharply but was still only 1.4 percent o f GDP in2005.l` The bulk occurs at the state level, consistent with the fact that the primary responsibility for basic service delivery lies with state governments. Thus the growth is drivenby rising federal transfers to Northern states, though the growth inhealth spendinghas beenlower (39 percent annual growth in2005 relative to 155 percent for total statetransfers). A major concern is the extremely low level o f investment spending inthe health sector, with over 95 percent going for current expenditure. In 2005, development spending rose to 9.3 percent o f health expenditure, but from negligible levels inboth absolute and relative terms. However, assessing the appropriate balance between recurrent and capital expenditures will require more detailed sector-specific analysis than is currently available. The recent North Kordofan Basic Services Study demonstrates some early work inthis direction (see Box 3.2), and is available inthe PERbackgroundvolume (see footnote 1) to promote further studies of the outcomes o fpublic expenditure and sub-national service delivery. 10Among countries with 2004 per capita GDP less than $2,500 the expectedlevel of per capita public spending on health by a country of GDP per capita of $1,000 would be around 3 percent of GDP, although with wide variation. 28 Table 3.5: Federal and Northern States Health Spending," 2000-2005 2000 2001 2002 2003 2004 2005 SDD billions Total Health Expenditure 23.2 32.5 37.9 49.1 75.2 98.4 I- spending Current 21.7 30.5 37.2 47.7 71.7 88.4 II-Development spending 1.5 2.0 0.7 1.4 3.5 10.1 I-FederalExpenditure 7.1 9.0 8.8 11.7 20.8 19.4 Current spending 6.0 7.3 8.7 11.3 18.8 16.7 Development spending 1.1 1.7 0.1 0.4 2.0 2.7 11- State Expenditure 16.1 23.5 29.1 37.4 54.3 79.0 Current spending 15.7 23.2 28.5 36.4 52.9 71.7 Development spending 0.4 0.3 0.6 1.o 1.5 7.3 Breakdown of TOTAL health expenditure (YO) Current spending 93.5 93.7 98.1 97.2 95.4 89.8 Development spending 6.5 6.3 1.9 2.8 4.6 10.2 Breakdown of FEDERAL health expenditure ( O h ) Current spending 84.1 80.7 98.9 96.5 90.4 86.0 Development spending 15.9 19.3 1.1 3.5 9.6 14.0 Breakdown of NORTHERNSTATES health expenditure (YO) Current spending 97.7 98.7 97.9 97.4 97.3 90.7 Development spending 2.3 1.3 2.1 2.6 2.7 9.3 Source: State FinalAccounts Annual Reports,MOFNE. Note: total healthexpendituresexcludethe GOSS. Box 3.2: Insightson Sub-nationalHealth Spending The Federal and State Ministries o f Health andthe World Bank undertook a study in2006 o fthe financing of basic health services at the local level inNorth Kordofan. Financial data were collected from the State and Locality administrations, a facility survey measured service quality and utilization indicators, and focus groups revealed patient perceptions. The main messages are as follows. Most state/locality public spending on health is for salaries, with negligible investment spending. Health services in the state, including primary health care and hospital services, are largely the responsibility o f the State and Localities. Total estimated government spendingon these services in2005 was around SDD 2,400 million, or $10 million. This representedabout SDD 1,500 per capita, or $6.20. Of this, about $4.00 was spent by the State government and USD 2.20 by Localities, mostly on salaries financed by transfers from the State. It i s estimated that 60 percent o f public spendingon health i s for salaries. Investment spending i s negligible. This health spending, although steady as a proportion o f total government spending, is considerably increased from previous years in line with the overall growth in public spending, linked to increased Federal transfers to the state. In 2003, for example, State government health spending was SDD 700 million, about halfthe 2005 figure. Higher spending is correlated with higher levels of service utilization. Estimatedpublic spending onnon- hospital services varies considerably by locality and is well-correlated with utilization rates. Without accounting for possible confounding factors such as differing levels o f socio-economic status between localities, utilization increases by 0.05 patients per capita annually with each increase in spending on non- hospital services of SDD 10. Utilization is similarly well-correlated with numbers o f health workers, and this 11Dueto data limitations, the calculations do not include foreign finance, direct federal expenditure on medical equipment, expenditures from revenue collected directly by health facilities, and other expenditures not captured inthe MOFNEbudget figures. 29 i s the case both with higher and lower-skilled workers. This can be seen, for example, inUmRwabalocality, where overall utilization is relatively high and facilities are largely staffed by community health workers. Similarly, utilization of public sector non-hospital facilities in rural areas, mostly staffed by lower-level cadres, is higher than inurbanareas. Spending is concentrated on urban health services in hospitals. About two-thirds of government health spending in the state is allocated to the 14 hospitals, concentrated in urban areas, in particular the State capital. Seventy percent of spending on hospital services is for salaries. Non-hospital basic services receive only about a third of public spending on health in the state, and of this about 40 percent is for staff remuneration. Facilities rely on cost recovery for significant financing. Officialuser fees are a significantrevenuesource for the financing of basic health services, representingabout a fifth of total expenditures. In addition, many fees retained by health facilities are not included in revenue reports, but certainly represent a significant source of financing, likely covering most non-salaryrecurrent costs. By policy, most drug costs are borne by patients. As would be expected, higher fees are charged in higher-level facilities staffed by more skilled personnel. The reported cost to patientsof treatment for an episode of malaria rangedfrom SDD 500 ($2.20) inprimary healthcare (PHC) unitsto SDD 870 ($3.80) inhospitals. ... thus the rural poor have less access to services. Fees for services represent a significant barrier to access to care, particularly inrural areas. Health facilities reported that 25 percent of patients fail to pay for services inrural areas, compared to only 5 percent inurban areas. O f those who failed to pay inrural areas, only 1percent was subsequentlysupportedby a social support mechanism, comparedto 62 percent inurban areas. Focus groups generally indicatedthe fees for services were higherthan ability to pay andthat this was an important barrier to access. It could be expectedthat higher public spending would be accompaniedby lower user fees, which inturn would leadto higher utilization, but such apattern is not evident. Overall, this study sketches a picture of an under-financed but moderately-performing public sector heath system that privileges hospital services in urban areas. Most public financing goes to salaries and higher financing (and greater numbers of health workers) are reflected by higher service utilization. A large proportion of total resources are concentratedin Sheikan Locality, which contains the main referral hospital in the State capital, but there is also evidence that lower-level services in rural areas and some Localities function well enoughto attract moderatelevels ofutilization. The main policy implications of this study are that further increases in government health spending(largely tied to increases in Federal transfers) should: (i) be better allocated to non-hospital basic services in rural areas and under-served Localities; (ii) improve spending on non-salary recurrent costs and investment; and (iii)betiedtoreductionsinthefinancialburdenonhouseholds,whichimpliesbothreducinguserfeesand startingto subsidizedrug costs. w c e : North Kordofan Basic Services Study (2007), included inPERbackgroundvolume, Chapter 6. 3.21 Table 3.6 shows the federal and Northern state allocations on education. The trend and composition i s broadly similar to health, albeit with more modest growth and slightly higher state and recurrent shares. Total education spending was 1.3 percent o f GDP in 2005.'* Over the period 2000-2005, education averaged 4 percent of federal and 23 percent o f state expenditure. As in the health sector, increases in state-level spending have been almost entirely absorbed by salaries, however, as mentioned with health spending, more detailed sector work is required to assess the right recurrentldevelopment balance. Capital ''Among countries with per capita GDP less than $2,500 (in2004) the expectedlevel of per capita government spendingon education by a country o f GDP per capita of $1,000 would be around 4 percent of GDP, although with wide variation. 30 investment (Le., school construction) i s possibly necessary but certainly not sufficient to improve education services. Improved teacher remuneration, as well as public financing of other recurrent costs (particularly textbooks) will be crucial to reduce the financial burdenon households among the poor and in under-served rural areas. In states and cities where enrolment i s relatively high, investment inquality improvement would appear necessary. Table 3.6: Federal and Northern States Education Spending, 2000-2005 2000 2001 2002 2003 2004 2005 Total EducationExpenditure 32.7 37.3 53.8 58.9 91.4 88.5 I- spending Current 32.0 36.3 44.8 57.3 85.2 80.7 11-Development spending 0.7 1.o 9.1 1.7 6.3 7.9 Federal Expenditure 14.3 17.0 21.7 29.1 40.7 21.6 Current spending 13.8 16.2 21.6 27.8 36.3 17.0 Development spending 0.5 0.8 0.1 1.3 4.4 4.6 StateExpenditure 18.4 20.3 32.1 29.8 50.8 66.9 Currentspending 18.2 20.1 23.1 29.4 48.9 63.6 Development spending 0.2 0.2 9.0 0.4 1.9 3.3 Breakdownof TOTAL education expenditure ( O h ) Current spending 97.9 97.2 83.2 97.1 93.1 91.1 Development spending 2.1 2.8 16.8 2.9 6.9 8.9 Breakdown of FEDERAL educationexpenditure (YO) Current spending 96.6 95.0 99.8 95.4 89.2 78.9 Development spending 3.4 5.0 0.2 4.6 10.8 21.1 Breakdownof NORTHERN STATES education expenditure(YO) Current spending 99.0 99.1 72.0 98.8 96.3 95.0 Development spending 1.o 0.9 28.0 1.2 3.7 5.0 Source: State Final Accounts Annual Reports, MOFNE. Note: total health expendituresexclude the GOSS. 31 D. BUDGETPROCESSAND CREDIBILITY 3.22 The GNU has an establishedbudget planningprocess inthe second half o f the fiscal year (July-December). The specific steps are outlined in Table 3.7. The steps generally occur on time and in a transparent manner, with broad representation in the various committees and participation o f the Northern states. 13 The time period for budget formulation i s compact, especially for coordination with sub-national levels. The current practice o f contemporaneous preparation of central and state budgets in the North does not give the states a reliable estimate of central transfers intime for state budgetformulation. I Formulation o f macroeconomic committees 2"dto 4" week of July Preparation o f macroeconomic framework 31dand 4" weeks of July Submission o f the budget to the National Assembly EndNovember to early December Amroval of budget bv the National Assemblv IIMidDecember 3.23 While the budget preparation process is orderly, the budget has credibility concerns stemming from systemic deficiencies in the upstream and downstream budget processes. Efforts to improve upstream budget planning activities, particularly the macroeconomic framework, may provide the best short term gains. As mentioned earlier, the following areas are key: (i)sector planning needs to look at both current and capital expenditures; (ii) allocations need to link to policy anchors, such as the on-going Five Year Plan; (iii) concrete estimates o f policy actions are required; and (iv) predictability in execution i s a necessary incentive for investment inthe planningprocess. A longer term issue i s the need to ensure all revenues are collected in the NRF. It i s unclear whether the full extent o f fee collection by lineministries is reflectedinGNUrevenues. 3.24 A second source of budget credibility concerns is execution performance (Table 3.8). Expenditure deviations in annual aggregates from approved budget have tended to improve 13Southern Sudan has an autonomous budget process, which is discussed inthe Section 7. 32 over time, though 2006 experienced a low 87.5 percent performance due to oil revenue shortfalls. However, monthly variation o f MOFNE releases is a significant issue. Without a predictable flow o f resources to spendingunits, execution o f spending plans is hampered and o f particular concern are the effects on financing o f capital expenditure. Roughly one-third o f foreign disbursements for national development projects for 2006 were reported as released in the month o f December. Nearly 20 percent o f 2006 state development transfers also occurred in December. Thus while the GNU has increased its budget commitments to pro-poor spendingvia fiscal decentralization, execution inaggregate terms was low for 2006 and the timing o f releases erratic. This has a devastating effect for state budgets which rely on central transfers for a significant share o f revenues. Overly optimistic revenueprojections and expenditures across budget categories having little or no relationship to the amounts budgeted. It i s therefore not uncommon for there to be revenue shortfalls o f around 30-40 percent on an annual basis along with similar variability on the expenditure side. Table3.8: GNUBudgetExecution(actualexpenditures as percentof budget) Items 2000 2001 2002 2003 2004 2005 2006 TotalPublic Expenditure 96.5 90.7 86.6 105.4 102.3 98.8 87.5 Wages and Salaries 104.1 96.2 96.5 96.0 95.5 86.1 94.6 Goods and Services 95.3 103.1 77.5 121.4 108.2 124.5 83.4 Capital Expenditure 79.1 65.2 88.7 99.6 106.7 109.3 96.9 o/w National Development 77.3 57.2 80.2 77.4 100.1 106.9 97.1 Intergovernmental Transfers 118;2 89.5 89.6 85.9 85.0 81.2 82.2 o/w GOSS transfers 118.2 89.5 89.6 85.9 85.0 95.5 75.2 o/w N.State transfers 69.4 92.6 Source: MOFNE and staff estimates. E. FINANCIAL MANAGEMENT 3.25 The authorities have made a number of efforts to improve financial management and public expenditure. Efforts by the authorities to improve financial management systems and public expenditure include restructuring the budget to reflect sub-national transfers, a computerizing budget performance reporting system, training on functional budget classification, and establishment o f a high level committee to help ensure allocation and monitoring o f federal government resources in line with budget priorities. Nonetheless, a significant number o f key deficiencies remain. 3.26 Key aspects of financial management are being tackled by the on-going Country Integrated Fiduciary Assessment (CIFA). Inthe North, there is wide variability between the performance of the system at the federal level and in the states.14Notwithstanding the aforementioned budget and accounting classification issues, there i s regular and effective monitoring o f expenditures at the federal level. However control systems are weak. The Auditor General's report for 2004 states that internal controls systems are particularly weak at the federal government resulting in practices such as complete or partial offshoring o f revenues or depositing them in creditors accounts and then spending them and imposing of 14An assessment of PFM issues in Southern Sudan is covered inSection 7. 33 unapproved fees and regularly issuing o f unofficial receipts. At the State and locality levels financial management capacity i s far weaker, which i s a particular concern giventhe increase intransfers to Northern states. Muchgreater attention is required on the ability to effectively use these resources. (See section on state case studies.) 3.27 There is a consistent and large build up o f arrears in the system as the commitment control system i s not effective. Expenditures are committed on the basis o f appropriations, rather than availability o f cash resources. Cash management is weak, and for other than salary payments, there i s a large amount o f unpredictability regarding resource flows at the federal, state and locality level. Due to this, there i s a regular build up o f arrears. Information regarding commitments or arrears i s neither collectednor analyzed. 3.28 Regarding fiscal reporting, accounting i s the responsibility o f the Chamber of Accounts (CoA). The CoA has around 8000 staff posted in various spending units. Reconciliation o f banking and fiscal records i s undertaken satisfactorily on a monthly basis. Inaccordance with the 1995 regulations, each spendingunit reconciles its accounts with the bank statement and submits the same to the CoA by the 15th o f the following month. The CoA in turn uses this information to compile the monthly account for the Government as a whole. Monthly government finance data i s produced on time, with a six week lag. The utility ofthis data is undermineddue to shortcomings inthe classification system. 3.29 Annual public accounts are produced on time both at the Federal Government level as well as in the States. This is a relatively recent development. There has been a marked improvement in closing o f accounts since 2002 when all 16 states in the North closed their accounts on time. However, the Auditor General's report for 2004 records that even though States did manage to close their accounts on time, this was through not fully recording either revenues or expenditures inthe final accounts. Accounts are audited in a timely manner and submitted to the legislature. However there are widespread capacity constraints at both the Federallevel and inthe states. The quality o f auditingneedsto be improved to bringit inline with international standards. 3.30 Insummary, the mainagendafor the 2008 budgetincludes: Accelerating reforms to address the basic information deficiencies that undermine effective public expenditure management. Two key dimensions are consolidated budget planning and monitoring using the GFS functional budget classification, and building capacity for fiscal management and reporting. The latter is on-going, including through World Bank support, and will be further guided by the on-going CIFA. 0 Rationalize the current expansionary fiscal policy, in light o f revenue mobilization difficulties. This should include focusing on using existing expenditure levels more efficiently, especially national development expenditures and state-level expenditures. A much greater focus on accountability and the outcomes o f public spending is required. Development planning and execution should be a priority inthis regard. Improve budget credibility through: stronger and more transparent linkages during budget preparation (e.g., via the macroeconomic framework) to sector and thematic 34 policies, including poverty reduction; more transparent expenditure prioritization and predictable budget execution, including checks on the significant spending beyond budgeted allocations (e.g., operation and maintenance, and large development projects); and monitoring and reporting of arrears. 35 36 4, TOWARDA PRO-POORPUBLIC BUDGET 4.1 In a post-conflict, decentralized Sudan, equitable and transparent budget allocations (how much is spent and on what) are key. Progress on this front will support successful implementation o fthe CPA, broad-based growth, and poverty reduction and sustained human development towards the Millennium Development Goals (MDGs). Improved fiscal decentralization with state and local governments fulfilling their service delivery responsibilities i s thus critical. These are clear motivations for measuring and monitoring pro-poor spending, as the Sudanese people expect public expenditures to address the underlyingstructural causes o f conflict and underdevelopment. Additional fiscal effort is not the only constraint to better outcomes, but recent studies point to significant underfundingas a critical obstacle to progre~s.'~ And the gap is large, given the weak national record on human development outcomes, especially relative to income per capita. While per capita GDP nears $1,000 or roughly 25 percent higher than the SSA average, recent estimates place Sudan's health outcomes on par with countries near $400 per capita GDP.16 Raid economic growth i s relatively recent, and programs and policies will affect outcomes with a lag. At the same time, a concerted effort is needed to ensure this opportunity i s used to put the country on a path for significantly improved pro-poor outcomes. A. A PROXY FORPOVERTY REDUCINGEXPENDITURES 4.2 There are several challengesin Sudan that make international good practices in monitoring pro-poor spending difficult to apply. A conceptual understandingo f pro-poor spending in the relevant development literature l7is that pro-poor spending refers to "spending that benefits thepoor more than the non-poor; spending that actually reaches the poor; and spending expected to have an impact on the welfare of the poor over time". Identifyingpro-poor spending in Sudani s constrained by: the quality and classification ofbudgetdata; 0 the decentralized responsibility for basic service delivery; 0 the lack ofdata on outcomes and a missingculture of accountability; and 0 the lack of a coherent set o f policies that can anchor a definition o f poverty-reducing allocations (e.g., via a national poverty reduction strategy). 4.3 Recognizing these constraints and building on the JAM, the PER exercise derived a feasible definition of GNU pro-poor spending in Sudan. Over the past two years better information on public expenditures has become available, including via PER I 5See section 6 of this volume, as well as Darfur: Dimensions of Challengesfor Development - A Background Volume (World Bank, 2007). l6 North Kordofan Basic Services Study, 2007, included inthe PER backgroundvolume (see footnote 1). l7 See for example, PRSP Sourcebook, (2004, Klugman editor) or Tim Williamson and Sudharshan Canagarajah (2003) "Is There a Place for Virtual Poverty Funds in Pro-Poor Public Spending Reform? Lessons from Uganda's PAF." 37 state case studies, allowing for an improvement in identification o f pro-poor categories relative to the 2005 JAM. The PER working definition o fpro-poor expenditure for the GNU includes: selected federal recurrent expenditures-recurrent spending on basic health care; primary education; water supply and sanitation; parts o f agriculture, irrigation and livestock; rural infrastructure (roads and bridges); energy and electricity; and the social subsidy; selected federal development expenditure-local financing o f pro-poor projects identified based on broad international experience (in lieu o f beneficiary information); this includes national development projects inbasic education, primary health, and drinking water and sanitation, and most basic infrastructure; and selected transfers to Northern states-three-quarters o f current and block transfers (based on observed expenditures at the state level via the state case studies (see Section 6), plus all local financed development transfers, plus all transfers to the three Areas. 4.4 While recognizing this definition is only an approximationbased on newly available data, further improvements are expected during the process o f PRS preparation and the National Strategic Plan. However we also recognize that very little is known on the actual use of the funds classified as pro-poor, which i s a gap to be filled inthe future. 4.5 During the PER dialogue, MOFNE took ownership of defining and measuring pro-poor spending. Identification o f pro-poor spending is o f course highly context- dependent, and i s most useful when country-driven, developed through a wide participatory discussion, and used as a means to help guide and track government efforts. The PER exercise motivated the government to start developing its own definition o f pro-poor spending. While differing from the PER working definition described above, a MOFNE working definition i s an important step to formalizing monitoring in budget reporting and will hopefully be incorporated in the 2008 budget preparation process. The MOFNE definition i s a work inprogress and currently includes: from federal recurrent expenditures-all recurrent spending except wages/salaries/ pensions for ministers and politicians, and all expenditure on goods and services, excluding defense; 0 from federal development expenditures-all development spending, both locally and foreign financed; and from transfers to Northern states-80 percent o f current and block transfers to northern states, plus all regional development projects both locally and foreign financed. B. APPLYING THE DEFINITION 4.6 GNU pro-poor spending has increased significantly. The pre-CPA period (2000- 2004) exhibits a low level and flat trend o f pro-poor spending averaging roughly $16 per capita (19.3 percent of total GNU expenditures and 2.8 percent o f GDP). Since the CPA, GNU poverty-reducing spending increased significantly to $68 per capita in 2006 (24.8 38 percent o f total GNU expenditures and 5.5 percent o f GDP). This represents a commendable increase, but i s below budget plans (6.6 percent o f GDP), JAM commitments (5.9 percent of GDP), the average for HIPC countries (7 percent o f GDP), and, for example, neighboring Ethiopia (19 percent o f GDP) and Uganda (11percent o f GDP). 4.7 As noted above, we don't have information directly from the spending units on the use o f funds for the intended purposes and the quality/efficiency o f spending. This is especially true at the state level. Moreover, the growth inpro-poor spendinghas occurred in the context of rapidly expandingtotal expenditures, and did not involve reallocation inlevel terms from other spending categories (Le., all categories experienced real growth). It will be important to protect and strengthen pro-poor spending as a policy reflected in the budget, especially given fiscal pressures. The formulation o f a national poverty reduction strategy should help inthis regard. 4.8 Composition of Pro-Poor Spending. To determine the underlying factors behindthe aggregate trends, three aspects are analyzed: 0 By level of government. The increasing pro-poor efforts over the last two years are largely explained by higher transfers to Northern states, which account for more than half o f total pro-poor spending in2006. The J A M called for a large majority of GNU pro-poor budget efforts in the form o f transfers to the Northern states, to support decentralized delivery o f basic services. The increase in pro-poor transfers was, however, less than the government committed to in the JAM. This feeds into an overall trend in pro-poor spending which i s skewed toward the federal level, suggesting under-funding o f poverty-reducing expenditure at the state, and local levels, where basic services are financed and delivered. 0 By economic budget classification. Most of the federal share of pro-poor spending i s development spending, while state pro-poor spending has increasingly been allocated to recurrent expenditure (two-thirds in 2006). The dominance o f recurrent pro-poor spendingover development spendingat the state level raises concerns, since improving service delivery to the poor in the longer term requires investment in schools, clinics, roads and so on. The relative emphasis on recurrent spending leaves little room for items that have a direct impact on the quality o f services (e.g. investments infacilities, equipment and capacity development). 0 By functional budget classification. Sectoral expenditureanalysis inSudanhas been constrained by both the economic classification o f the budget and the lack o f a consolidated budget. An attempt was made at overcoming these challenges by using the available functional dataset as well as new data on state expenditure covering GNUtransfers as well as expenditures financed by own-revenues. 4.9 The preliminary functional data show the bulk o fpro-poor spending at the federal level i s devoted to economic affairs: 0 Agriculture and livestock has beenthe largest line item on average, although year- to-year allocations have fluctuated significantly. 39 0 Total infrastructure spending (roads and bridges, energy and electricity, and water) has represented about a fourth o f federal pro-poor spending, on average, as have social subsidies. Health, education and water shares remain stable at 8 percent, 4 percent and 0.3 percent o f total federal pro-poor spending. This reflects the fact that the primary responsibility for health education and water lies with the state governments, and does not therefore constitute a significant share of federal poverty reducing spending. Nonetheless the state case studies found that state spending on pro-poor facilities remains very low, especially given the eroded state of infrastructure in the country, and remains way below what i s needed to meet the MDGs, a situation which needs to be addressedthrough improvements inresource reallocations. 4.10 State-level allocations to expected pro-poor activities were mostly for general public services, reflecting support for wages and salaries: 0 There is a significant "General Public" category, which absorbs about 30 percent o f the budget (although it decreasedto about one-fourth in2005). 0 Expenditure to support the agriculture sector has beenrelatively low (only about 5 percent). 0 State infrastructure spending (roads, bridges, electricity, and water) has been slightly lower than at the federal level as a share o f total expenditure, at about 20 percent o fthe total budget. 4.11 Pro-poor spending on health and education. Unlike the federal level, where allocations to the health and education sectors constituted less than 15 percent of pro-poor spending over the period 2000-2006, state-level expenditure on health and education amount to about half the state budgets (one-fourth each). This i s consistent with the fact that the primaryresponsibility for basic service delivery lies with state governments. A key concern i s the extremely low level o f investment spending in the health and education sectors, with over 95 percent going for current expenditure. Improved service delivery to the poor inthe longer term will require larger investment in health infrastructure. In 2005, development spending rose to 9.3 percent o f health expenditure and to 4.9 percent of education expenditure, but from negligible levels and in both absolute and relative terms remains very low. 4.12 The Three Areas-South Kordofan, Blue Nile and Abyei-have been accorded special status inthe CPA, in recognition of war-affected status and challenges. However, in contrast to the increase in overall transfers to Northern states, financing for development directed to the Three Areas remains very low, especially given the eroded infrastructure and weak public financial management inthese areas. 40 c. INCREASING THE LEVEL AND QUALITY OF PRO-POOR SPENDING 4.13 The PER exercise has identified three entry points for increasing the level and quality of pro-poor spending: Improving planning and analytical underpinning. Strategic pro-poor planning appears to be weak at the national level. Lack o f integration o f public investment planningwith the annual budget allocations process is a key problem inthis context. Low capacity inplanningand prioritizing pro-poor spending can in turn be traced to weak coordination within federal level (MOFNE and line ministries), and between federal and state level in preparing and executing the poverty reducing budget, in particular information sharing (e.g. budget and execution data). The PRS, and the ongoing Five Year National Strategic Plan, can be used to match expenditures to specific poverty-reducing efforts and establishholster the monitoring and evaluation systemfor public expenditures and promote a results culture. 0 Enhancing budget credibility. Budget credibility, as defined by the degree to which actual expenditures deviate from budgeted levels, i s important for effective implementation o f government's poverty reductionprograms and developmentefforts more generally: Pro-poor budget credibility is relatively low, compared to execution o f non- pro-poor expenditures. This i s likely to remain a challenge in the current fiscal environment. In 2006, the execution rate for pro-poor spending deteriorated by 17 percent points. In this context, we can observe lower likelihood to protect poverty related sectors during resource envelope shortfalls, with development expenditures the hardest hit area. 0 Weak budget credibility is even more acute at the state level. State transfers deviated on average deviation by 18.4 percent. The annual execution for development poverty related transfers to states was only 71 percent. This pattern has serious repercussions for the effectiveness o f the budget as an instrument to address regional disparities and the success of fiscal decentralization inthe country. 0 There is a disappointing track record on budget credibility on federal allocations on basic social services, especially education and health. Weak predictability strongly limits the ability o f line ministries to plan and implement policies and programs. This underlines the urgency of improving the budget processes at both the federal and sub-national levels through clarifying the roles o f the different levels o f government in resource managementand strengtheningthe involvement o f line ministries, Parliaments and broader stakeholders inthe budgetprocess. 0 Reallocating from non-pro-poor spending. The current fiscal pressures in Sudan underscore the importance o f increasing expenditure efficiency and effective use o f scarce resources. The growth in pro-poor spending has occurred in the context o f rapidly expanding total expenditures, which have allowed increases without reallocation. Government revenue leapt from $1.3 billion in 2000 to $6.9 billion in 2006, and an estimated at $9.1 billion in the 2007 budget. However, there has been 41 only a weak effort to shift expenditure priorities o f government during this period. Non-poverty reducing expenditure dominate, at 85 percent o f total federal expenditures. Federal spending on pro-poor sectors increased over the same period, but inabsolute terms by muchless-from $0.137 billion to $0.364 billion. Within an overall prioritization of GNU expenditures, the authorities need to look for opportunities to increase pro-poor spendingthrough reallocations. 4.14 Further review o f the composition o f federal non-pro-poor spending in 2000-2006 shows: 0 Expenditures on goods and services, especially on "centralized items,"'* account on average for 46 percent o f the federal total non-pro-poor budget; Non-pro-poor spendingon wages and salaries, including the defence and security wage bill as well as chapter one reserves, account on average for 35 percent o f the federal total non-pro-poor budget; defence spending had declined significantly in 2005, but jumped to 72 percent o f non-pro-poor wages and salaries in2006; and 0 Non-pro-poor development spendingaveraged around 19 percent o f total federal non-pro-poor spending. 4.15 Next Steps. Increasing the level and quality of pro-poor efforts will require a concerted effort inthe near term, and insummary, the reform agenda includes: Development of MOFNE's definition, and monitoring, of pro-poor expenditures. Ownership of this process is necessary for meaningfulpro-poor expenditure management. Inthis regard, the precision o f the definition is less important than its use in budget planning and monitoring, and further engagement with the authorities will aim for a reasonable MOFNE definition that is presentedinthe 2008 budget and subsequently monitored. Two specific areas for possible adjustment to the definition presented inthe PERmay involve including foreign financing o f national development projects deemed to be pro- poor as well as significant components o f the Merowe Damproject assuming a detailed breakdown o f activities supports their focus on the poor as primary beneficiaries. Data constraints. Improved underlyinganalytical base-it i s hard to identify pro-poor spending without knowing where the poor are and what the most basic needs are. The key analytical gaps include poverty analysis, house budget survey, development planning, sector approach, national prioritization, etc. There is also need for improved budget data, including a comprehensiveness coverage o f budget data (consolidated budget); disaggregated budget classifications (GFS); skills in the statistics and PFM areas; timeliness, Centralized items include domestic and external debt repayments and debt services; delegations and conferences; subscription inregional and international organizations; custom duties; and government travel and hospitality. 42 availability and readability o f budget information; feedback loops from poverty analysis to budgeting; and monitoring o f program budgets and other results- oriented policy-making tools to encourage evidence-based budgeting. Budget processes. Improve the budget processes at the Federal and sub- national levels, starting with the 2008 budget process, including stronger poverty-motivation in the budget, strategic thinking at the national and especially sector planning, and a culture o f accountability to use o f resources. This might be through clarifying the roles o f the different levels o f government in resource management; and fostering the involvement o f line ministries, Parliaments and private sector in budget process. The PRSP process should contribute significantly, if adequately linked to the budget (e.g., monitoring o f pro-poor expenditures in the budget and spending). Targeted efforts are needed to improve the budget credibility through concerting efforts to lower budget deviations; and improve revenue estimation. State budgets need reliable estimates from the center on transfers, which are the single most important factor behind severe credibility problems at the sub-national level. Decentralization. Since responsibility for basic services devolved to the state and locality levels, this is the frontline for addressing poverty. As more resources are getting to the sub-national level, it is imperative to gather information at the state/locality level on the use o f funds, in a manner, for example, similar to the North Kordofan Basic Services Study (PER background volume, Chapter 6; see footnote 1). At the same time, financial management capacity building efforts must be strengthened. Concentration on development spending. There i s an urgent need to direct resources to development at the state level, where it will directly benefit the poor and marginalized areas. Making development planning more effective through better coordination between the federal and state levels; information sharing (e.g., budget and execution data); and developing guidelines for more streamlined procedures at the state level. 43 44 5. INTERGOVERNMENTALTRANSFERSAND FISCAL DECENTRALIZATION 5.1 With the recent peace agreements and the emphasis on equitable sharing of national wealth and resources, the issue o f intergovernmental fiscal relations i s central to the policy agenda in Sudan. The roles and responsibilities o f sub-national governments, and associated public resources, have grown significantly in the past few years. In this context, the GNU has presented a vision for a more equitable and transparent system o f intergovernmental resource allocation across its northern States, and well as between the north and south o f Sudan. This vision is anchored in the 2005 Comprehensive Peace Agreement (CPA) and InterimNational Constitution(INC). 5.2 This section summarizes the motivation for fiscal decentralization, reviews the current state o f intergovernmental fiscal arrangements and presents a forward-looking set o f challenges and policy options to inform the reform dialogue. The focus i s on the Northern states, with the nascent structures o f decentralization in the South and its unique challenges addressed in Section 7. This section also provides the backdrop for Section 6, which looks inter alia at the linkages between deficiencies o f the intergovernmental transfer system and state-level budgets and fiscal management. A. MOTIVATION 5.3 The CPA and INC define the vision of fiscal decentralization in Sudan, and underscore its vital importance for successful implementation of wealth and power sharing agreements during the Interim period. The vision has beeninitiated with a range o f objectives, including post-conflict stabilization and the improved access and delivery o f public services for a broader range o f beneficiaries across all o f Sudan's regions. Marginalization and neglect o f the periphery and the dominance o f the center have been central to the various conflicts in Sudan. Wide cross-regional variations in development outcomes, emphasized in Section 1 (and in more detail in the PER background volume, Chapter 2 on pro-poor spending; see footnote l), reflect the large differences inthe allocation of resources across states and regions. Fiscal decentralization is expected to redistribute resources to match public service delivery responsibilities, and thus improve inter-regional equity, responsiveness in public services to local level demand, sub-national autonomy and accountability and regional economic development more generally. It also promises to help ameliorate continued and potential conflicts such as in Darfwr and Eastern Sudan by supporting the near term peace dividendthrough poverty reduction and economic growth. 5.4 The primary component of successful fiscal decentralizationis a more equitable and transparent system of intergovernmentalresource allocation between the North and South o f Sudan, as well as across and within its Northern states. The specific objective o f a good transfer system i s to address vertical imbalances between the center and sub-national levels o f government with respect to revenues and responsibilities (as expenditure 45 responsibility for basic services shifted to state and local levels), as well as horizontal (inter- state) imbalances due to differing own-revenue potential and differing needs. 5.5 International experience suggests that it i s important to evaluate the impact o f an intergovernmental transfer system in terms o f its incentive effects on sub-national governments. These include the overall efficacy o f sub-national public service delivery and accountability, sustainable fiscal policies, and own source revenue mobilization. In post- conflict settings, the promotion o f state legitimacy and national unity tends to represent an especially important ingredient for intergovernmental fiscal design. The analysis discussed below i s a baseline assessment o f ongoing issues in Sudan's various levels o f fiscal decentralization that will hopefully inform strategies and prioritization by Sudan's policy makers, at various levels o f government, and the engagement by international development partners. This understanding will enable the identification of the capacity and technical supportthat is likely to be needed. B. CURRENT STATE OF THE INTERGOVERNMENTAL TRANSFERSTHE NORTH IN 5.6 The intergovernmental structure itself reflects the legacy of conflict and regional grievances in the Sudan. Inclusiveness, in its various aspects, and equitable sharing o f wealth and power thus became central pillars o f the CPA. Figure 5.1 depicts the principal intergovernmental relations and fiscal transfers emerging from the CPA and ensuing legislation. The June 2005 Interim National Constitution (INC) mandates three tiers o f government (federal, state and local) inthe north, with an additional level-the Government of Southern Sudan (G0SS)-in the south. Figure5.1: IntergovernmentalRelationsin Sudan Northern States SouthernStates (15/16) I I Local Local Governments Governments (40-80 Counties,Payams, Bomas) 46 5.7 GNUresources transferred to sub-nationalgovernments have increasedsharply, and Northern states have become increasingly dependent on central transfers. In2006, 36 percent o f GNUexpenditures were fiscal transfers to either GOSS or Northern states, with budgeted levels rising to over 38 percent in 2007 (see Section 3). This represents a very large increase, both in level and share terms. The share o f GNU budget transferred to sub- national levels, also known as the "vertical share," was only 8 percent in 2000 (for Northern states). Data on states' own source revenue collections i s not readily available for all states, but case study evidence from six states suggests that own source revenue shares in 2005 rangedfrom 66 percent inRed Sea to 12 percent inNorth Darfur (Section 6). Major sources o f state own source revenues include non-tax items, such as departmental fees and charges collected by the ministries and localities in addition to state public enterprise profits, as well as taxes such as the business profit tax, property tax, capital gains tax, animal tax, and state stamp duties. 5.8 There is a complex array of revenue and transfer types.'g Table 5.1 summarizes the main types of revenues ofNorthern States, rankedbroadly by the level of autonomy sub- national governments enjoy over these revenues. Own source revenues imply that local governments have a high degree o f autonomy (e.g., rate setting authority) over revenue sources (see Ebel & Yilmaz 2004). In practice, they can also be understood as those that accrue directly to state accounts/treasuries, rather than passing through the national budgetkreasury. Shared revenues are transfers returnedon a derivation basis. Finally, other types o f transfers could be determined on the basis o f a number o f criteria, ranging from formula basedto purely discretionary. Table 5.1: State RevenueSourcesinNorthern Sudan RevenueType Revenue Items Determination of Collection/Allocation Own Source Revenues State land andproperty tax and royalties; service Combination of fiscal baseand ( O W charges for state services; licenses; state personal effort by individual states income tax; levies on tourism; state governmentprojects and national parks; stamp duties; agriculturaltaxes; Potentialbasesprovidedby grants-in-aid and foreign aid; excise duties; border trade Article 193 of the INC charges or levies inaccordance with national legislation; other state taxes, which do not encroach on National or Southern Sudan Governmenttaxes, many other tax as may be determinedby law. Shared revenue (SR) VAT Sharing State share of revenue based on 2 percentof petroleum revenuesby derivation derivation basis (and other criteria), established by CPA Grants and Transfers Historical offset for eliminated states agriculturaltaxes May be determinedby formula, (TR) Current transfers (largely for wages) existing establishment costs (e.g., Development transfers wages), or are in a sense ad hoc. Other transfers Social transfers Emergencytransfers Ad hoc transfers Specialtransfers to the NorthernAreas Additional fee on benzene Borrowing Loans/borrowinginaccordance with the Constitution 19The PERbackgroundvolume contains a fuller descriptionsofthe various types oftransfers (see footnote 1). 47 5.9 Table 5.2 presents fiscal transfers to the Northern states by specific line items, along with execution rates for 2006 (actual over budgeted) and pro-rated developments based on 2007 Q1 data. VAT transfers account for almost a quarter o f budgeted transfers in 2006. While in principle these are allocated on a derivation basis, in practice different horizontal criteria are used (see below), suggesting that these revenues can best be understood as transfers rather than derivation-based shared revenues. The remainder o f transfers are allocated for current transfers, compensations for historically abolished state revenues (agricultural compensation), and development transfers. There are significant differences in execution rates across the different transfer items. While the execution rate for aggregate fiscal transfers in 2006 was 75 percent, rates for capital expenditures (61 percent) were significantly lower than for current transfers. Table 5.2: Fiscal Transfers to Northern States (2006/2007) 2007 Q1 2005 2005 2006 2006 2007 Execution Budget Execution Budget Execution Budget * (SO (% (SO (% (SO (% Billions) Budgeted) Billions) Budgeted) Billions) Budgeted) GNUbudget, part I11 NorthernStates Transfers - 224.8 95.6% 477.8 75.2% 602.3 53.9% a. Current Earmarked Transfers 55.5 90.5% 62.6 84.3% 75.1 76.2% Wages (Chapter I), (Judiciary, Police, Higher Edu.) 47.9 100.0% 48.8 94.5% 60 90.0% Operation(Chapter II), (Judiciary, Police, Higher Edu.) 7.6 30.3% 7.8 32.1% 4.2 28.6% Social Subsidy (Chapter 11) 6 68.3% 10.9 18.3% b. Block Transfers 130.3 86.8% 273.6 80.4% 288.6 80.5% Additional Excise on Benzene 10 Transfers to Oil ProducingNorthern States 1.3 53.8% 5.3 94.3% 4.9 130.6% Current Transfersto Northern States 90 97.2% 124.8 86.2% 134.8 82.8% Develop.Trans. to Nuba Mountains & BlueNile Area 5 0.0% 10 68.0% 12 0.0% Developmentand Construction Transfersto Abie Area 3 0.0% 6 8.3% 6 446.7% National ReconstructionFund for War Affected Area 3 13.3% Agricultural Tax Compensation 18 91.7% 22.5 99.1% 24 46.7% Value Added Tax 105 74.2% 106.9 71.5% c. DevelopmentSubsidy 39 132.1% 141.6 61.2% 169.7 31.3% States DevelopmentProjects(Local Component) 27.2 183.1% 92.1 75.2% 124.7 38.2% States DevelopmentProjects(Foreign Component) 11.8 15.3% 49.5 34.9% 45 11.6% d. Reconstruction Funds - - 69 0.6% Source: MOFNEand staffestimates. 2007 ExecutionRates are based on Q1 data only * 5.10 Horizontal (Le., inter-state) allocation criteria vary across the different types of GNU transfers. While the NSSFdocumentation suggeststhat VAT transfers are made on a derivation basis, the reality appears to be more complex. While a formula based allocation 48 appears to be underlying current transfers, it i s not immediately possible to reproduce these allocations. The aggregate vertical share for overall NSSF reported transfers therefore seems to represent the summation of these various individual transfer pools, rather than some aggregate vertical share that i s consequently allocated on a transparent formula basis across states. The newly established FFAMC has in this context proposed a number o f reforms to the prevailing transfer distributions. Prior to turning to options for implementing these or other comparable reforms, it i s useful to describe the workings o f the present system using the latest available data from 2007. (See Figure 5.2.) Figure5.2: Transfer to the NorthernStates (YOGDP, 2000-2006, and budgeted2007) Northern State Transfers (% GDP) 4% 3% 3% 2% 2% 1% 1% 0% 2000 2001 2002 2003 2004 2005 2006 2007* Current Agr. Compensn Development W V. Added Additional Social Funds Emergency Others Source: NSSFAnnual Reports 5.11 Historically the National State Support Fund (NSSF) has acted both as an intermediary and monitoring institution for state transfers. The NSSF has been mandated to determine the distribution criteria o f VAT, as well as some capital and current expenditures. It annual monitoring reports providing a more detailed review o f allocations made under the various flows. While some o f these flows are first transferred by MOFNE to the NSSF for onward execution, others are transferred directly to the states. Increasingly, however, the reports by the NSSF only encompass part o f the transfers to the states as other direct flows from MNFE for current and capital expenditure allocations have gained prominence. 5.12 The newly established FFAMC has advocated making transfers to the Northern states more rules-based. The CPA and the INC mandate the creation o f a new institution- the Fiscal and Financial Monitoring and Allocation Commission (FFAMC)20-as the entity with key responsibility for implementing the vision o f fiscal decentralization. Recent reports by the commissionpropose transfer levels quite close to those reflected inthe NSSF data (SD 227.5 and 265.7 billion in2006 and 2007, respectively). However, the reports do not provide a clear statement o f how these vertical pools are arrived at, but presumably they are in line with the absolute aggregate amounts determined below. The FFAMC proposes a more *'Seearticle 198 ofthe INC (reproducedinthe PERbackgroundvolume, Annex B; see footnote 1). 49 explicit horizontal allocation across states, based on (i) percent minimum required for 40 government responsibility; (ii)40 percent population; (iii)15 percent development indicators; and (iv) 5 percent fiscal effort by the state.21 5.13 Actual allocations to states are characterized by significant differences. Figure 5.3 reveals large variation across states in the volume o f the transfers per capita for 2006 (using NSSF data). In 2006 there was more than a five times difference between Northern State which received SD 19,573 (US$ 98) per capita and South Darfur that received SD 3,636 (US$ 18) per capita. In 2000 Northern State (the top recipient) received over eight times what was receivedby Gedarif(the lowest ranking recipient). River Nile and BlueNile (along with Northern State) are the highest recipients of transfers on a per capita basis. Prior to turning to a more detailed examination o f the proposed horizontal distribution, we detail the practice of actual reportedtransfer to the statesto-date. Figure5.3: State TransferLevels (Per Capita, 2006) Per Capita Fiscal Transfers (SO) 25,000 20,000 15,000 10,000 5,000 0 VAT rn AgriComp 0 CurrentTr 0 DevTr rn AdditonalTr 5.14 Disparities in per capita allocations do not necessarily imply poor equalization. More careful analysis i s needed to assess the extent of fiscal capacity and expenditure needs o f the individual states (Hofman, Kaiser, and Suharnoko, 2006). Fiscal capacity for example would be reflected by an ability to generate own revenues. However, Figure 5.3 does not even reflect actual revenues currently being collected from the states. Expenditure needs could encompass population, but also other factors such as area, cost o f service delivery, etc. However, primafacie it appears that some states are getting less (others more) than would be justifiedby any equalization criteria. 5.15 Fiscal transfer realizations have varied significantly by state. Using NSSF data for planned and actual fiscal transfers, the annual realization ratios for the total volume o f transfers have varied quite extensively. Figure 5.4 shows the realizations by state for 2005 and 2006. Evidence to suggest any systematic differences in realizations across states is, *'See also"Interview with FFAMC Chairman: The CommissionEnsuresTransparent and Prompt Transfers of Funds from the FederalGovernmentto the States," SudanVision, Sunday, June 24,2007. 50 however, quite weak. Furthermore, rank correlation between realizations in2003 and 2004 i s 0.35 and between2003 and 2005 i s 0.34. However, there i s virtually no correlation between the ranking in 2004 and 2005. Khartoum has remained at the top o f the rankings across the three years whereas West Darfur has been consistently at the bottom. Based on available information, it i s not possible to reach any robust conclusions as to whether differences are a result o f states' performance/capacity or a result o f actions on the part o fthe NSSFand/or the MOFNE. More generally, however, the poor predictability o f fiscal transfers, given state dependence on this source of revenue, will in turn significantly undermine overall budget credibility and financial management at the sub-national level (see Section 6). Figure 5.4: State Transfer RealizationRatios (2005/2006) IcnoL State Transfer RealizationRates (ActuaVRealized) l J V l 0 130% 110% 90% 70% 350% 0% 10% 2005 ' w2006 Source: NSSFAnnualReports c. GOING FORWARD-CHALLENGES AND POLICY OPTIONS 5.16 The design and implementation of intergovernmental finance stands to promote both good public finance as well as peace. In Sudan, political-economic considerations articulated inthe CPA and the DPA makethis even more critical than would ordinarily be the case. The systemmust be seen to be equitable and fair, inparticular insharing oil wealth, and should address the severely uneven development that has resulted from a combination o f inequitable sharing o f resources and prolonged civil strife in certain areas. In Sudan, sustaining the peace (an extremely valuable public good with huge positive externalities) should be viewed as an important objective o f the federal grantdtransfers. This justifies the case for special transferdarrangements for this purpose. 5.17 Three topics addressed below pose the highest challenge to implementation o f a transparent, predictable, and equitable intergovernmental transfer system, with policy options discussed to inform future dialogue: rationalizing the types o f individual transfers; the formula for horizontal distribution o f transfers; and institutional arrangements. Additional areas to support the effective service use o f these transfers by states are added to conclude this section. 51 Rationalizing the types of individual transfers 5.18 By international standards, the current modalities of allocation are quite complex and hence may lack more broad based buy-in by stakeholders, notably at the state Greater transparency and simplicity will help in mitigating the risk o f further conflicts and promote legitimacy. The points below are meant to set out a possible prioritization and subsequent sequencing o f reforms, with the larger goal o f an improved systemover the mediumterm (3-5 years). 5.19 VAT Sharing: It is not clear whether the horizontal allocations for VAT are made on an annual derivation basis, or based on some historic collection estimates. Sharingrevenues on a derivation basis potentially provides a more direct link between revenues transferred, respective fiscal capacity, and associated economic developments in the state. This may provide some incentives for the state to provide pro-growth policies, although the actual impact o f such incentive effects in practice would need to be more closely examined. However, allocating revenues on some measure o f fiscal capacity also results ingreater fiscal disparities across states. To better address incentive effects and resulting fiscal disparities, greater clarity concerning the exact allocation criteria for VAT would be recommended. If VAT is allocated on a sharing basis, this could call for a more explicit and graduated approach on how this fiscal capacity i s considered for other transfers (e.g., current transfers for states such as Khartoum). Inturn, ifVAT transfers are allocated across states on the basis o f some other criteria such as needs, this would alter how these revenues would be considered for other revenue sources (see Section 4). Generally, each revenue transfer should be best understood as part on an intergovernmental fiscal transfer system intended to achieve a range o f objectives. These principles may be weakened if different transfers appear to evolve over time ina more ad hoc fashion. 5.20 Agriculture Compensation: Similar to the VAT, the agriculture compensation transfer was introduced to compensate the states for an abolished tax. However, there i s little clarity as to the current determination o f the vertical pool or whether horizontal distribution o f these allocations i s linked to current agriculture production levels. Over time, it may no longer make sense to distribute these resources based on some measure o f agriculture production. Hence, it makes sense to either introduce new model o f tax sharingor fold these resources into the overall block grant. It i s also not clear whether the rationale for these transfers reflects a higher expenditure need o f these states given their agricultural sector, or it i s being used as a proxy for fiscal capacity. Inthat case, it may make more sense to expect states to tax agriculture directly (without reproducing the poor features o f the abolished taxes). 5.21 Current and Development Transfers: Current transfers appear to be based largely on existing public sector establishment costs and are provided as a compensation. Simply paying states for existing establishments suggests states have limited autonomy over the actual management o f their establishment sizes and the wage and salary structure, and that these grants may be setting poor incentives. The basis for the development transfers is not 22 For example, formula based allocations inIndia and Nigeria typically use far fewer factors in allocations and are thus more easily understood. 52 easily understood. It is not clear whether these provide federal support to ongoing state schemes based on an annual determination or whether these reflect federal schemes that are being implementedin states but are accounted inthis way. Allocating recurrent and capital budgets in parallel processes may also be leading to poor incentives. Policy analysis i s needed to help determine the extent to which these transfers can be made more transparent for a formula basis. 5.22 Specific Conditional Grants: These transfers appear to have declined through 2005, then increased significantly in the 2006 budget (from SD 5.5 to 36.4 billion). One part includes "earmarked" transfers for items that were previously on the national budget. One question i s whether these will be regularized as part o f the standard transfers. Specific grants can play an important role in intergovernmental fiscal systems, for example to addressed externalities or promote specific national programs. However, a series o f non-transparent special grants can be problematic, and individual specific grants should have a clear policy goal. Policy work i s needed to identify the demand and appropriate design for a limited number ofthis type o ftransfer. An improvedformula for horizontal distribution of transfers 5.23 Ongoing policy discussions in Sudan and inthe FFAMC suggest that the Government may be considering moving to a fully formula based transfer for northern states. Such an approach would, in effect, abolish the current transfer windows (e.g., VAT, current, development, other), and consolidate these in a single pool for further rule-based allocation. For the purpose o f discussion, three purely illustrative simulations and benchmarks are presented, which allocate the overall vertical poll (Le., SD 248 billion realized in2006): (i) based on population as a broad proxy for need; (ii)based on a mix o f needs proxies, Population(50 percent), Area (20 percent), Primary Child Enrolment (20 percent), Ruralness (10 percent); and finally (iii)allocating part on fiscal capacity/derivation basis (VAT+Ag transfers), with the remainder using the needs formula used for (iii). different formulas The simply represent different scenarios on possible horizontal distributions.23 International experience suggests that a range o f transfer designs are used in practice (cf Boadway and Shah 2006). Some examples such as Indonesia's general equalization block grant incorporate both measures o f expenditure needs, and proxies for fiscal capacity (for own source revenue mobilization and sharedrevenues). 23These simulations do not address the related question of the adequacy of the overall vertical pool, given prevailing expenditure assignments for states. 53 Figure5.5: Simulationsof AlternativeTransfer Formulasto States(2007) cSimulated4ActualHorizontal Allocation (%) g200% 180% =I 160% 2 2 140% 120% 100% 80% 0 60% < 9 40% ij 20% Needs(Pop) W Needs(Various) 0 FiscCap+Needs 5.24 Figure 5.5 presents the ratio o f transfers simulated in accordance with the threes scenarios described above to the current transfers. Under each o f these scenarios, Khartoum, Red Sea, South Darfir and West Darfur are currently under-resourced and would be significant gainers from changes in this direction. Clear losers would include River Nile, BlueNile, and Kassala. These results should, however, be treated with caution, as we have only partial information as to the own source fiscal capacity o f the various states, and what would happen if a full-fledged fiscal gap formula i s introduced. Some of these results may well be reversed to some extent as result that is more starkly achieved by the current approach o f not assigning any current transfers to Khartoum and Red Sea. This raises the question o f whether some o f the proxies for expenditure needs fail to capture special needs for these states. In contrast, the fact that states like Khartoum are under-resourced by these measures might be accounted for by the fact that these simulations do not take account o f other fiscal capacity (e.g., property and personal income taxes assigned to the states). 5.25 The simulations highlight the sensitivity of selecting a different formula for states. For example, Northern state would lose less on a needs based formula relative to a purely population based one. The work highlights the need to have a more concerted discussion o f actual state fiscal capacity and expenditure needs (based on functional allocations) to arrive at a preferred set o f formulas for horizontal allocations. These simulations touch on horizontal allocations, but a clearly related question i s whether states on aggregate receive enough (i.e., their vertical share) to carry out their expenditure responsibilities. 5.26 The FFAMC proposed a simplifiedformula for new state allocations, but the precise criteria need to be clarified. The FFAMC proposed a more explicit horizontal allocation across states, based on (i) 40 percent minimumrequired for government responsibility; (ii) 40 percent population; (iii) 15 percent development indicators; and (iv) 5 percent fiscal effort by 54 the state.24 While the numbers for the vertical pool allocation used by FFAMC differ slightly from those o f the NSSF figures (although are also far smaller than those used by the MOFNE), the magnitudes are almost the same. This suggests that the FFAMC has used these numbers to determine its vertical pool. Figure 5.6 simulates the FFAMC allocations against the realized 2006 allocations using the same vertical pool &e., SD 248 billion). This suggests that the main winners o f the FFAMC formula would be the Darfurs and Khartoum. In contrast, Red Sea, Kassala and Blue Nile would be the "loosers" from these revised transfers. The second simulation simply increases the overall vertical pool to the NSSF budget number for 2007 (SD 268 million). While the ratios used are the same, the greater absolute amounts mean that even with the formula changes some states would under this scenario would not loose out inabsolute terms (e.g., Gedarif). Figure5.6: FFAMCProposedFederalTransfers 140% Ratio to 2006 NSSFRealizations 120% 100% 80% 60% 40% 20% 0% FFAMC 2007* HFFAMAC2007I 5.27 The foregoing simulations highlight the sensitivity o f state transfers to various formula designs. The initial simulations tried to make explicit the expenditure and fiscal capacity measures used. These generated quite different allocations to the currently prevailing ones, suggestion that any move to a formula would need to be subject to a phased inintroduction. Finally, the simulations highlight the fact that current allocations appear to deviate significantly from a formula basis. The FFAMC formula has the merit o f being simplerthat that presumably applied byNSSFto-date. However, the documentationdoes not separately state the underlyingdata used for each o f the factors. Greater transparency inthis regard would be useful to specify the most appropriate formula and a phasing in path that would be acceptable to all states. Institutional arrangements 5.28 Another importantelementof intergovernmentalfiscal design is the institutional arrangement by which intergovernmental transfers are determined and made. A numberof agencies and commissions have beenengaged inthe determination, execution, and 24See also "Interview with FFAMC Chairman: The Commission Ensures Transparent and Prompt Transfers of Funds from the FederalGovernmentto the States," SudanVision, Sunday, June 24,2007 55 monitoring o f intergovernmental transfers in Sudan. The foregoing review suggests that overall quantum o f transfers (the "vertical" share) and the allocation ("horizontal") and disbursementto states is decided and implementedby some combination o f the MOFNE, the NSSF and a "Supreme Council" that includes all the state Governors, with a growing role of the FFAMC. The current system o f determining and implementing intergovernmental transfers can be perceived to lack transparency and to perform sub-optimally in terms o f equity and efficiency. 5.29 The roles of the various institutional players such as the MOFNE,NSSF and the FFAMC need to be clarified and understoodby all stakeholders. The Presidential Decree o f 2006 does not fully address this, and inJuly 2007 a new Presidential decree dissolved the previous FFAMC and established a new commission under the same name and general responsibilities, replacing the previous chairman with an ex-head o f the NSSF. This latest development raises questions on the standing o f previous recommendations of the FFAMC, and particularly whether the formula discussed inparagraph 5.26 remains valid for continued monitoring and the 2008 budget. Inthe end, most important for the transfer system i s that the institutions ensure that ultimate allocations conform to good intergovernmental fiscal designprinciples and that their determination and delivery is transparent to all stakeholders. Agreement in the context o f intergovernmental forums in principle provide a means to achieve this, but independent grants commissions can also play an important role in infusing ultimate agreements with good design principles. Inthis context the legislative framework for the FFAMC i s o f critical importance. 5.30 Existing good practices should be retained and enhanced. The NSSFhas a record o f regularly publishing annuals report, which provide information on existing allocations. However, greater transparency i s needed. For example, South Africa's periodic intergovernmental fiscal reviews provide an example o f good practice, and have increasingly been extended to incorporate not only allocations but performance in such areas as service delivery. Mandates o f any new institutions that are established and the revised mandates of existing ones should be clearly spelt out. For example, to be effective, the FFAMC should have legal legitimacy with the legislation clarifying its roles and responsibility, powers and autonomy. Other areasfor effective use of transfers 5.31 Reforms in the design and application o f intergovernmental fiscal systems are only one ingredient to potentially promoting conflict resolution, improved state legitimacy, and broad-based improvements in service delivery. A number o f other reforms will need to enhance mechanisms o f accountability at all levels o f governments. 5.32 Strengthening state level data for monitoring and evaluation, as well as fiscal transfer design. Currently, no comprehensive data is available concerning state own source revenuemobilization. General efforts to strengthencomparable GFS-consistent reporting on the revenue side would be a first step towards addressing this information gap. On the expenditure side, reporting by functional (e.g., across health, education, roads) and economic lines (e.g., wages) would also allow for a more systematic analysis the fiscal position o f states. In turn, improved data on sectoral outcomes would highlight potential disparities 56 across states. At the same time, a few basic and credible indicators for the design o f intergovernmental transfers will need to be updated on a periodic basis.25 5.33 Strengthening sub-national public financial management and accountability. Most o f the discussion has focused on fiscal transfers. However, a critical need will be to strengthen the commensurate horizontal accountability (e.g., to state legislatures) and external accountability (to audit agencies) for sub-national expenditures.26 Capacity building in this regardare needed in the Northern states, and also in particular the South. The on- going CIFA work i s the starting point to engage in more systematic assessments o f even benchmarkingsub-national PFM for all states, and ultimately also local governments. 5.34 Making decentralization work for service delivery. The reform on intergovernmental fiscal relations in Sudan has the potential to promote peace and national cohesion. Ultimately, however, these outcomes will need to ensure that national and sub- national expenditures deliver a peace dividend in terms o f service delivery outcomes throughout Sudan. This will require a host o f reforms, only starting with greater clarity and transparency in allocation and use o f fiscal resources. At the sectoral level (e.g., education, health, basic infrastructure), this will require special attention to sub-national capacity and implementation strategy, as well as role o f line agencieddeconcentrated structures. *'Generally, it is inadvisable to used administrative data reported by the states for this purpose, as linking finance to indicatorsmay create incentivesfor misreporting. 26 Accounting and audit i s set out in Articles 204 and 205. Calls for a National Audit Chamber and a Southern Audit Chamber, both to be further specified by Law. 57 58 6. BUDGETSAND FISCALMANAGEMENT INNORTHERN STATES 6.1 Improving fiscal decentralization and increasing pro-poor spending at the state and local levels i s a key focus o f the current phase o f the PER, which has been informed by Northern state case studies.27 The main motivation o f the sub-national PER work i s fourfold:(i) to have a better understanding o f the budgets inthe Northern states; (ii) have a to better understanding o f pro-poor spending at the national and lower levels o f government, including the linkage o f sub-national transfers; (iii) to identify key constraints to better spending; and (iv) to identify ways to addressthese constraints. 6.2 Case studies were undertaken in six states-namely Gezira, Red Sea, North Kordofan and North, West and South Darfur-covering 2000-2005. In each case, the budget preparation, execution and monitoring inthe state as well as the overall fiscal performance at the state and the locality level were analyzed. Special attention has been placed upon the scrutiny o f the two major components o f the states' budgets: the intergovernmental transfers on the revenue side and the pro-poor component o f the expenditure side. The main messages which emerged are as follows. A. LIMITEDPROGRESS FISCALDECENTRALIZATION ON 6.3 Successful fiscal decentralization depends on sufficiently autonomous and accountable sub-national governments with credible budgets and sound fiscal management practices. The state case studies show that by these criteria, fiscal decentralization remains a work inprogress and requires additional reform and capacity building efforts. These reforms have become more imperative as increased central funding has been delivered to Northern states, which shoulder the bulk o f expenditure responsibility for service delivery to the poor. 6.4 The fiscal and decision-making autonomy of states and localities is limited. This stems from the high degree o f reliance on the resource transfers o f higher levels o f government and the central control over a significant share of these transfers. While central transfers are a general feature o f fiscal decentralization and necessary to effect redistribution to poorer states, an increasing share o f state revenues comes from central transfers, with unpredictable and non-transparent distribution. Figure 6.1 shows the large majority o f state revenues come from the federal government, with the exception o f Red Sea state. Federal transfers account on average for roughly 70 percent o fthe case study states' revenue in2005, and existing transfer practices o f little information on expected transfers and erratic delivery *'Between April and July 2006, draft case study reports were completed for the Red Sea, Gezira, and North Kordofan states covering fiscal trends and budget management. Case studies of the three Darfur states were also conducted in support of the Darfur Joint Assessment Mission during June-July 2006, followingthe same methodology as the PER case studies. Results were discussed in a November 2006 workshop in Khartoum, with representativesfrom all 15 Northern states, Thus assessments for six of the fifteen Northern states have beencompleted,yieldinga representativesample acrossthe range ofstate-levelincomes. 59 prevent longer term planning by the states. The level o f reliance on transfers is greatest for the locality level. InGezira and North Kordofan states, over 80 percent o f locality revenues on average come from the state level. Adding to the lack of autonomy is the high share o f transfers that are earmarked (Figure 6.2). Development transfers, for example, require submission of proposals from the states and approval from the central government. States also mention the influence o f central policies o f wage determination and employment o f management personnel on the state labor bill, which i s the single highest component o f state budgets. Comparison of other states to Khartoum state suggests these influences may be inversely related to the ability to raise own revenues. Figure6.1: Share ofFederalTransfers in State Revenues Red Sea S. Darfur Gezira N.Kordofan W. Darfur I 0 20 40 60 80 loo Percent Source: WorldBank staffestimates inthe individualstate case study reports. - . - _ _ _ - I-.-.--.--.-- I-_ 90 - 80 70 60 I 50 40 30 20 10 0 Red Sea Gezira S. Darfur N.Kordofan N.Darfur Source: World Bank staff estimatesinthe individual state case study reports. 60 6.5 Budget credibility is a major problem. Budget performance, as defined by the degree to which actual revenues and expenditures deviate from budgeted levels, has been poor by international standards for most years in most states. Figure 6.3 shows the extent o f deviations for annual state expenditures. 6.6 The key causes o f poor budget performance appear to be threefold: data limitations, weak capacity to estimate a budget constraint, and political pressure to exaggerate budgets. Key data limitations includethe lack o ftimely and accurate information from higher levels of government on expected transfers, a lack o f coordination between states and central authorities, a lack o f information on state-level economic activity and thus sources o f own- revenue, and poor quality data bases on which to base state and locality tax imposts. Capacity issues include both limitations in carrying out skill-intensive processes (e.g., economic monitoring, project costibenefit analysis, consolidated monthly cash planning, etc.,) as well as technical constraints from lack o f computers and equipment. On the political side, a consistent pattern emerges across the case studies whereby budgets are often used for political signaling rather than fiscal management. Expenditures tend to be estimated on a needs basis, and revenues are inflated to depict a balanced budget, thus serving in effect as a residual. For example, inNorth Kordofan state, locality own revenues were nearly doubled between the initial 2006 locality budget proposal and the final state budget approved by the State Assembly, from 2,807 to 5,000 million dinar. Figure6.3: State Budget Credibility PlannedVs. ActualExpenditures - 10 II:::I 0 + + z I + +A A I 4 r + z X I X m -30 - X 0 X + X !i + + -40 - 0 0 -50 - -60 - 0 -70 ' I 2000 2001 2002 2003 2004 2005 1 +Khartuom I R e d S e a AGezira XN.Kordofm +N.Darfur 0S.Darfur I Source: World Bank staffestimates based on data from state Ministries of Finance. 6.7 The effect o f weak budget formulation shows up quickly during budget execution. Consistent revenue shortfalls force a misalignment between expenditure assignments and resource allocations, and spending is normally rationed early in the fiscal year, with development expenditures the hardest hit area. Federally mandated increases in wages and salaries were the major source o f the noted rapid rise in expenditure and received priority in 61 budget execution, crowding-out other budget allocations, especially for investments in development. 6.8 Financial management at state and locality levels is weak. The type and availability o f budget information showed a poor financial management environment in Northern states. Little information on the effectiveness o f public expenditure i s either produced by the state Ministry o f Finance or expected by the public, which compounds budget credibility problems and may also partly explain limited and falling local tax collection since constituents can't see what comes from taxes they pay. State budget classification i s in economic terms, following the national budget practice, which seriously impedes the transparency of fiscal operations and the ability to plan, execute, and monitor poverty reduction efforts. State officials seem frustrated at not having functional information available. 6.9 The lack o f clear guidelines for financial management and intergovernmental coordination also limits financial management. Three areas stand out in this regard. First, there are significant problems with budget planning and execution with respect to transfers. Between states and localities, precise revenue sharing arrangements require more clarity, for example on what taxes are shared and at what rate. Second, budget authorization processes and payment systems could not be substantiated in the case studies, and cash management procedures were unclear. Third, communication and coordination between levels o f government are limited. Guidelines developed across government levels would enhance development planning and implementation by clarifying roles and responsibilities. State and locality officials queried duringfield visits were often unaware o f items included inthe list o f federally supported on-going projects inthe states provided to the mission by MOFNE. B. SUB-NATIONAL PRO-POOR EXPENDITURES RISING 6.10 An attempt was made in each case study to apply the methodology used to define GNU pro-poor spending to the state level, to understand the composition of spending. Estimates were made in three states, but not for the Darfur states due to data and time limitations. According to case study calculations, per capita pro-poor spending estimates in Red Sea, Gezira, and North Kordofan reveal strong growth over recent years, rising to $38 - $46 in 2005 (Table 6.1). This mirrors the rise in overall expenditures, and confirms that a significant share o f state-level expenditures i s directly poverty-orientated. 6.11 Analyzing the composition o f pro-poor spending shows roughly 75 percent o f state- level pro-poor spendingi s current innature, which by assumption includes all locality current expenditures since functional breakdowns were not available at this level. Pro-poor development spendingvaried significantly by ministryacross years but was largely drivenby increased expenditures on roads, water, and electricity in the Red Sea and education and health inGezira. 62 Table 6.1: Per CapitaPro-Poor Spending SDD Mn us$ 2003 2004 2005 2003 2004 2005 Red Sea Total 5,763 8,366 11,128 22 33 46 Current 4,329 6,503 8,139 17 26 33 Development 1,434 1,863 2,988 6 7 12 N.Kordofan Total 3,399 6,586 9,096 13 26 33 Current 2,448 4,080 6,498 9 16 27 Development 951 2,505 2,598 4 10 11 Gezira Total 4,377 7,222 9,222 17 29 38 Current 3,665 5,786 7,167 14 23 29 DeveloDment 712 1.436 2.054 3 6 8 Source: World Bank staff estimates in the individual state case study reports. 6.12 There is a disappointing track record on development spending in all the case study states. Figure 6.4 shows budgeted development allocations versus outturns for the six states for five years. The 45 degree line represents full budget execution, with planned development allocations equal to actual development spending. Clearly, the execution inthe majority o f state budgetshas resulted indevelopment spending far less than planned. This i s especially acute in the Darfur states, where the situation i s exacerbated by the ongoing conflict. Figure6.4: Planned and ActualDevelopmentSpending X N.Kordofan V N.Darfur 0 1 2 3 4 5 6 7 Budgeted spending, In Bn SD Source: World Bank staff estimates inthe individual state case study reports. 6.13 A consistent message from the case studies was that communication betweenthe state and federal ministries (including functional line ministries)regarding development projects is weak and that guidelines are needed to clarify the specific roles o f each level o f government, interms of planning, execution, and monitoring development projects. Insome cases, state 63 finance officials did not have knowledge o f national development projects under implementation in their own states. This situation highlights the need for a more strategic overall public investment program. C. WEAKFISCALMANAGEMENT CAPACITY 6.14 The limited capacity for fiscal management at the subnational level was the most frequently cited constraint to more effective resource allocation and use. For example, only 11 percent o f the labor force in the public sector in North Kordofan state has above higher secondary school qualifications (technical diploma, bachelors, masters or PhD). Some proponents o f continued central control in key fiscal areas and limited autonomy for the states have argued their case based on existingadministrative weaknesses. However the only way to achieve successful implementation o f the CPA inthe long run i s to build capacities at the subnational level to fulfill roles and responsibilities laid out in the legal framework of fiscal decentralization in Sudan. 6.15 The foregoing shortcomings suggest an agenda for capacity building in Northern states, though a more detailed assessment would be useful (e.g., in the context o f the upcoming Country Integrated Fiduciary Assessment. Immediate needs include capacity building in overall budget processes, revenue estimation and collection procedures, and development planning and execution. However, it i s important to underline that these efforts would be largely wasted inthe event o f failure to move decisively on the reform agenda. 6.16 The poor track record on development spending has gone hand in hand with very limited capacity at the state level to plan and maintain development projects. Most of the planning is centralized at the federal line ministry level, with weak communication to state line ministries. To ensure the satisfactory returns and sustainability o f current increases in public investment, there i s a critical need to involve and strengthen the state and local development institutions. The case study work suggests capacity building should include strengthening state-level project appraisal and management capacities, encouraging and utilizing the existing locality capacities to identify development priorities, and beginningto strengthen fiscal management at the locality level. An improved approach could involve a combination o f the state and local levels driving the identification o f priorities, and in turn driving the funding and strategic management, alongside an active capacity building push that will help sustainability over the mediumterm. 6.17 Financial management capacity is also weak on the revenue side, which in turn contributes to the states' poor own revenue collection efforts. Also, states appear unable to accurately forecast either own or transferred revenues, which help drive budget credibility problems. There i s limited ability to estimate and monitor basic economic activity in the state, which i s necessary for estimating own tax and non-tax revenues. Gezira i s an exception, where state officials were upgradingtheir ability to collect and monitor economic indicators. Some o f the localities visited did have or were starting to make databases for revenues sources and expenditures. However, the process is just beginning and the need for a more reliable system o f recording o f all sources o f tax- and non-tax revenues and inputting them into a computer for documentation and easy retrieval and presentation purposes is widespread. 64 6.18 Development partners can play a significant role in meeting Sudan's capacity building challenge in the Northern states. The recommendations o f the PER subnational work are planned to feed into the Bank-supported Capacity Building and Decentralization Project. Also, the World Bank Institute has designeda program to support training activities on budget management inthe Darfw states, as part o f the DJAM. The six proposed themes to be covered are the project cycle, project management, procurement, local government financial management, intergovernmental fiscal relations, and broader orientation to development concepts to raise awareness o f pro-poor policy options (including participation and gender issues), andprinciples o f good governance. D. KEYNEXT STEPS 6.19 There are important reforms needed to improve pro-poor spending and budget management overall at the state and locality level. Below are specific areas presentedat a November 2006 workshop attended by policymakers at the federal, state, and locality levels.28 It i s recognized that these measures are focused on supply-side accountability, and there is a large gap on the demand-side context, including the specific needs o f the expected beneficiaries and resulting value-added. This was beyond the scope o f an initial round o f case studies, but i s generally quite relevant for the next phase o f PER work. > To build capacity for effective resource allocation and use: Federal level: Expand and deepen federal training problems currently in place for state and locality government officials. State level: Identify priority areas for training and capacity enhancement, including a reviewo fneeds for training activities on budget management. > To improve budget processes: Federal level: 1. Accelerate GFS adoption in the states to enable a consolidated budget. 2. Make the transfer o f fund from the Federal to sub-national levels formula- based, predictability and transparent to recipients. 3. Provide data directly to the States on monthly expected transfers for the remainder o f 2006. 4. Establish procedures for the 2008 annual budget, and beyond, to give States firm estimates of intergovernmental transfers and expected monthly cash flows to establish more credible State and Locality budgets. Estimates should distinguishbetweenblock and earmarked transfers. These procedures may consider suggested changes in budget calendars. State level: 1. Prepare guidelines to clarify revenue and expenditure assignments between the state and locality levels and lower budget deviations. These guidelines should incorporate budgeted levels o f transfers consistent with the abovementioned federal estimates. 2. Improve revenue estimation, including development o f analytical basis and database o f supporting information. The latter may consider longer term instruments such as use o f the national census as an instrument for data **The recommendationsbenefitedfrom a workshoponthe Dark case studiesheldinKhartoumon September 34,2006. 65 collection. 3. Outline steps for states and localities to follow the federal reporting format. 4. Summarize current actions taken to make budget plans publicly available. Develop plans for similar treatment o f budget outturns. 5. Prepare an empirical review of civil service size andcomposition, to clarify the trends over time. >For moreeffectivedevelopmentplanning/management: Federal level: Provide data for each state on development transfers (budgeted and outturns), including source o f financing. State level: States to provide a functional/sectoral breakdown of 2000-2005 expenditures and the 2006 budget, to enable an estimate o f pro-poor spending. Joint: Prepare guidelines for a more streamlined and effective procedure for development planning, including roles o f various parties indevelopment planningand execution (MOFNE, federal line ministries, state MoF and line ministries, and localities). 66 7. EXPENDITUREAND REVENUEPOLICYINSOUTHERN SUDAN: PROGRESSAND CHALLENGES 7.1 The Government of Southern Sudan (GOSS) was established only two years ago, facing the daunting task o f rapidly improving the well-being of the people, and getting the region onto an accelerated development path consistent with its rich endowments. One key advantage o f Southern Sudan i s the large domestic resources flowing from oil production in the region, such that there is muchless aid dependencethanmany post conflict economies- e.g. Afghanistan, Mozambique and Cambodia-provided that public resources are managed appropriately. 7.2 This section provides an overview of the progress and status of budget management inSouthern Sudan. Itbeginsby reviewingthe institutional set-up, thenthe GOSS budgetfor 2006 with particular attention to plans against execution, and analyzing the plans for 2007. The final section presents a series o f key challenges-namely salaries and the payroll, revenue forecasts and predictability and debt. We do not separately address the issue o f management o f the defense budget, nor the role o f budget information and accountability, thoughthe latter is a cross-cutting theme. A. BUILDINGBUDGET INSTITUTIONSFOR GOODGOVERNANCE 7.3 The GOSS has been establishing a system of public financial management virtually from scratch. The former SPLM Secretariat o fFinance, which managed resources o f around $100,000, has transformed itself into a Ministry responsible for managing over one and a half billion dollars annually, including significant MDTF financing. In2005 and 2006, the Ministry of Finance and Economic Planning (MOFEP) was characterized by very weak capacity and undeveloped structures, with Directors being formally appointed only in July 2006. As in most ministries,there were no appointments below the Assistant Director level, where staff are beingpaid, but have not beenformally appointed. 7.4 It is importantto recognizethe evolutionover the past 18 months or so, where at the time o f this writing some major progress could be seen in operationalising key aspects o f the system: 0 InOctober 2006, a Single Treasury System was initiated to enhance accountability of GOSS ministries and units and to exercise control over their spending. GOSS ministries cover one fourth o f the annual operational budget which can be recouped on the basis o f subsequent adjustments and submission o f vouchers. 0 Interim Procurement Regulations were submitted to the SSLA as a part of the budgetary proposals for 2006. These regulations were subsequently approved by the Council o f Ministries inJune 2006. A Procurement Act i s being drafted. 67 InFebruary2007, aprocurement agent (Crown Agents) became operational. 0 A Public Financial Management Act has beendrafted and submitted to the SSLA. 0 The appointment o f anExternalAudit Agent is scheduled inOctober 2007. 7.5 In order to ensure sound public resource management and to guard against corruption and diversion, the MOFEP has made a series of commitments. As noted above, some o f these have been fulfilled, but there is a significant way to go, and the following steps remain priorities: Adopt and implement the public finance bill that i s currently under consideration by the Southern Sudan Legislative Assembly (SSLA); Put in place a transparent procurement system, by applying interim procurement regulations, enacting the new procurement law and properly employing the recently appointed Procurement Agent as well as urgently appointing an External Audit Agent; 0 Work to effectively and fully operationalise the Treasury Single Account as a mediumterm objective for a transparent consolidation and accounting of all revenues and expenditures and for ensuringoversight over operational expense; appropriately utilizing ex ante controls, including the recently established Cash Management Committee; 0 Build on the healthy parliamentary debate around the 2007 budget, and establish appropriate systems and structures for regular reviews o f public expenditure and pro- poor spending by Cabinet, Parliament and the public, with minimal lags inreporting; and; 0 Fully operationalize mechanisms for executive accountability, as per the Southern Sudan Interim Constitution, including the Southern Sudan Audit Chamber and the Anti-Corruption Commission. B. PERFORMANCEOFTHE 2006 BUDGET 7.6 The planned allocation of funds in 2006, as laid out in the budget documents passed by Parliament, was broadly consistent with the developmenttargets laid out in the JAM, even ifthe allocation of funds to the SPLA (around 40 percent of the budget) was larger than initially expected. Sector programs were launched in 2006 for infrastructure, health, education and water and sanitation. This process has been bolsteredby budget sector working groups (BSWGs) which have been functioning since 2006 to assist the Government in sector planning and budget formulation, currently with a three year per~pective.~'With joint participation o f external partners, BSWGs help GOSS define objectives and activities 29There are ten such groups each clustering around related functions; accountability, economic function, education, health, infrastructure, justice, law and order, natural resource and rural development, public administration, security and social and humanitarianaffairs. 68 for individual ministries and are also a forum for training participants inthe budget cycle. In 2007 the functioning o f the BSWGs has been strengthened and enlarged to include participation by the state governments thereby enhancing the comprehensiveness o f the planning exercise and extendingcapacity building inbudget formulation to the states. 7.7 However, there was limited fiscal discipline over much of the year and actual execution bore little relation to the plan. Table 7.1 depicts the overall fiscal position and execution performance in2006. The overall balance in2006 moved sharply into cash deficit due to inadequate spending discipline. Aggregate spending was driven by outlays on wages and operations that were roughly double planned amounts, while capital expenditures were cut sharply. Cash deficit before grants summed up to $430 million, which was augmented by suspense and direct expenditures that await clearance and was financed through significant draw-downs on accumulated reserves. Table 7.1: GOSS OverallBalanceDuring2005-2006, millionUSD . , Itec enuc and grnots 822 1494 1330 I Ye/* ie\enues 822 1.304 I267 97' Y Oil rwcnucs 1300 1266 97"o Transfers 1300 899 69Oo ORSA drawd m ns 327 L'nrec~~nciled receipts, plus adjustnien~ 40 Son.ul rc\tnucs 4 2 sou0 -h.-....tinm-. -I m r grants 181 63 310'0 :mnt .-...r.niw 6 I) a/" I Expenilitures 295 13113 1697 123% Nages 14 253 546 216% >perations 277 296 542 183% >evelopment2 4 835 45 1 54% Ither expenditures 0 158 Suspense expenditures 132 Imprest -1 Approved direct expenditures' 24 I Reserves4 459 367 Expenditurefrnanced by GoSS revenues ' 1377 1539 Expenditures fundedfrom other sources 6 Development expenditure frnanced by MDTF,excluded above 147 63 Notes: 1. Transfers includethe US$53 million receivedby oil producingstates. Budget for donor grants excludenon-MDTF sources, which amounts to US$242 million. Accounting for this budget results to revenue and grants execution of only 77 percent instead of 89 percent. The 2007 GOSS Budget document excluded this amount, thus dropped from the above presentation.2. Development budget are net of MDTF, total amounts to US $982 million. Unreconciled receipts reportedby KPMG, pertainingto the US$34 million difference between the amounts that GNU transferred and received by the GOSS. 3. Direct expenditures that have not yet been approved amounts to $13 million. 4. The amountof reserverundown to finance the deficit is derivedfrom the change inthe endperiodcash balance in 2005 and 2006, net of donor grants. The end periodcash balancefigures were taken from the KPMG 2006 report. 5. Taken from the 2007 Budget document. Source:KF'MG, MOFEP, GoSS Budgetdocument2006 and2007, MOFNE PetroleumUnitReportDecember2007. 7.8 Existing revenue cushions were exhausted in 2006. Withdrawals from the national Oil Revenue Stabilization Account (ORSA) accounted for about a quarter o f the GOSS revenues in 2006, together with virtual depletion o f GOSS's own reserves, compensated for the shortcomings on oil revenue transfers and donor finance. Inaggregate, revenues were 89 percent o fplans. Delays inbringing significant Dar blendproduction on-line as well as poor 69 initial export sales prices lowered oil revenue to the South, only $953 million in2006 against a budgeted amount o f $1.3 billion. Actual transfers from the GNU were even lower, at $846 million, due to arrears in GNU payments ($64 million) as well direct expenditures by GNU. A shortfall indonor grants was another contributing factor to revenue problems, with only a third of expected support realized. The grants came solely from the MDTF. There was no financial support received from non-MDTF sources which was expected to finance about US$242 million3'worth o f development program in South Sudan. 7.9 Aggregate spending was nearly one-quarter above targets, with uneven adjustments in budget priorities. Table 7.2 shows the largest proportionate reductions can be observed for infrastructure (share falling from 20 to 11 percent), education (10 to 4 percent) and health (8 to 4 percent). On the other hand, the share under the broad label o f "accountability" substantially increased-mainly because o f significant over-spending by MOFEP, which i s included inthis category. The agencies accounting for the largest spending were, in rank order, the SPLA, MOFEP and interior affairs, collectively comprising 60 percent o f the outlays. Resources to health and education together accounted for only 7 percent o f total expenditures. Available information about the execution o f capital spending suggests that little was initiated by way o f development projects, and that the bulk o f this spending so far has gone on vehicles for GOSS ministries and agencies. There were also significant operational expenditures. Table 7.2: ExpendituresRelativeto Plans, 2006 USD millionand percent ExecutionRate Budget Actual (YO) Accountability 14 285 1982 FinanceandEconomicPlanning 5 278 5139 Economic Functions 18 21 119 Education 137 104 76 Health 109 63 57 Infrastructure 279 166 60 NaturalResources& RuralDevelopment 59 55 93 Public Administration 26 85 328 RuleofLaw 74 162 218 Security 533 555 104 Ofwhich SPLA 526 552 105 Social& HumanitarianAffairs 13 39 308 UnconditionalTransfers 115 3 3 Reserves Total Ministerial Expenditures 1377 1539 112 Plus:Other items (budget financedby other sourcesandother expendituresincluding suspenseitemsanddirect expenses) 6 158 Total GoSS Expenditures 1383 1697 123 Source: KPMG and MOFEP. 30This figure is takenfrom the 2006 GOSS budgetdocument. 70 7.10 Spendingincreasesin 2006 were driven by outlays on salaries and operations of ministries and public administrators that were larger than planned. The increase in the wage bill reflected the characteristics o f a transition of South Sudan to a post conflict environment, including establishing a government and the popular expectations associated with the provision o f support for members o f the SPLA and the states. It appears that the escalation in the size o f the organized forces31(prison wardens, wildlife, police and war veterans) to about 80 percent o f the civil service i s associated with some de facto demobilization o f the SPLA. The number o f personnel supported by GOSS in the states i s unknown; conditional transfers by the ministries o f agriculture, health, education and legislative assembly to cover personnel compensation accounted for 16 percent o f total wage bill in2006. The purchases o f general office supplies and the cost o ftravel and stafftraining were incurred to carry the operations o f the ministries. The operations inthe states and other levels o f cabinet affairs cost $166 million. The food supply for SPLA also contributed substantially to operational expenditures. 7.11 Weak administrative controls were associated with an expansion in the size of the bureaucracy. The progress in building an effective civil service has been slow and constrained by the lack o f a functioning payroll, shortages o f qualified staff and the absence o f proper procedures for recruitment and incentives. Administrative controls are weak and perpetuate a system whereby appointments, due process o f recruitment and verification o f qualifications are not prerequisite for pay. The lack o f administrative controls in face o f further expansion in the bureaucracy and in existing operations raise the need for urgent control measures inorder to become fiscally sustainable. 7.12 Major wage bill measures exerted an upward pressure on salary expenditures. The adoption o f a higher salary scale during the second half o f 2006 substantially raised wage expenditures. The decision to use an exchange rate o f 250 Sudanese dinar to the US dollar for staff compensation inearly 2007 i s also expected to raise wage costs by about 25% relative to the budgeted amount. 7.13 The slow process of the DDR program and the lack of social security schemes to support war affected persons put pressure upon employment as a safety net for children, orphans, widows, disabled soldiers and demobilized soldiers. However, this could undermine the quality o f the civil service with prevalence o f low-skilled persons. Donor assistance for social security and livelihood schemes could be encouraged to target support to this group to mitigate the fiscal burden o f public employment. 7.14 Bolder initiatives are required to reform the civil service. A comprehensive payroll and payment system for both the civil service and army is an important element inthe assertion o f fiscal discipline. At this point, it appears that the payroll reform requires several key elements, including control over the appointment process, matching o f approved posts to appointed persons, release o f monthly payments o f the correct amount after appropriate 31This excludes SPLA. 71 application o f deductions, and a stoppage o f salary payments on death, retirement or attrition. We are unaware o f any explicit GOSS initiatives inany o fthese areas. 7.15 The overspending on wages and operations appeared to squeeze aggregate capital expenditures in 2006 (Table 7.3). The execution for capital spending has been skewed in favor o f the MOFEP and the ministries o f environment, telecommunication and internal affairs and at the expense o f agriculture, forestry, water resource and irrigation, and road and transport infrastructure. The poor implementation and disbursement o f funds for capital in above-stated sectors might be partly explained by the long duration o f time requiredto preparethe projects and follow procurement regulations. Table 7.3: Execution ofWages, Operationsand Capital, percent of plan, 2006 Salaries Operating Capital Accountability 93 4,503 766 Finance and Economic Planning 78 6,978 9,320 Economic Functions 81 141 130 Education 82 66 71 Health 107 61 50 Infrastructure 27 110 58 Natural Resources& Rural Development 209 109 39 Public Administration 214 535 183 RuleofLaw 221 224 212 Security 360 229 14 SPLA 363 235 14 Social & Humanitarian Affairs 131 189 442 Unconditional Transfers 0 Reserves Total 216 183 54 Source:KPMG and MOFEP. 7.16. The composition of budget expenditures diverged from the stated emphasis on development and pro-poor expenditures. Execution of pro-poor expenditures, here defined as the expenditures by the ministries except by SPLA, MOFNE and Council o f Ministers, ran below plans. The disproportionate spending incurred for the functioning o f rule o f law, public administration, social affairs and economic activity happened alongside serious cuts in the planned support for infrastructure, health, education and rural development. Figure 7.1 illustrates the magnitude o f deviations in actual pro-poor spending from the budget. The spendingby the MOFNE andthe Council o fMinistries onthe other handwas above plans. 7.17. The observed overspending is a byproduct of myriad factors, with lack of capacity for public financial management being key. The situation was characterized by underdeveloped accounting practices, absence o f a well-defined system o f controls, low humanresource and governance capacity andunderdeveloped financial institutions. 72 7.18. The non-oil revenue base in South Sudan is very thin. At the same time, there are non-oil revenue items that are not reported inthe GOSS accounts, such as the non-tax income generated by the ministries for the services they render. For 2007, these ministries are estimated to have a budgeted income of US$17 million. There are other sources o f income that appear off-the budget such as the police registration fee and travel taxes. There are also reportedly taxes collected by the GNU inthe South, 50 percent o f which are due to the GOSS, butthe amounts are unknown. Figure 7.1: Deviations of ActualPro-poor Spendingto Plans, 2006, in percent Propoor Education Health Cooperative and Fisheries Agriculture and Forestry Transport and Roads Housing Lands and Public Utility Water and Irrigation UnconditionalTransfers Other Propoor MOFNE and Council of Ministers SPLA -160 -100 -60 0 60 100 160 200 260 300 Note: Only the unconditional transfers are depicted in the illustration. Other pro-poor expenditures refer to public administration, rule of law, economic function and social and humanitarian affairs. Above expenditures exclude suspense anddirectexpendituresthat haveno functionalclassification. Source: World Bank estimates. 7.19. The significant appreciation of the Sudanese dinar, which combined with the difference in the exchange rate used in valuing revenues and expenditure, effectively raises GOSS spending and lowers income. While the budget expenditure plans were made in dollars, the revenue transfers were in dinars (which then had to be at least notionally reconverted, since the budget had beenplanned in dollars). Since the oil transfer out-turn for the year was only 97 percent ofplans, and spendingwas 7.20. In sum, performance of the 2006 budget underlined the need to exercise greater control over budgetary outlays and to significantlylimit spendingthat was not provided for in the approved budget, and to improve revenue estimation, jointly with the GNU. Revenue predictability and associated issues around oil sector transparency (noted above) pose a major challenge to the GOSS. We have noted that the burgeoning cost o f staff salaries i s a major concern. The identification o f root causes (ghost workers, slow processing of appointments) i s an important step, alongside reforms to preclude its perpetuation, e.g. payroll reform) in order to make serious progress on the public financial management agenda. 73 7.21. Improvinginformation and accountability remains a priority, at several levels: (i) better transparency in oil revenue sharing with the GNU, and more timely information to allow predictability in resources and expenditure planning; and (ii) better, and much more timely, accounting for expenditure, by spendingunit, and the MOFEP. The present lags not only inhibit parliamentary and public scrutiny, but also make sound budget management much more difficult. C. AN ANALYSIS OF2007 GOSSBUDGET 7.22. The GOSS 2007 budget was passed by Parliament in January 2007. On the revenue side, the planned envelope was similar to 2006, at $1.622 billion. Inaddition to oil revenue, the budget was premised on operationalization o f the new revenue authority and non-oil revenue o f $234 million, or about 5.5 percent o f total GOSS income. This i s an ambitious target. The SSLA increased the target for non-oil revenues by some $150 million, without specifically identifying sources. Moreover, since half of customs and excise duty (about one fourth o f the estimated non-oil revenue for 2007) i s shared with the GNU, this would require a huge incremental effort. Given the nascent state o f the revenue authority and the lack ofproper taxation rules andlegislation, and the fact that non-oil revenue in2006 was only $2 million, this will be a difficult task to complete. 7.23. Turning to the expenditure side, the stated priorities for 2007 were outlined by the Minister in his budget speech as follows: (i)address priority development needs, notably agriculture, rural water, education, health and infrastructure development; (ii)control the wage bill and ensure that the payroll and salary scales are adhered to; and (iii) budget enforce discipline and control extra-budgetary expenditures, and eliminate suspense accounts. There will be a significant rise intotal spending, which is projected to be over 10 percentage points higherthanin2006. The breakdownby economic category suggests a large increase insalary costs, which at $865 million, amount to half o f the total budget. This represents more than a tripling from the 2006 budget plan, and a 58 percent increase from the actual outturns in 2006. Development spending i s planned to drop significantly, from almost $1 billion inthe 2006 budget to about $541 million for the coming year, and include a significant foreign financing component. Thus for 2007, the MDTF would account for about 40 percent o f total development spending ($218 million), compared to 15 percent in the 2006 plan and 10 percent inoutturns. 7.24. The 2007 budget involvedseveral major changes inthe spendingpattern relative to 2006. Most notably, spending on public administration and economic functions increase by almost 150 percent, which appears to respond to the over-spending in 2006. Justice and law enforcement, accountability functions, and spending on services for the disadvantaged were also given large spending increases. On the other side o f the spectrum, spending on education and health was significantly reduced, o f the order o f $25 million in each case. Transfers to southern states are plannedto increase by 10 percent, to $125 million. 7.25. The GOSS budget that was adoptedby the SSLA is balanced only because of the projected increase in non-oil revenue. The original budget presented to the SSLA by MOFEP was more realistic, but deliberations resulted in additional funds being allocated across different ministries, and the SSLA directed MOFEP to look at options for increased 74 taxes. This issue raises potential disincentives to private sector investors. The draft Public Financial Management Act aims to prevent Parliament from changing the resources envelope. 7.26. Data on mid-year execution of the 2007 budget shows continued oil revenue shortfalls, with expenditure rationing given reserves and oil savings are exhausted. Dar blend production shortfalls continued to cause oil revenue shortfalls into 2007. Production has since picked up considerably, with the completion o f the dedicated export terminal in Port Sudan. Dar blend has also been getting more favourable prices on the export market toward the middle o f the year. In fact, oil revenue surged from a six month average o f $86 million to $167 million in July 2007 which was well above the budgeted prorata level o f $108 million. Future inflows will depend on market forces, and continued volatility i s expected. The fiscal crunch appears to have been borne so far by capital and operating expenditures. For the first six months the execution rate i s 33 percent for operational expense (excluding block grants to states) and 15 percent for capital expenditures. D. KEY CHALLENGES 7.27. The foregoing has revealed a series o f major challenges for Southern Sudan that are related to the budget. Looking forward, the following are priority areas for further dialogue withthe GOSS authorities. 1. Buildingan effective cash management system requires the setting up of commitment and expenditure controls, a well functioning payment system and sound government banking and institutional arrangements. Strengthening the Treasury Single Account, streamlining andrevising the accounting andpayment processes and the utilization o f freebalance could enhance the accounting and payment system. Government banking arrangements for oil revenue deposits and transfers to states must be studied in greater detail. 2. Reforming the payroll system for SPLA and organized forces, GOSS civil servants and state personnel is key to addressing the excessive outlays on wages. Eliminating ghosts and ineligible employees, enforcing workforce ceilings, adopting affordable and uniform pay and allowance structures and collection o f personal income tax constitute the immediate actions for payroll reform on civil servants. Accelerating the DDR process is integral in managing the size of the payroll. Donor assistance is needed inproviding support packages for retrenched employees. 3. Broadening the fiscal base by increased mobilization o f non-oil revenues reduces the vulnerability to oil volatility. Conservative revenue planning on both oil and non-oil revenues i s prudent. Revenue diversification requires the establishment o f an enhanced Tax/Revenue Department in MOFEP to raise collections from customs, corporate tax and VAT, the modification to the Revenue Act to address overlaps in tax responsibilities across the various levels o f government, its adoption and implementation, the training and computerization and the imposition o f additional taxes or user fees. 75 4. Improved transparency and understanding o f the oil market, industry dynamics and relevant operations o f the oil industry i s required in improving oil revenue management. This includes bridgingthe knowledge gap on the PSA, revenue sharing agreements, and on production and investment plans inoil fields. The development o f an effective National Petroleum Corporation is crucial in the coordination o f GOSS and GNUand thus inimproved governance o foil revenue. 5. Intergovernmental relations between the North and the South and between the GOSS and the states are integral to effective fiscal management and to development achievement. The creation o f a directorate o f intergovernmental relations within MOFEP with a sole task o f monitoring and managing fiscal relations between the GOSS and GONU can facilitate their coordination on revenue planning and execution. Increasing the knowledge on the fiscal situation in the states, particularly on the pattern o f their expenditure, provision of pro-poor services and revenue generation i s a short-run concern. The level o f autonomy o f states and county governments must be ascertained and must be the basis behind their shares to tied and untiedfunding. The distribution of grants mustalso consider the fiscal capacity o fthe states and their relative need for services. 76 ANNEXES 77 L 1 5 r ? i z c r c 7 7-77 -77 7 - 7 -7 7 7 - 7 -7777 7-7-7 - 7 Y m m 4 - 7 E2 t ! 2 ! E S I I i : I 3 00 eL 13 I , I I I YJ ' I ' ' '" ' ' N 0 0 N-j - 9 I II I I ,I I I ,6 m X m3 - - Annex Table B.2: GNURevenueand Expenditure: First Quarter 2007 Performance,in Economic Classification(SDD billions) ITEM BUD. Total Total Total Total 2007 J A N Fab. M a r Q1 p 1,880.6 84.9 108.7 114.5 308.1 1,824.6 84.9 108.2 114.5 307.6 A- Tax Revenues 739.5 38.7 54.6 48.8 142.1 1 Direct Tax - 119.5 I.1 9.2 7.2 23.5 2 -Customs & Excise 380.0 19.7 25.7 28.4 73.7 - 3- VAT. 240.0 12.0 19.8 13.2 44.9 B-Non-Tax Revenue 1,085.1 46.2 53.6 65.7 165.5 1 DepartmentalFees 26.0 1.4 1.4 1.3 4.1 2 National Revenues -- 1,059.1 44.8 52.2 64.4 161.4 Oil Revenues 929.1 44.2 46.7 59.4 150.2 Non Oil Revenues 130.0 0.6 5.5 5.0 11.2 C-Grants 56.0 0.5 0.5 2,358.9 102.6 146.0 125.1 373.7 T 1: N- 1,460.4 61.6 104.3 83.7 249.6 Chapter One 520.8 28.1 36.8 31.6 96.4 A Wages& Salaries 479.5 27.6 33.0 29.9 90.5 B InsuranceBenefits(Pen.&Soc. Insu.) -- 41.3 0.5 3.8 1.7 5.9 ChaDter Two 522.2 16.6 39.8 32.1 88.6 A- Debtservices -- External 82.0 0.2 3.2 7.0 10.4 44.0 0.2 1.8 1.4 3.4 Internal 38.0 1.4 5.6 7.0 B-GMCs/ GICs Repayment 52.5 14.1 14.1 C - Goods & Services 121.0 1.9 9.4 8.0 19.4 D-GeneralReserve 94.5 6.4 1.7 5.5 19.6 E-Other Obligations 93.4 4.9 3.5 2.8 11.1 F -Social Subsidies 78.8 3.2 1.9 8.8 14.0 Chapter Three : Capital Services 23.3 1.1 1.4 1.2 3.7 Chapter four :National Development 394.1 15.8 26.3 18.8 60.9 A- National Development projects -- Foreign 340.8 15.8 26.0 18.5 60.2 Local 207.0 9.6 13.6 13.0 36.1 133.8 6.2 12.4 5.5 24.1 B-StrengtheningofDevelopment FinancingInstitutions 5.0 C- GovernmentCapital Contribution 31.3 0.1 0.1 0.1 0.2 D-Agricultural Season Support 17.0 0.2 0.3 0.5 P A R T 11: GOSS 296.2 16.3 13.4 13.2 42.9 GOSS Oil Share 284.6 15.8 12.1 12.7 40.6 SouthernOil Producing States Share 11.6 0.5 1.3 0.5 2.3 P A R T 111: Transfers To Northern States 602.3 24.7 28.3 28.2 81.2 A Agricultural Taxes Compensations 24.0 1.4 1.6 2.0 5.0 B Transfers to Other Oil Producing States 4.9 0.2 0.9 0.5 1.6 C 134.8 9.4 9.8 8.8 27.9 D DevelopmentTransfersto NubaMountainsandBlue --- Current - Transfers to Northern States Nile Areas 12.0 0.0 0.0 E DevelopmentTransfers to Abeyi Area 6.0 F --- Foreign -- DevelopmentProjects 169.7 4.8 4.4 4.2 13.3 Local 124.7 4.1 3.9 3.9 11.9 45.0 0.6 0.4 0.3 1.3 69.0 0.1 0.0 0.1 G Wages TransferredTo : -ReconstructionFund 60.0 3.6 3.6 6.3 13.5 - Higher -- Interior Judiciary 19.0 1.6 1.7 3.4 6.7 Education 24.0 2.0 2.0 4.0 84 ITEM BUD. Total Total Total Total 2007 JAN Fab. M a r Q1 - Graduates 17.0 1.9 0.9 2.8 H-Goods & ServicesTransferredTo : - 4.2 0.1 0.1 0.3 --Higher Judiciary Interior 2.4 0.1 0.1 0.3 Education 1.8 I-SocialSubsidies 10.9 0.1 0.4 0.5 J- VAT Transfers 106.9 5.3 7.8 5.9 19.1 Source: MOFNEand Bank staffestimates. 85 0 2 d 0 - 0 - 0 - 0 - 0 - 0 8 ?"i \o OQ o m c 0 8 8 8 8 -0 B t z 24 v! c f v; vr % W m N d z zr:2 2 e I- d i 2 2 d 2N 0 c 0 vr d N 0 x N 'E z 2 d c QI d N N i 9 ;2 d c c QI .d 0 0 N 'E c? d 2 x2 x2 i - 09 09 4 2 W c W rr, 0 N 0 'E c? c? d s c c W W i - - t f d 2 vl vl N 0 N - 0 'E c? c? d 2 c W W - i 9 c? d I d m 3 m 0 N 0 'E c c d z d w d d i d 2 2cu 2 N 0 0 0 N 'E c c d x v; vi N N k W E Annex Table B.4: GNUExpenditures: 2007 BudgetinFunctionalClassification(SDD billions) ITEM 2007 Bud. TOTAL GENERALEXPENDITURE 2358.9 a) Federal Government Expenditure 1460.4 GeneralPublic Services 617.0 Executive andLegislative Organs 188.6 Defense andNational Security 311.9 Public order and safety 116.5 Economic Affairs 407.5 Agriculture & Livestock 188.4 Agriculture 175.8 Livestock 12.7 Transport, Roads and Bridges 54.8 Energy and Mining 95.0 Industry and Construction 13.9 Communication 9.9 Other Economic Services 45.4 Labour 45.2 Trade 0.2 Social Affairs 173.6 Health 51.0 Education 42.2 Environmental Protection 0.2 Housing and Community Amenities 26.1 Recreation, Culture and Religion 8.1 SocialProtection - 46.1 Others 262.3 External Loans Repayment 44.0 LocalDept Repayment & Other Repayment 90.5 Reserves/ Contingencies 119.1 Developmentreserves/conthgencies 2.0 Othersnot classified by Item 6.7 b) Intergovernmental Transfers 898.5 Northern States Transfers 602.3 Governmentof SouthernSudan Transfers 296.2 Source: MOFNE and Bank staffestimates. 88 ANNEXC: NORTHERN STATESEXPENDITURES Annex Table C.l: NorthernStates Expenditures: Economic/Functional Classification(SDD billion,2000 2005) - Item 2000 2001 2002 2003 2004 2005 TOTAL GENERALEXPENDITURE 73.6 92.7 119.5 148.8 247.5 313.7 Wages and Salaries, Improvement and Privileges 33.9 44.3 52.5 65.7 110.7 144.7 I-WagesandSalaries 33.8 44.0 52.4 65.2 110.5 143.8 Agriculture Sector 2.3 2.9 3.2 3.8 7.3 9.8 Transport, RoadsandBridges Sector 2.5 2.8 3.0 3.5 5.8 7.2 The Sovereignty Sector 1.0 1.1 1.o 1.8 1.6 1.7 FinancialandEconomic Sector 1.7 2.6 3.4 3.2 7.7 7.3 Health Sector 10.8 17.0 20.8 25.3 38.9 54.8 EducationSector 14.9 16.6 19.5 25.1 43.8 57.7 Administrativeand Social Sector 0.6 1.o 1.5 2.5 5.4 5.3 11-ImprovementandPrivileges 0.0 0.3 0.0 0.4 0.2 0.9 Goods and Services and Centralized Items 33.1 38.4 45.1 45.8 76.2 89.8 I-Goodsand Services 28.4 31.0 35.8 40.9 58.4 65.7 Agriculture Sector 2.7 1.9 2.2 2.3 3.5 3.6 Transport, Roads and Bridges Sector 2.7 3.2 3.3 3.3 6.2 7.3 The Sovereignty Sector 2.5 3.6 4.2 2.9 5.2 7.0 Financialand Economic Sector 9.9 10.4 11.1 13.4 18.6 18.5 HealthSector 5.0 6.2 7.7 11.1 13.9 16.9 EducationSector 3.3 3.5 3.6 4.3 5.1 5.9 Administrativeand Social Sector 2.4 2.3 3.8 3.6 6.0 6.6 11-CentralizedItems 2.2 3.0 4.0 2.4 8.8 12.3 111-CapitalServices 2.5 4.4 5.3 2.4 9.0 11.8 Total of Development Expenditure 6.6 10.0 21.9 37.4 60.6 79.3 Agriculture Sector 0.5 0.2 0.5 0.7 1.7 2.9 Transportation, Roads & Bridges Sector 3.9 5.8 2.1 21.0 33.2 47.9 Energy Sector (Electricity) 0.0 0.1 0.2 0.1 0.4 0.4 Social Development Sector 1.9 3.6 18.5 12.9 8.2 23.5 Capacity building and human resourcesdevelopment projects 0.5 3.0 1.0 11.1 2.2 9.4 Health 0.4 0.3 0.6 1.0 1.5 7.3 Education 0.2 0.2 9.0 0.4 1.9 3.3 Social Care and administrativeSector 0.8 0.1 7.8 0.5 2.7 3.5 General Public Services Sector 0.2 0.1 0.3 1.7 0.6 0.2 Water'sProjectsfinanced through loans and grants 0.0 0.1 0.2 0.2 0.5 1.1 Developmentreserveskontingencies 0.0 0.0 0.0 0.0 0.1 0.1 CentralizedItems 0.0 0.0 0.0 0.1 13.8 2.8 Others 0.0 0.1 0.2 0.7 1.9 0.2 Source: Ministry of FinanceandNationalEconomy, NorthernStates FinalAccount Reports, and Bank staff estimates. 89