Report No. 34582-AF Afghanistan Managing Public Finances for Development (In Five Volumes) Volume I: Main Report December 22, 2005 Poverty Reduction and Economic Management Sector Unit South Asia Region Document of the World Bank Table of Contents ACKNOWLEDGEMENTS ....................................................................................................................... V EXECUTIVE SUMMARY VI CHAPTER1 PUBLICFINANCEINAFGHANISTAN'SDEVELOPMENT . ....................................................................................................................... ...................................... 1 A.ECONOMICPERFORMANCEAND CHALLENGES ANDDEVELOPMENT STRATEGY ................................ 2 Progress and Problems.......................................................................................................................... 2 B.Development Strategy .......................................................................................................................... 4 C.THECONTRIBUTIONOF PUBLICFINANCE............................................................................................ 4 KEYPUBLIC FINANCEISSUES ............................................................................................................... 7 CHAPTER2 ASSESSINGPERFORMANCEINPUBLICFINANCIALMANAGEMENT . ...........10 A.A FRAMEWORK ANALYSISAND ACTION.................................................................................... FOR 10 B.THEPERFORMANCEOFAFGHANISTAN'S PFMSYSTEM.................................................................... 11 CHAPTER3 PROGRESSINGTOWARD OVERALLFISCALSUSTAINABILITY . ...................... A. FISCALSTRUCTUREANDTRENDS..................................................................................................... 15 15 Budget Structure and Recent Trends .................................................................................................. 15 Implications ........................................................................................................................................ 16 B.A FRAMEWORK ACHIEVINGFISCALSUSTAINABILITy................................................................ FOR 17 Fiscal Sustainabilityinthe Context o f Afghanistan ........................................................................... 17 c.MANAGING Medium-Term Fiscal Framework: What It Means. How ItWorks ................................................... 18 EXPENDITURES FROMA MEDIUM-TERM PERSPECTIVE................................................... 19 The GovernmentNon-Security Payroll ($187 millionin2004/05) .................................................... 19 Non-wage Operating Expenditures ($96 millionin2004/05. excluding security) ............................. 22 The Development Program ($300 millionCore. $1.467 millionExternal in2004/05) ...................... 22 D.Security Sector Spending ($1.328 million in2004/05. of which $292 million Core) ........................ 23 FINANCINGRECONSTRUCTIONAND MANAGING FISCALRISKS ........................................................ 26 Managing Budget Financing over the MediumTerm ........................................................................ 26 Strengthening Municipal Finances ..................................................................................................... 27 E.Managing National Assets including State-Owned Enterprises......................................................... 28 MEDIUM-TERM FISCALSCENARIOS.................................................................................................... 29 CHAPTER4 MOBILIZINGDOMESTICREVENUES . ...................................................................... 32 A.REVENUE STRUCTURE ........................................................................................................................ 32 B.IMPROVINGREVENUEPOLICY AND TAX ADMINISTRATION .............................................................. 34 General Approach............................................................................................................................... 34 Customs Administration Reform........................................................................................................ 35 Domestic tax administrationreform................................................................................................... 35 Moving toward a broad-based consumptiontax................................................................................. 36 C.Taxation and investment incentives.................................................................................................... 37 MEDIUM-TERM REVENUE PROJECTIONS............................................................................................ 38 CHAPTER5 PRIORITIZINGEXPENDITURES . ................................................................................ 40 A.SECTORALEXPENDITURE PATTERNS: WHY ITI S IMPORTANTTO PRIORITIZE ................................... 40 The Allocation o f Spending Across Sectors and Recent Trends ........................................................ 40 B.Why TOPRIORITIZE:ALLOCATING Prioritize? Key Trade-offs ....................................................................................................... 41 HOW EXPENDITURESACROSS SECTORS............................................. 42 C.HOW TO PRIORITIZE: ALLOCATINGEXPENDITURES WITHIN SECTORS ............................................. 44 CHAPTER6 MAKINGTHE NATIONALBUDGETTHE CENTRALINSTRUMENT OF . POLICYANDREFORM ......................................................................................................................... 47 A.BUDGET PREPARATIONPROCESS....................................................................................................... 47 B.POLICY ORIENTATIONOFTHEBUDGET .............................................................................................. 48 i Comprehensiveness and Classification............................................................................................... 48 c.ENHANCINGOWNERSHIPANDPOLITICALB w - 1............................................................................. From Policy to Budget........................................................................................................................ 49 ~ 50 D.ROLEOFDONORSINIMPROVINGTHEBUDGET PROCESS.................................................................. 51 CHAPTER7 ENHANCINGTHE EFFECTIVENESSOFBUDGETEXECUTION . ........................ 54 A.GETTING FUNDSTO SERVICEDELIVERY UNITS ................................................................................. 54 B.CONTROLLINGTHEUSEOFPUBLICFUNDS........................................................................................ 55 Treasury Functions ............................................................................................................................. 56 c.PROCURING Internal Audit Functions ..................................................................................................................... 57 GOODSAND SERVICES ..................................................................................................... 58 D.HOLDING GOVERNMENTACCOUNTABLE THEUSEOFPUBLICFUNDS....................................... FOR 59 External Scrutiny o f the Budget ......................................................................................................... 59 FightingAgainst Corruption............................................................................................................... 60 CHAPTER8 DELIVERINGSERVICESTOTHE AFGHANPEOPLE . ............................................ A.SIX MODELSPUBLICSERVICEDELIVERY 63 OF ..................................................................................... 63 Centralized Government Service Delivery......................................................................................... 63 Outside Government........................................................................................................................... 65 Government Outsourcing to NGOsPrivate Sector............................................................................. 66 Regulation and/or Contracting of the Private Sector .......................................................................... 67 Community-based ............................................................................................................................... 69 B.MunicipalAND IMPLICATIONS ............................................................................................................. 71 Services ............................................................................................................................. 70 LESSONS Financingo f Service Delivery and the Role o f Cost Recovery .......................................................... 72 C.Monitoring and Evaluation................................................................................................................. 74 SUB-NATIONAL STRATEGY................................................................................................................ 74 CHAPTER9 INSTITUTIONAL REFORMSAND CAPACITY DEVELOPMENT . ........................ 77 A.DEVELOPING GOVERNMENT'S THE CAPACITY .................................................................................. 77 Three Strategic Components .............................................................................................................. 78 Two Key Challenges .......................................................................................................................... 79 B.REFORMING THEMINISTRY F~ANCE OF ............................................................................................ 80 C.BUILDINGPFMCAPACITY THROUGHOUT GOVERNMENT ................................................................. 81 CHAPTER 10 KEY CHALLENGESAND A ROADMAPFORPFMREFORM . .............................. 83 A.EIGHTMAINCHALLENGES................................................................................................................. 83 B.A FIVE-POINT REFORMAGENDA....................................................................................................... 85 STATISTICAL APPENDIX ..................................................................................................................... 93 Table A. 1: Macroeconomic Indicators.............................................................................................. 93 Table A.2: Social Indicators (2003) ................................................................................................... 93 Table A. 3: Fiscal Summary (Afs Million) ......................................................................................... 94 Table A.4: Expenditure Comparisons................................................................................................ 95 Table A. 5:Revenue comparisons ...................................................................................................... 96 Table A. 6: Fiscal Scenarios ............................................................................................................... 97 Table A.7: Expenditures by Program from 2003104 to 2005106 ....................................................... 98 LIST OFADDITIONAL DOCUMENTSPREPAREDAS PARTOFTHE PFMREVIEW .............99 REFERENCES ........................................................................................................................................ 100 .. 11 BOXES Chapter 1 Box 1. 1: Highlightsof the PFMReview Process......................................................................................... 1 Box 1.2: BadPublic Finance Management i s Bad for Development........................................................... 5 Box 1. 3: The Challenges o f Increasing Aid Effectiveness........................................................................... 9 Chapter 3 Box 3. 1: Possible Elements o f a Medium-Term Fiscal Framework for Afghanistan ................................ 19 Box 3.2: Salient Features o f Afghanistan's Security Sector ...................................................................... 25 Box 3. 3: Recommended Guidelines for Responding to Budget Proposals for SOEs ................................ 29 Chapter 4 Box 4. 1: Summary o f Major Tax Instruments inAfghanistan................................................................... Box 4.2: Afghanistan's Trade Regime ....................................................................................................... 33 36 Chapter 5 Box 5. 1:Prioritizing When Everythingi s Critical ..................................................................................... 46 Chapter 6 Box 6. 1:Who Decides on the Budget? ...................................................................................................... 50 Box 6. 2: The Critical Role o f the Afghanistan Reconstruction Trust Fund............................................... 53 Chapter 7 Box 7 1:Trachng Public Expenditures inEducation................................................................................ 54 Box 7.2: Can E-GovernmentPlay a Role inAfghanistan? ........................................................................ . 57 Box 7. 3: Procurement for Highways.......................................................................................................... 58 Box 7.4: Vulnerabilities to Corruption....................................................................................................... 61 Chapter 8 Box 8. 1:Forms o f Private Sector ParticipationinInfrastructureKJtility Sectors....................................... 68 Box 8.2: Natural Gas Pricing ..................................................................................................................... 73 Box 8. 3: Improvementsand Challenges in Subnational Administration ................................................... 75 Box 9. 1:Harnessing Afghan Expertise...................................................................................................... Chapter 9 77 Box 9.2: Using Technical Assistance Effectively ...................................................................................... 79 FIGURES Chapter 1 Figure 1. 1:Economic Growth and Inflation................................................................................................. 3 Figure 1.2: Linkages betweenPublic Finance and Strategic Objectives...................................................... Figure 1. 3: Domestic Revenues and External Assistance. $ million............................................................ 5 7 Figure 1.4: Expenditure Structure and Financing, 2004/05 Estimates......................................................... 8 Figure 2. 1: Linkages Among the Six Core Dimensions o f PFMPerformance .......................................... Chapter 2 11 Chapter 3 Figure3. 1:BudgetImplementation (2004/05 actuals as % o fbudget) ...................................................... 16 Figure3.2: KeyDrivers o fAfghanistan's Public Expenditures, 2004/05 (% o ftotal budget)...................20 Figure 3. 3: Pay Scales ($/month) ............................................................................................................... Figure 3.4: Forecast Security Sector Spending Against ProjectedRevenues ............................................ 21 24 Figure 3. 5: Three Illustrative Fiscal Scenarios (% o f GDP) ...................................................................... 30 Figure 4. 1: Total tax payable (% o f gross profits)...................................................................................... 37 Chapter 4 Figure 4.2: Medium-tern revenue forecast (% o f GDP)............................................................................ 39 Figure 4. 3: Comparative Growth inRevenue Collection (% o f GDP)....................................................... 39 ... 111 Chapter 5 Figure 5. 1: 2004/05 Expenditure Composition ("XI o f total o f each category) ........................................... 40 Figure 5.2: Composition o f Security Expenditures. 2002/03-2004lO5 ....................................................... 43 Figure 5. 3: Net Enrollment Ratios. Grades 1-6 (2003) .............................................................................. 44 Chapter 6 Figure 6. 1: Comprehensiveness o f Budget and ExternalAssistance ("A GDP) ......................................... 48 Figure 6.2: Operating and DevelopmentBudgets ($ million) .................................................................... 49 Chapter 7 Figure 7. 1: Share o fBudget Allocated to Provinces (% o f Total Expenditures) ....................................... 55 Figure 7.2: Share o f Expenditures Reimbursed by ARTF ......................................................................... 56 Figure 7. 3: Six Dimensions o f Governance ............................................................................................... 60 Chapter 8 Figure 8. 1: Service Delivery Framework................................................................................................... 63 Figure 8.2: Service Delivery by Central Government (Education) ............................................................ 64 Figure 8. 3: Enrollment Growth (Grades 1-12) ........................................................................................... 64 Figure 8.4: Service Delivery Outside Government (Humanitarian) .......................................................... 66 Figure 8. 5: Service Delivery by Outsourcingto NGOs (part o f Health services) ...................................... 66 Figure 8. 6: Service Delivery by Private Sector (Telecommunications) ..................................................... 68 Figure 8.7: Service Delivery by Communities (National Solidarity Program) .......................................... 69 Figure 8. 8: Service Delivery by Municipalities ......................................................................................... 70 Figure 8. 9: Administrative and Political Structure o f Afghanistan............................................................ 75 TABLES Chapter 1 Table 1. 1:Fiscal Outcomes ($ million) ........................................................................................................ 6 Chapter 2 Table 2. 1:Public Finance Management Performance Indicators............................................................... 12 Chapter 3 Table 3. 1:Wage Bill inAfghanistan.......................................................................................................... 20 Table 3.2: Three IllustrativeFiscal Scenarios ($ million) .......................................................................... 31 Table 4. 1: Central Government Revenues ("XI o f GDP) ............................................................................. Chapter 4 32 Table 4.2: Domestic Revenue Collection................................................................................................... 33 Chapter 5 Table 5. 1:2004/05 National Budget Expenditures by Program ($ million) .............................................. 41 Table 5.2: Cost-Effectiveness o f Different Health Interventions............................................................... 45 Chapter 6 Table 6. 1:Key Steps and Issues inBudget Preparation............................................................................. Table 6.2: UsingNational Budget Systems................................................................................................ 48 52 Chapter 7 Table 7. 1: Constraints on Local h v a t e Sector Participation to Public Procurement ................................ 59 Chapter 8 Table 8. 1: Contracted Costs for Delivering BPHS ($ per capita o fpopulation served) ............................. Table 8.2: DABM:Effective Tariffs and Costs by Region, 2004/05 ......................................................... 72 67 Chapter 10 Table 10. 1:Roadmap to Strengthen PFMPerformance............................................................................. 87 iv ACKNOWLEDGEMENTS This report, a product o f the World Bank's Public Finance Management Review for Afghanistan, was prepared by a team led by William Byrd (SASPR) and StCphane Guimbert (SASPR) and including Julia Fraser (SASEI), Soraya Goga (SASEI), M. Qahir Haidan (SASEI), Michael Haney (SASEI), Q u a m 1 Hasan (SARPS), Wall Ibrahimi (SASEI), Masami Kojima (COCPO), Benjamin Loevinsohn (SASHD), Peter Middlebrook (SASPR), Keiko M i w a (ECSHD), Kayhan Natiq (SASHD), Asta Olesen (SASES), Habib Rab (DFID), Arlene Reyes (SASPR), Robert Rothery (ADB), Paul Sisk (SARFM), Michael Stanley (COCPO), Juliet Teodosio (SASPR), The0 Thomas (IMF), Nicole Ball (Consultant, DFID), Robert Batt (Consultant), Anne Evans (Consultant, AREU), Kimo Karini (Consultant, EC), Yasin Osmani (Consultant, AREU), Peter Trepte (Consultant), Christopher Ward (Consultant), and David Webber (Consultant). Invaluable support was received throughout the process from William Dorotinsky (PIMPS), Satendra Prasad (SASPR), and Nicola Smithers (PEFA). Responsibilities for individual volumes are as follows: Volume I(Main Report): William Byrdand StCphane Guimbert. Volume I1 (Public Financial Management Performance Assessment and Procurement): Paul Sisk, Q u a m 1 Hasan, StCphane Guimbert, The0 Thomas, William Byrd, and Peter Trepte. Volume I11 (Key Cross-Cutting Issues): Anne Evans, Soraya Goga, StCphane Guimbert, Yasin Osmani, Habib Rab, and The0 Thomas. Volume IV (Case Studies o f Selected Sectors): William Byrd, Julia Fraser, M.Qahir Haidari, Michael Haney, Wall Ibrahimi, Kimo Karini, Masami Kojima, Benjamin Loevinsohn, Keiko Miwa, Kayhan Natiq, and Michael Stanley. Volume V (Improving Public Finance Management in the Security Sector): Peter Middlebrook, Nicole Ball, William Byrd, and Christopher Ward. Peer Reviewers for the PFMReview, whose comments are greatly appreciated, include: Richard Bontjer (former adviser to the Government o f Afghanistan), Santhanam Krishnan (SARPS), Rajat Narula (LOAG2), Zmarak Shalizi (DECRG), Vinaya Swaroop (SASPR), and Steven Symansky (IMF). This work was camed out with guidance and supervision from Kapil Kapoor (Sector Manager, SASPR) and on procurement and financial management from Els Hinderdael (SARPS) and Robert Saum (SARFM), respectively. Overall leadership was provided by Sadiq Ahmed (Sector Director, SASPR) and Alastair McKechnie (Country Director, SAC01). Comments received at the Concept Review meeting (September 27,2004), duringa mid-ternreview, at a Review meeting (October 24,2005), and at various times on the different backgroundpapers prepared for the PFM Review are gratefully acknowledged. All volumes o f this report were processedbyJulietTeodosio. The collaboration of the following agencies with the World Bank inworkmg on specific parts o f the P F M Review is gratefully acknowledged: IMF (revenue), DFID (state-owned enterprises, security), the European Commission (highways), and ADB (procurement). The process o f carrying out the PFMReview involved extensive interactions with the Ministryo f Finance and other Government agencies, whose valuable cooperation and support was indispensable for completingthe review and also greatly benefitedthis report. V EXECUTIVE SUMMARY 1. Afghanistan's reconstruction has made considerable progress during the past four years. Led by the Government with international support but relying most on the energy and initiative o f the Afghan people, reconstruction has resulted in solid achievements - rapid economic growth, unprecedented primary school enrollments including for girls, great expansion o f immunization, rehabilitation o f major highways, a new and stable currency, promulgation o f a new Constitution, Presidential and Parliamentary elections, return o f refugees, and demobilization o f militias. Public Finance Management (PFM) has made a major contribution to these successes. Yet the challenges remain enormous. ii. ThemainobjectiveoftheAfghanistanPFMReview,conductedbytheWorldBankandpartner agencies in close cooperation with the Government o f Afghanistan, i s to consolidate, deepen, and present inan accessible, action-oriented form the knowledgebase on Afghanistan's public finance system, review recent progress, analyze key challenges, and put forward options and recommendations for moving forward. This main report i s supplemented by four additional volumes, covering PFM performance and procurement (Volume 11); key cross-cutting issues (Volume 111); selected sector studies (Volume IV); and security sector expenditures (Volume V). PUBLICFINANCE MANAGEMENT AND AFGHANISTAN'S DEVELOPMENT ... 111. Fiscal structure and trends. Afghanistan has an unusual fiscal structure reflecting its historical legacy (e.g. traditionally low revenue mobilization), the long period o f conflict (e.g. civil service wages greatly eroded and compressed by hyperinflation), and current reconstruction activities (very high spending) and priorities (e.g. massive expenditures on security). More specifically: 0 Public expenditures are extraordinarily high; total budgetary expenditures in 2004/05 were equivalent to 57% o f GDP (excluding the drug economy). 0 This is accounted for by extremely highdevelopment spending; unlike inmost countries, the development budget ($2.8 billion in2004/05) dwarfs the operating budget ($0.6 billion). 0 Operating expenditures (9% o f GDP in2004/05) are well within internationalpatterns. 0 There are however large recurrent expenditures in the development budget, including health services, technical assistance, salary payments, and grants, among others. 0 There i s substantial underspending on non-wage operation and maintenance (O&M). 0 Most spending (three quarters in2004/05) occurs outside Government channels and oversight through the External Budget (as opposed to the Government-controlled Core Budget). 0 The security sector (39% o f total expenditures in2004/05) i s a major driver o f spending. 0 A disproportionate share o fpublic spending occurs inKabul (e.g. 70% o fnon-wage O&M). 0 Domestic revenues (5% o f GDP) are very low (likely the lowest ratio in the world for a sizable country) and pay for only about 8% o f total expenditures. 0 The fiscal deficit i s extremely highand entirely financed by aid, mostly grants. iv. Turningto fiscal trends since 2001/02, (i) budgetary expenditures have increased sharply during 2002/03 to 2004/05, with operating expenditures almost doubling and development spending growing even faster; (ii)revenues also have risen rapidly, more than doubling during the same period; (iii) since revenues started out from a much lower base, the absolute gap between recurrent spending and revenues more than tripled; (iv) this means that external financing o f recurrent expenditures has been a fast- growing component o f the budget; (v) the budget deficit (before grants) has increased sharply, reflecting large inflows o f external assistance predominantly on a grant basis; and finally, (vi) actual development expenditures (in both core and external budgets) have fallen far short o f budget targets (reflecting past inclusion o f unhnded projects in the development budget, confusion in some cases between commitments and disbursements, overly optimistic projections o f disbursements, insufficient project screening, large contingencies inthe original budget, and implementation constraints). vi V. Contribution of PFM to Afghanistan's development. Public finance has played and will continue to play a critical role contributing to economic stability, state-building, growth, and poverty reduction, including in particular enabling Government to perform effectively and deliver public services to the Afghan people (Chapter 1). To support these objectives, public finances mustbe: (i) affordable; (ii) well-prioritized in accordance with the national strategy; (iii)efficient in terms o f value for money and service delivery; and (iv) fair, accountable, and transparently reported to the Afghan public, private business, and other stakeholders. On the negative side, weaknesses in public finance can enormously harm development. Poor fiscal management can result in hyperinflation (as occurred in Afghanistan in the 1990s). Failure to prioritize spending among programs or problems inbudget execution (payments, procurement, accounting, etc.) can result in waste o f resources and poor value for money. And lack o f accountability and transparency can result in corruption and loss o f public support. This i s an area o f serious concern for many Afghans; there are widespread allegations o f various hnds o f corruption, including massive corruption associated with the drug industryas well as petty corruption. vi. Facing a difficult post-conflict situation, Afghanistan has achieved remarkable progress on the fiscal front. Despite pressures, Jiscal discipline has been strictly enforced and maintained, notably through control over the Government wage bill. This has been the essential ingredient o f macroeconomic stability. Domestic revenues have increased substantially, growing at an average annual rate o f 60% per year during 2002/03-2004/05 and rising from 3.2% o f GDP to 4.5% in this period. Major efforts have been made to improve the budgetprocess, and it has come a long way since the first post-Talibanbudget o f 2002/03. Budget execution also has greatly improved inmany respects - not least, most civil servants are now being paid, on time. Finally, the Government has made a strong commitment to Jinancial transparency and accountability, resulting in improving fiduciary performance. There have also been institutional improvements and capacity development, especially in the Ministry o f Finance (MoF). These achievements have provided confidence to donors resulting in mobilization o f high levels o f external support for the national budget, mainly through the Afghanistan Reconstmction Trust Fund (ARTF) which finances most o f the civilian recurrent budget. These achievements are commendable compared to experience in other post-conflict countries. However, improvements have been to a large extent dependent on external capacity brought in on an "emergency" basis to get things going. The shift from this situation to sustainable core Government capacity for key public finance functions has barely begunfor the most part, and there are also daunting weaknesses inPFMcapacity inthe line ministries. vii. PFMperformance measurementframework and assessment. As discussed inChapter 2, a PFM performance rating system has been recently developed by the Public Expenditure and Financial Accountability (PEFA) multi-agency partnership program (including the World Bank, IMF, EC, and a number o f other agencies). It i s intended to provide an objective, internationally comparable framework for assessing the performance o f a country's P F M system. This does not attempt to measure fiscal outcomes, the substantive appropriateness o f public expenditure policies and decisions, or the actual impacts and value for money achieved through public expenditures. The PFMperformance measurement framework instead should be seen as assessing the P F M system, which i s a crucial enabling factor for achieving broader development goals and substantive outcomes. Structured around six core dimensions o f PFM performance (budget credibility; comprehensiveness and transparency; policy-based budgeting; predictability and control inbudget execution; accounting, recording, and reporting; and external scrutiny and audit), the fi-amework includes 28 high-level PFM performance indicators and three indicators o f donors' performance. The performance o f Afghanistan's PFM system against this framework as o f June 2005 has been assessedas part o f the PFMReview (see Table 2.1). viii. ... This assessment generally portrays a public sector where financial resources are, by and large, being used for their intended purposes as authorized by a budget which is processed with transparency and has contributed to aggregate fiscal discipline. The expenditures and financial position o f the Government are reportedregularly in an understandable format. Performance regarding the allocation of spending across programs and agencies and the efficiency o f operations is not as good, however. Given that the starting point in 2001 was extremely weak, this is overall a remarkable achievement - yet vii pointing to the need for further progress. Using these ratings as a baseline, future progress in improving the performance of the PFM system can be monitored. However, measurable progress in terms o f changes in ratings i s more likely to occur over periods o f a year or longer than on a short-term basis. Moreover, since the ratings assess the system's current performance with external support (both advisory and operational), questions arise about sustainability. Further improvements in many cases may be manifested not inhigher ratings but rather inmaintenance o f current ratings based on sustainable national capacity and with declining levels o f external support. The ratings also provide an assessment for the donor community o f the potential capacity o f Government systems (currently supported by external assistance) to implement the operations which are now occurring outside Government systems. EIGHTCHALLENGESFORPUBLICFINANCE MANAGEMENT ix. Despite recent progress, Afghanistan's P F M system faces daunting challenges, which can be summarized as follows. x. Progressing towardfiscal sustainability. Afghanistan's annual budgets have been prudent, and the Government has a strong policy to avoid domestic borrowing (relying instead primarily on external assistance and to a much lesser extent on growing domestic revenues to pay for its expenditures). From a medium-term perspective, however, progressing toward fiscal sustainability (defined very simply in this report as reaching a fiscal position where domestic revenues are large enough to cover total recurrent expenditures) will be a daunting challenge (discussed in Chapter 3). kght now, domestic revenues cover less than half o f total recurrent budgetary expenditures and only a quarter o f all recurrent expenditures (if the large amounts ofrecurrent expenditures inthe entirely donor-financed Core Development Budget and External Budget are included). Failing to keep public expenditures within affordable limits would have dire consequences. A large influx o f aid, combined with lack o f domestic capacity and low domestic revenues, can result in an unaffordable pattern o f expenditures if not well-managed. While international support gives Afghanistan an opportunity to temporarily soften its national budget constraint, a visionary medium-term framework (Le. a Medium-Term Fiscal Framework, MTFF) i s needed to ensure that the country moves toward a sustainable fiscal position. Afghanistan has taken the first step in this regard by approving its first MTFF in October 2005. Progressing toward fiscal sustainability will require, inter alia, (i)sustained rapid growth o f domestic revenues (discussed below); (ii) containment o f overall expenditures - notably for non-discretionary expenditures such as wage bill and pensions - in line with medium-term resource constraints; (iii) ensuring that downstream expenditure liabilities created by public investments and other spending decisions are affordable; and (iv) incorporating recurrent expenditures funded directly by donors through the External Budget (notably those in the security sector) in the fiscal sustainability equation and bringingthese expenditures progressively into nationalbudget channels. xi. Mobilizing domestic revenues. Sustainedrapid growth o f domestic revenues will be the lynchpin o f progress toward fiscal sustainability (Chapter 4). Afghanistan's revenue-to-GDP ratio at around 5% i s well below half the level even in other very poor countries. In order for revenues to catch up with recurrent expenditures over a period o f a decade or so inline with the Government's objectives, revenues will have to grow much morerapidly than expenditures on a sustained basis-most likely at least twice as fast. At the same time, a sound revenue system shouldbe conducive to development o f the private sector. Key challenges inraisingrevenues include: (i) o f capacity inthe tax administration system as well as lack among taxpayers to calculate and pay taxes; (ii) the existence o f numerous, low-yielding "nuisance taxes", illicit revenue collection by many local authorities, instances o f double-taxation (mainly between municipal and national taxation), and ambiguities in the tax laws; (iii) the dominance o f agnculture and the informal sector in the economy, which are difficult to tax in any country; and (iv) widely perceived corruption in the tax administration. The Government has already made progress in reforming the Customs tariff structure and improving the income tax system. Further progress will depend on focusing efforts on larger taxpayers and high-yielding revenue sources; eliminating "nuisance taxes" and craclung down on illicit levies as well as clarifying ambiguities inthe tax code; and building capacity inthe form o f a well-managed professional tax cadre with integrity and associated physical, IT (information ... Vlll technology), and business process assets. Increasing domestic revenue mobilization needs to be given very highpriority: it will reduce aid dependence over time, create fiscal space for development, and send a very strong signal o f the Government's commitment to progressing toward fiscal sustainability. Full implementation o f the current policy and administrative reform plans would raise the revenue to GDP ratio from 5% to 8% by 2010. More forceful implementation o f administrative reforms, reaching quickly to provinces, and additionalrevenue measures could bringthis ratio to 10-11%in2010. xii. Prioritizing expenditures. Strategically prioritizing expenditures so they effectively support national development objectives i s both one o f the most important and one o f the most difficult aspects o f public finance management (Chapter 5). Pnoritization i s critical because: (i)like any country, Afghanistan faces major resource constraints over the medium term; (ii) with the bulk o fpublic spending externally financed and most o f it donor-executed, if the Government does not prioritize the allocation o f expenditures, priorities will be determined inan ad hoc manner by the fragmented priorities o f others; and (iii)capacity i s limited. Prioritization i s a political process; an orderly budget process (see below) i s the means by which decision makers (Cabinet, Parliament) make trade-offs under the constraint o f limited resources. This process must be anchored innational and sectoral strateges and based on solid evidence coming out o f the monitoring and evaluation system. xiii. ... There are no easy answers on how to prioritize expenditures across sectors. Simple rules o f thumb (identifying and mitigating gross anomalies, looking at international patterns of spending and outcomes, identifying bottlenecks and directing expenditures to alleviate them) are likely to work better than sophisticated technical approaches. For example, the allocation to the justice sector appears to have been grossly inadequate in recent years. It is very important that inter-sectoral prioritization be comprehensive - no sector (e.g. defense) should be considered sacrosanct or immune from scrutiny. Prioritization o f expenditures within sectors and major programs needs to be firmly anchored in sector strategies. Within sectors, prioritization can rely on analysis o f the costs and benefits o f various interventions (e.g. highways, dams). Of particular importance i s the review of the costs o f operating and maintaining (O&M) public investments (e.g. hospitals, roads). For instance, new investments are being made in the power sector, even while existing power supply inKabul depends on off-budget provision o f fuel by an external donor. A simple rule of thumb suggests that investments made in 2004/05 could generate an additional $100 million in annual O&M costs (i.e. 40% o f all domestic revenues collected in that year). Taking into account complementarities (where the effectiveness o f spending on one activity i s dependent on another activity) and sequencing issues also can help guide intra-sectoral prioritization. xiv. Making the national budget the central instrument of policy and reform. While much progress has been made in strengthening the budget formulation process, there i s a long way to go toward a policy- based, well-prioritized budget with strong political buy-in. As discussed in Chapter 6, key constraints include: (i)limited capacity in M o F and especially in line ministries; (ii)the dominant share o f expenditures directly by donors outside Government systems; (iii)serious timing dilemmas and a compressed budget schedule which weakens the quality o f the budget; and (iv) lack o f a medium-term strategic and fiscal perspective to guide annual budget formulation. An MTFF (Medium-Term Fiscal Framework), integrally linked to annual budgeting, i s critical to improve the budget process. K e y cross- cutting issues such as gender and counter-narcotics need to be mainstreamed in budgetary decision- malung. Meaningful processes o f political engagement and approval (including by Parliament) and public communications are required. There i s also an urgent need for greater inputs and participation in budget formulation by lower levels o f the Government administration(provinces and districts). xv. Enhancing the effectiveness of budget execution. While considerable progress has been made in some areas o f budget execution, a great deal remains to be done in four respects - which would improve the quality o f spending and the trust that the Afghan people, the private sector, and donors place in the Government's systems (Chapter 7). First, financial management processes need to ensure that funds reach service delivery units; this i s a serious weakness in most sectors (e.g. education). Second, further strengthening o f the control framework for public spending in particular requires implementing the new Public Expenditure and Financial Management L a w and creating a strong internal audit function. Third, ix while the existing procurement framework includes basic financial controls and stresses competitive bidding, it places an excessive emphasis on price (at the expense o f quality o f goods and services), and there is too limited participation from the private sector. Progressive implementation o f the new Procurement Law, combined with substantial training o f civil servants as well as private firms, will create the foundations for a fair, transparent, and effective procurement system. Finally, external scrutiny and audit are necessary to hold Government accountable for the use o f public funds. Much remains to be done to buildthe external auditor's capacity. The election o f the Parliament also creates an opportunity - with its own challenges - to further strengthen accountability and participation. Progress in these four areas will help greatly inthe fight against corruption, and should be a key part o f a holistic anti-corruption strategy. Corruption i s considered to be a symptom o f poor governance, and international experience demonstrates that anti-corruption strategies work better through addressing the underlying governance problems (e.g. inthe PFMsystem), rather than through investigationand prosecution o f corruption cases. xvi. Delivering services to theAfghan people. As discussed inChapter 8, the ultimate outcome o f the PFM system i s service delivery, essential for public spending to achieve results and extremely important for the credibility and legitimacy o f the State. However, as recognized by the Government's strategy documents, it i s not required for the Government to deliver all public services itself. Other possible Government roles include financing, setting policy, contractinghegulating, and monitoring. Various stakeholders - non-Government providers, private sector, users, and citizens - also have important roles to play. With a few exceptions, public service delivery in most sectors i s poor, as demonstrated by available data on outcome indicators, giving rise to frustrations and weakening the credibility o f the Government. xvii. The experience inAfghanistanprovides some important lessons. Performance-based contracts in the health sector appear to be a promising way to deliver services effectively, while keeping full regulatory control and sufficient visibility for the Government. Private sector participation, supported by direct payments from users, has stimulated rapid growth o f telecommunication services, while state- owned enterprises have been much less successful. As part o f the National Solidarity Program, half o f the villages o f Afghanistan now have elected development councils empowered to make local investment decisions using block grants. Afghanistan can also learn from less positive experiences, such as the failure to provide non-salary budgets to most schools and the unsustainability o f financing for electricity. Some key lessons include: (i) the critical need to focus on service delivery, put people at the center, and get accountability and incentives right; (ii) choices and expenditures should depend on sector strategies and circumstances and be guided by adequate monitoring and evaluation; and (iii) sustainable financing and, where appropriate, cost recovery are essential. xviii. Institutional reforms and capacity development. Meeting the challenge o f sustainable national capacity development i s critical for success in improving Afghanistan's PFM system (Chapter 9). At present, national capacity in the PFM system i s limited, virtually nonexistent in many areas. External capacity has been brought in to support core functions on an "emergency" basis, and large amounts o f technical assistance ($400 million per year according to OECD data) injected (for all areas o f activity, not just PFM). But this i s not sustainable, and there are serious risks that slower but more sustainable national capacity development will be undermined, malung Afghanistan highly vulnerable to a reduction in aid flows. Developing capacity requires public administration reforms (recruitment, pay, human resources), organizational restructuring and improvements, and training. Better management o f external technical assistance i s needed to ensure that it supports longer-term national capacity development xix. Finally, underlying all these challenges i s the needfor coordination and donor alignment, within an overall national strategic and budgetary framework. Coordination within Government needs to be improved (this can be accomplished around a widely-owned national development strategy and a sound budget process), and public communications and transparency will be extremely important in fostering Government accountability vis-$-vis citizens and their elected representatives. Coordination i s a major challenge for donors as well, which also needs to occur around the national development strategy and budget process; the low share of external assistance going through national budget and treasury channels n i s a serious constraint in this regard. Improving aid effectiveness requires a continuing and deepening partnership between the Government and the donors. A FIVEPOINT AGENDAFORACTION xx. The Government has fully recognized these challenges and the centrality o f the budget process and PFM performance for Afghanistan's future. As the Government is preparing its national development strategy, a five-point agenda to improve PFM i s put forward inthis report for consideration (Chapter 10). xxi. First, the path towardfiscal sustainability needs to be mapped out and steady progress in this direction achieved. Based on a sound MTFF(rudimentary at the outset but improved over time), this will require particular attention to (i)increasing domestic revenue mobilization (through administrative improvements leading to better compliance as well as sound tax policies); (ii) maintaining control over aggregate expenditures within the overall resource envelope; (iii) setting and implementing clear policies for key expenditure components notably the Government wage bill (including the non-civilian component, much o f which i s currently outside the Core Budget); (iv) monitoring and controlling fiscal risks (including pension liabilities, external debt, and liabilities o f SOEs); and (v) ensuring that strategic and public expenditure decisions are affordable over the medium-term, including downstream O&M implications. Progress toward fiscal sustainability will be measured by achievements against fiscal targets under the MTFF. xxii. Second, a number ofpolicy decisions and actions are needed to improve service delivery. This will require clarifying the roles of the State (operational, financing, policy, and regulation) in service delivery in different sectors; strengthening accountability; ensuring sustainable financing arrangements including cost recovery where appropriate; and determining the responsibilities o f various levels o f Government administration (central ministries, provinces, districts). A variety o f different models o f service delivery are possible within Afghanistan's Constitutionally-mandated unitary state structure, and the choice among them should depend on sector-specific conditions. Although "one size does not fit all," the effectiveness o f service delivery depends largely on the degree o f ownership and the coherence o f the models in terms o f accountability, incentives, and financing. Donors also need to ensure that their interventions strengthen rather than undermine national service delivery mechanisms. Progress in improving service delivery will be measured by the articulation o f the Government's strategy - with a clear set o f indicators, baseline data, and targets - and achievements against these indicators. xxiii. Third, the capacity of the key PFM institutions should be developed. Priority areas include training, recruitment practices, pay structures, organizational reforms, detailed business processes, specific IT systems, etc. While these issues cut across all levels o f Government, MoF's capacity i s o f particular importance. As chief custodian o f the PFM system, its capacity to develop and implement policies i s at the core o f PFM performance. Progress toward an M o F Strategic Plan - initiated in M a y 2005 - will be very important in developing MoF's capacity and performance. PFM capacity development inline ministries also i s very important but will take more time. Progress inthis area will be measured by the finalization o f an M o F Strategic Plan with monitoring indicators and achievements against these indicators. xxiv. Fourth, and related to the previous recommendations, the Government should adopt a clear action plan to improve PFM performance, with time-bound milestones, quantified objectives, and clearly identified responsible agencies. This proposed PFM action plan would consist o f a set o f mutually supportive measures that are feasible, realistic, and sustainable, and that can generate a step- change in PFM performance each year. The action plan should in addition include (i)a reform communication plan (with staff in MoF, with the Cabinet and other Government agencies, with the Afghan people, and with the international community); (ii) a solid organizational and institutional plan; and (iii) a process to monitor actions and performance, review progress, and adjust the program. A detailed roadmap for improving P F M performance, which i s linked to the PFM performance indicators and could form the basis for annual action plans, i s presented in the form o f a matrix at the end o f this xi report (see Table 10.1). Progress in implementing the action plan could be measured by the key fiscal outcomes as well as progressive improvement o f the MTFF, and by the P F M performance indicators presented in this report. Monitoring o f progress, along with a meaningful feedback loop into subsequent decision-making, will be as important as the action plan itself. xxv. Finally, continuing and further deepening collaboration between the Government and the international community will be critical for the success of the reform program. The Government's development strategy - which may include plans toward fiscal sustainability, more effective service delivery, capacity development, and strengthened PFM performance - should serve as the basis for a framework o f mutual accountability. Donors would align their support to national strategic objectives and the national budget and increasingly use the Government's systems, while the Government would implement the decisions it has announced and further improve the performance o f the PFM system. In addition, the content, mode, and capacity-building contribution o f technical assistance will be very important and needs to be well-managed and well-coordinated in line with Government leadership and national budgetary and other capacity-buildingprocesses. xxvi. Wayforward. Not surprisinglyin view o f the major challenges that Afghanistan faces, this i s a multi-faceted, demanding, and ambitious medium-term agenda. Each o f the five elements includes difficult sub-agendas and important actions required for success. This points to the need to prioritize and focus on a set o f realistic yet meaningful short-term measures that will achieve significant improvements and step-changes, thereby setting the stage for further progress. A package of prioritized short-run actions could thus comprise a "platform" from which the next set o f actions would take off, with monitoring and feedback to guide the process at each stage. Moreover the package, rather than any single measure or mechanical target, would be what the Government takes responsibility for and could form the basis for dialogue and agreements with the international community. xxvii. Ineach o fthe five points o f the agenda outlined above, a small number o f key elements stand out as critical for a first platform (see table below). For instance, the MTFF provides a foundation on which the Government can subsequently anchor filly costed sector strategies. With implementation o f the PFEM Law and Procurement Law, the Government will be able to establish its credibility in terms of managing public funds, better analyze patterns o f public expenditures, and introduce more sophisticated PFM practices over time. Following the establishment o f a track record o f mutual accountability with donors, a higher share o f external assistance could be channeled through the Government's systems. xxviii. Inthis process, the Government first needs to develop and reach agreement on its strategy. The Afghanistan National Development Strategy (ANDs) needs to include or to be complemented by an MTFFandPFMaction plan adoptedbythe Government, as well as anMoFreform strategy. Clear sector strategies also are needed to guide prioritization and sector expenditure programs. The ANDs and sector strategies need to specify the roles o f the State and put forward sound approaches to ensure effective service delivery, differentiated by sector as appropriate (see Chapter 8). The sector strategies will take varying amounts o f time to develop and, like the interim ANDs, MTFF, and PFM action plan, will need to be improved over time and adjusted in the light o f experience. It i s critically important that these strategies have wide national ownership, both within and outside Government. xxix. The strategy as it applies to Public Finance Management would then be translated each year into annual plans that would consist o f three building blocks: (i) an updated medium-term view o f the fiscal path (MTFF); (ii) an annual budget consistent with the MTFF and ANDs; and (iii) "annual action an plan" (a list o f a small number - certainly less than 20 - o f key actions to be accomplished during the year). Successful implementationo f this package (not necessarily all o f the measuresbut at least a critical mass o f most o f them) would build the "platform" on which the following year's action plan could be based, leading to further progress subsequently. This type o f process would help in coordinating the expectations o f partners as to what the Government realistically will be able to achieve each year and, with adequate monitoring, would help the Government establish a strong track record o f implementation. Adequate monitoring and evaluation should also feed back into the next year's annual plan and revisions to the overall strategy -this requires significant development o f statistical and analytical capacity. xii xxx. In conclusion, the five-point reform agenda outlined above is ambitious, requiring sustained efforts over a considerable period o f time based on meaningful steps each year. But, as seen earlier, the challenges in building an effective, accountable, financially self-sufficient state that facilitates sustained economic growth, ensures adequate delivery o f services to the Afghan people, and reduces poverty are enormous and pressing -requiring a commensurate response. The Government, with strong support from the international community, has demonstrated its willingness and commitment to embrace an agenda along these lines and move forward with implementation in a determined yet realistic manner. Building on the achevements of the past several years, the table below suggests key short-term priorities, progress indicators, and responsible entities to move toward the next level o f strengthened PFMperformance. Key Priorities and Progress Indicators for Improving Public Finance Management 1.MovingToward Fiscal Sustainability Progress will be measured by: Increase inrevenue to GDP ratio Decline inratio o f operating expenditures to for settingbudget envelopes (Chapter 3) revenues Implementtax measures already gazetted, Progress according to fiscal targets inMTFF focus administrative reform on Customs and 2. Improving Service Delivery Progress will be measured by: Suggested short-termpriorities include: Implemented by: Adoption o f I-ANDs with monitoring Improve alignment o f budget and ANDs by Cabinet framework correcting significant mismatches (Chapter 5) Progress according to indicators inANDs Develop selected sector strategies with a Line monitoring framework service delivery focus (Chapter 8) ministries Improved alignment between budget and Enhance statistical capacity for monitoring cso ANDs and evaluation (Chapter 8) Progress against monitoring indicators in 3. DevelopingCapacity of PFMInstitutions Progress will be measured by: Suggested short-term priorities include: Implemented by: Adoption o fM o F reform strategy with Furtherdevelop and adopt MoFReform M o F monitoring framework Strategy (Chapter 9) Progress according to indicators inM o F Pilot reform o f administrative (finance / M o F reform strategy HR)department (including budget process) in M o F 4. Improving PFM Performance Progress will be measured by: Suggested short-termpriorities include: Implemented by: P F M Performance indicators (Chapter 2) Adopt PFM action plan(to be developed by Cabinet Progress according to fiscal targets inMTFF MoF, Chapters 6, 7, 10); this would include inter alia implementation o fPFEML a w Parliament (notably development o f internal and external and civil audit) and Procurement Law society, Develop means for public engagement on supported by budget matters (formulation and Government and implementation) donors 5. Deepening the Collaboration between the Government and the InternationalCommunity Progress will be measured by: Suggested short-termpriorities include: Implemented by: Proportiono f external assistance through Agree o n a framework o f mutual Cabinet and budget channels accountability donors Agree o n good principles for TA Donors management, use o f Government systems (Chapters 6, 7, 9) ... X l l l CHAPTER 1. PUBLIC FINANCE INAFGHANISTAN'S DEVELOPMENT 1.1 This report is the product o f the Afghanistan Public Finance Management (PFM) Review. The main goal i s to consolidate, deepen, and present inan accessible, action-oriented form the knowledge base on Afghanistan's public finance system, review recent progress, analyze key challenges, and put forward options and recommendations for moving forward. The report consists o f this main volume plus four other volumes, covering PFM performance and procurement (Volume 11); key cross-cutting issues (Volume 111); selected sector case studies (Volume IV); and security sector expenditures (Volume V). 1.2 This work has been carried out by the World Bank incollaboration with other agencies workmg on specific areas: IMF (revenue), DFID (state-owned enterprises, security), the European Commission (highways), and ADB (procurement). The team has worked closely with the Government o f Afghanistan, especially the Ministry o f Finance (MoF). 1.3 The objectives of the PFM Review as a whole have been to (i) respond to specific Government requests, in particular by MoF; (ii) provide inputs for and support to Afghanistan's annual budget cycle; and (iii) consolidate and deepen knowledge o f Afghanistan's public finances, including assessment o f fiduciary aspects (financial management and procurement). The process involved extensive interactions with MoF and other Government agencies, and produced numerous intermediate products providing technical inputs on specific topics o f interest to the Government (see list o f additional documents F pared, p. 99). Specific stages and activities during the PFMReview process are outlined inBox 1.1. Box 1.1: Highlights of the PFMReview Process The Afghanistan PFM Reviewhas been very much oriented toward real-time learning and interactions in partnership with the Government, particularly MoF. In this regard, responding in a timely manner with technical inputs and discussions on topics of immediate relevance to Afghanistan from a PFM perspective was given highest priority. Selectedactivitieswhich occurredas part of the PFM Review are listedbelow, alongwith some key milestones. Initial discussions with MoF and other Governmentagencies on P F MReview (July-September2004): The concept, content, and rationale for a PFM Reviewby the World Bank were discussedwith Government officials, and feedback was obtained. Specifically, MoF requestedthat the PFM Reviewinclude analysis of the securityandmining sectors. Initiation of P F M Review task (September 27, 2004): Internal World Bank meetingreviewedthe concept paper for the PFM Review, endorsedmoving ahead, and provided guidance from Peer Reviewers. Terms of reference for each componentofthe Reviewwere subsequentlypreparedand agreedwith MoF. P F M Review mission (November 28-December 19, 2004): A sizable team of World Bank staff, joined by team members provided by DFID, the EC, and other agencies, visited Afghanistan during this period for information gathering and discussions on the whole range of topics covered in the PFM Review, including institutional, cross- cutting, sectoral, and financial managementlprocurementaspects. In particular, in response to a request from MoF, a World Bank member o f the PFM Review team worked intensively with MoF oficials on the restructuring o f MoF Treasury and AccountingDepartmentsandpreparedatechnical paper outlining issues andoptions. Discussions on security sector expenditures (December 2004 - May 2005): Analysis of security sector expenditures was initiated during the December 2004 mission and continued thereafter. Major findings, in particular related to fiscal sustainabilityissues, were sharedwith the Government, anda shortpaper was prepared. Inputsfor 1384 (200.5/06) Budget (January-March2005): During this period, the teamproducedand shared with the Government severaltechnical papers on relevanttopics for the 1384Budgetthenunderpreparation. Examples include paperson budget formulation,the Governmentwage bill, medium-termfiscal issues, andstate-ownedenterprises. Internal reviews (mid-term:February6,2005 andfinal: October 24,2005) to assess progressandoutcomes. Meetings and workshopson MOPstrategicplanning and reform (May 22-30,2005): At the request of the Ministry, World Bank staff helped facilitate a series of meetings on the development of a strategic planningheform process in MoF. Earlier a draft "vision paper" for MoF had been prepared, and in May-June a more detailed paper outlining a possibleapproachto MoF strategic planningheformwas draftedand sharedwith MoF. Ongoing discussions on the P F M Performance Assessment (January-August 2005): a draft of the assessment (see Chapter 2 and Volume 11, Part 1) was shared and discussed with the Government, which decidedto release it on the website of the Afghanistan Development Forum (April 2005). Further discussions were held in May and August aroundthis assessment, and it was subsequentlypresentedby MoF to donorsat alocal ConsultativeGroupmeeting. 1 1.4 The scope o f the PFM Review i s intended to be comprehensive, including revenue, procurement, financial management, and fiduciary issues as well as more traditional Public Expenditure Review (PER) topics. There i s also a focus on institutional and capacity aspects. The coverage o f different sectors, however, i s not complete but i s intended to provide selected examples. Given the comprehensive scope o f the PFM Review, this volume inevitably i s highly condensed; more details are available in the other volumes and additional backgroundpapers. 1.5 This overview report has ten chapters. This first chapter discusses the contribution o f public finance to Afghanistan's reconstruction and development and key challenges. Chapter 2 provides a framework for assessing the performance o f a country's PFM system and an evaluation o f Afghanistan's current PFM performance. Chapter 3 focuses on the key imperative o f progressing toward fiscal sustainability in the medium-term, highlighting possible fiscal scenarios and fiscal risks. Given the critical importance of mobilizing domestic revenue, Chapter 4 i s devoted to this issue. Moving from the aggregate level to the sectoral and subsectoral level, Chapter 5 looks at prioritization and allocation o f expenditures across and within sectors. Turning to process issues, Chapter 6 discusses budget formulation and malung the national budget the central instrument o f policy and reform. Chapter 7 i s about the budget execution process, including getting funds to service delivery units, controlling use o f public funds, improving public sector procurement, and enhancing accountability. Chapter 8 looks at service delivery, presenting some models and examples and also discussing subnational strategy from a service delivery perspective. Chapter 9 analyzes the critical issues o f institutional reforms and capacity development in relation to public finance management. Chapter 10 provides a closing summary o f key challenges along with a roadmap for improving the performance o f Afghanistan's PFM system. The Statistical Appendix presents key macroeconomic, social, and fiscal data. 1.6 Afghanistan's reconstruction has made considerable progress during the past three years. Led by the Government with international support but relying most on the energy and initiative o f the Afghan people, reconstruction has already resulted in solid achievements - for example rapid economic growth, unprecedented primary school enrollments including for girls, great expansion o f immunization, rehabilitation o f major highways, a new and stable currency, promulgation o f a new Constitution, Presidential and Parliamentary elections, return o f refugees, and demobilization o f militias. Yet the challenges remain enormous. The role o f public finance in contributing to national achievements and in meeting challenges in the future i s the main subject o f this chapter. The chapter starts by reviewing recent economic performance and the development strategy pursued by Afghanistan. It then discusses the contribution o f public finance management to these achievements and the role that it will need to play in the future. The chapter ends by briefly outlining some key public finance issues faced by the country in its efforts to achieve national objectives o f growth, state-building, and poverty reduction. A. EconomicPerformanceand ChallengesandDevelopmentStrategy PROGRESSAND PROBLEMS 1.7 Economic growth. Afghanistan's economy has rebounded since 2001, stimulated by the end o f major conflict, good precipitation for agnculture, the initiation o f reconstruction activities, and strong demand and financing from opium, remittances, and reconstruction aid inflows. Economic growth has generally been at double-digit rates (averaging 20% pea.during2001-2004), and i s expected to continue in the 10% per year range (Figure 1.1). However, recent growth has been based largely on one-time, unsustainable, or undesirable (opium) factors and appears to have been concentrated in urban areas, notably Kabul. Moreover, the predominantly informal Afghan economy (outside the formal legal framework and largely outside the tax base) i s also a constraint on future development, despite its benefits in many respects. As argued by the Government's Securing Afghanistan s Future report, Afghanistan will needto achieve sustained economic growth on the order o f 9% per year over the next decade or so in order to phase out opium production while maintaining per-capita income growth at reasonable levels. The key challenge therefore i s to transform recent rapid growth into robust, sustainable growth over the 2 medium term, based on private investment and increasingly competitive production, with strong growth o f exports. 1.X Macroeconomic stability. In contrast with the hyperinflation Afghanistan suffered from in the 1990s and unlike experience in many post-conflict situations, inflation has been kept well under control (Figure 1.1), and the exchange rate has been stable. Particularly noteworthy has been the successful introduction o f a new currency to replace the multiple, severely devalued currencies inherited from the 1990s. Prudent, conservative fiscal policy (embodied in the "no-overdraft rule") and sound monetary policy have contributed to price stability, and external financing has covered Afghanistan's large trade deficit and contributed to exchange rate stability. One concern i s that the significant albeit controlled level o f inflation (on the order o f 10% per year) combined with exchange rate stability are symptomatic o f a "Dutch-disease'' type o fproblem, whereby resource inflows due to opium, remittances, and aid drive up domestic prices and costs, malung it difficult for the Afghan economy to produce for export or compete against imports, exacerbating the other obstacles to private investment and activity. Overall, however, the macroeconomic achievements compare favorably with other post-conflict countries (World Bank, 2005a). Figure 1.1: Economic Growth and Inflation 60 50 - 40 Estimates Estimates Estimates Projections 2002103 2003/04 2004105 2005/06 3 ranging in severity up to outright conflict in some areas. Reforms in the police and especially in the judicial system have lagged far behind the development o f the ANA. Another challenge i s that the development o f security forces i s proving to be costly, with serious questions about the fiscal sustainability o f security sector expenditures (see Chapter 3 and Volume V). 1.12 Infrastructure. Road investments, both rehabilitation o f Afghanistan's main highways and building smaller and rural roads, have moved ahead, although implementation has been in some cases slower than desired for various reasons including security, and costs o f highway rehabilitation have tended to be high (see Volume N,Chapter 5). Progress has been much slower in other sectors such as power and irrigation; financing o f infrastructure i s not on a sustainable basis; many state-owned enterprises are defunct; and there are major regional differences and gaps. 1.13 Human development. Primary school enrollments have reached unprecedented highlevels (54% in 2003) including for girls (40%), and basic health services are expanding while immunization campaigns have sharply increased coverage since 2001. Progress has also been made inreviving tertiary education. Nevertheless, primary enrollments remain far below international norms especially for girls, with wide regional gaps especially by gender. Access to health i s still problematic inmany rural areas, especially for females. There are serious quality problems at all levels o f education, and differences in levels o f service provision are dramatic (see Chapter 6). 1.I4 Narcotics. The opium economy - linked with insecurity, anti-state interests, poverty, weak governance, and economic distortions - i s an enormous problem and challenge for Afghanistan's development. Modest progress has been made in interdiction against drug trafficking, and opium cultivation has been greatly reduced in some areas, but it has grown rapidly elsewhere and i s now found in all of the country's 34 provinces. Large numbers of people - farmers, wage laborers, traders, militia forces, commanders, government officials - are involved inthe drug business. The nexus between drugs, insecurity, warlords, and weak government i s a profound threat to state-buildingand reconstruction. DEVELOPMENT STRATEGY 1.15 Progress achieved. The Government articulated a compelling vision o f the country's future in the National Development Framework promulgated in 2002, and then with a more specific program in Securing Afghanistan 's Future, presented at the Berlin Conference in 2004. More recently, the Government's documentation for the 2005 Afghanistan Development Forum presented a comprehensive overview o f development issues and challenges, as well as Government plans. Progress has been slower inbuildingownership for this vision across the Government, consultingwith civil society, and ensuring that Government policies and decisions are inline with national strategic objectives andpriorities. 1.I Preparation of the National Development Strategy. The Government i s currently reviewing, 6 updating, modifying, and refining its development vision and strategy, including through widespread consultations within the Government. The interim Afghanistan National Development Strategy (ANDs) i s expected to be approved by the Cabinet, presented to Parliament (for which elections were held on September 18,2005), and discussed at a high-level donor meeting planned for January 2006. 1.17 Strategicfoundation for publicfinance management. As will be discussed later inthis report, it i s essential that there be a close linkage between development strategy and medium-term fiscal planning as well as annual budgeting. Inthis regard, the Government intends that the ANDs will form the strategic anchor for the 2006/07 National Budget. It i s also very important that the ANDs embody a medium-term fiscal fi-amework (MTFF)to guide fiscal policy and reforms as well as public sector resource allocation. B. The Contributionof Public Finance 1.18 Public finance has played and will continue to play a critical role in contributing to economic stability, growth, and poverty reduction, and in enabling the government to perform effectively and deliver public services to the Afghan people. Key linkages between public finances and the strategic objectives o f growth, state building, and poverty reductionare shown inFigure 1.2. 4 Figure 1.2: Linkages between Public Finance and Strategic Objectives 1. A Rais Spe af economy stableand maintaindonors' trus 1.19 In order to promote sustained economic growth, building an effective, accountable state, and reducing poverty, public finances must be: (i)affordable (dynamically growing revenues and expenditures under control); (ii)prioritized in terms o f allocation across sectors in accordance with national strategy, and over time so that investments made today can be effectively operated and maintained in the future; (iii) efficient in the sense o f value for money and service delivery; and (iv) accountable and transparently reportedto the Afghan public and other stakeholders. 1.20 On the negative side, weaknesses in public finance can be very harmful for development. As discussed in Box 1.2, poor fiscal management can lead to hyperinflation. It can also result in failure to prioritize spending among sectors and programs and reduced development effectiveness. Problems in budget execution (payments, procurement, accounting) can lead to waste o f resources and poor value for money. And lack of accountability and transparency can generate corruption and loss o f public support. Box 1.2: BadPublic FinanceManagement is Badfor Development Afghanistan's history in the 199Os, as well as examples from many other countnes, highlights that poor public finance management adversely affects national development ina number o f ways: At the aggregate level, lack of a credible, realistic, policy-based (including a medium-termdimension), andreasonably comprehensive budget can result in high inflation and macroeconomic instability, and a crisis for the currency and for external relations. As inflation tends to hurt the poor most severely, poverty can be exacerbated. Reducing expenditures in such a crisis situation usually means cutting discretionary spending (including investments, non-wage operating costs), and being excessively aggressive in striving to raise domestic revenues penalizes economic activities. Lack o f affordability over the medium-term often is manifested in inability to operate and maintainpublic fac service delivery networks - deteriorating roads, inability to pay for fuel for power stations, hospitals unable to pay for running costs, health facilities without medicines, schools suffering from lack o f maintenance and educational materials, etc. Such adverse outcomes disproportionately harm the poor, as they typically have few alternatives whereas better-offpeople are often able to arrange pnvate alternatives to essentialpublic services. At the sector level, a weak budget process can result in msallocations of resources and msmatches with national prionties, less value for money, and, inthe end, inadequate service delivery. At the level of individualprograms and activities, lack o f a sound budget process can result inresources being wasted, roads built to inappropnate standards; power capacity constructed without finances for fuel and operations; hospitals opened without funding, medical staff, equipment, and medicines to runthem; etc. Weaknesses in procurement, insufficient financial data given to line managers, and initial delays in making payments outside Kabul have also been a constraint on the efficiency o f public expenditures. If the procurement system i s too complex, there is limited participation in particular from Afghan contractors, and contracts end up being more expensive. On the other hand, if the procurement system does not ensure adequate competition and sound technical review, there i s a risk that opportunities for higher quality standards and lower prices s t e m n g from competition are mssed. Lack o f fiduciary safeguards inprocurement can result inoutright corruption and poor value for money. Failure to generate information on development results, on contract awards, on financial accounts, or on audits contnbutes to suspicion and lack o f trust between the Government and the people. A mismatch between the Government's capacity and its policy decisions can also weaken the legitimacy o f the Government, and i s likely to reduce the willingness o f the international community to come forward with longer-term financial contributions. Source PEFA (ZOOS), especially Appendix 1. 5 1.21 PFM achievements. In a difficult post-conflict situation, Afghanistan has made remarkable progress on the fiscal front (see Table 1.1). Despite pressures,fiscal disciplinehas been strictly enforced and maintained, including control over the Government wage bill. Although the fiscal deficit (before grants) understandably has been very large in the context o f post-conflict reconstruction, expenditures have been contained within the level o f resources available from domestic revenues and external assistance, and there has been no resort to domestic borrowing to finance the budget. This has been the essential ingredient of macroeconomic stability (see above). Domestic revenues have increased substantially, growing at an average annual rate o f 60% per year from 2002/03 to 2004/05, rising from 3.2% of GDP to 4.5% in this period. Revenue mobilization has been supported by wholesale reform and rationalization of customs duties as well as the income tax regime, and initiatives are underway to strengthen the customs administration. Major efforts have been made to improve the budgetprocess, and it has come a long way since the first post-Taliban budget o f 2002/03 (see Chapter 6). An initial Medium-Term Fiscal Framework (MTFF) has been approved by the Cabinet. Budget execution also has greatly improved in many respects - not least, most civil servants are now being paid, on time. Finally, the Government has made a strong commitment tofinancial transparencyand accountability, which has been manifested in a number o f ways and has resulted in adequate fiduciary performance albeit based on an injection o f external capacity. These achievements have provided confidence to donors resulting in mobilization o f high levels o f external support for the national budget, mainly through the Afghanistan Reconstruction Trust Fund(ARTF) which finances most o f the civilian recurrent budget (Box 6.2). Table 1.1: Fiscal Outcomes($ million) 2002103 2002103 2003104 2004105 2005106 Budget Est. Budzet Est. Budget Est.a/ Budget (In milliousofUS dollars) DomesticRevenue 62 129 184 208 322 269 333 Donor Assistance Grants(to operatingbudget)bl 283 208 322 206 313 314 345 Donor Assistance Grants(to coredevelopment budget) 93 772 173 714 Total Expenditure(Core Budget) 345 342 505 614 1,706 874 1,887 OperatingExpenditure 34s 342 505 449 635 558 678 Wages and Salaries d 264 282 377 374 460 Purchase of goods and services 133 87 101 110 83 Debt snvice and Interestpayments L 6 2 8 Other recurrent 20 18 43 30 22 Capitalexpenditure 34 61 37 41 47 Reservesand Contingencies 53 70 57 Core budget development spending dl 165 1,072 317 1,209 By Program 345 342 505 614 1,706 874 1,887 HumanCapitalDevelopment 187 130 470 349 575 Physicallnfrasmcture 24 142 380 152 584 GeneralAdminisbation 48 106 129 81 258 Security 193 235 354 292 384 Reservesand Contingencies 53 374 86 Operatingbudget balance(excludinggrants) (322) (241) (313) (289) (345) Operatingbudget balance(including grants) 0 (35) 0 25 0 Core budget balance (including grants) (0) (107) (300) nla (494) Floatand Adjustment 0 3 300 53 2 Financing 104 66 492 ExternalLoans(net) 100 309 492 Domestic (net) 19 4 (243) 1 nn share of GDP) Revenues 1 5 3.2 4 0 4.5 5.4 4.5 4 7 Donor Grants 6.9 5.1 7 0 6 5 18 1 d a I 4 9 Total Expenditure(Core Budget) 8 5 8 4 11 0 13.4 28.6 14 6 26 5 Operating Expenditure 8 5 8 4 11 0 9 8 10.6 9 3 9 5 of which wages ... 5 8 6 2 6 3 6 3 6 5 DevelopmentExpenditure 3.6 17.9 5.3 17 0 Operatingbudgetbalance(including grants) (0 1) (0.8) 0.4 CoreBudeetBalance(including grants) (0 1) (0 0) (2 3) (5 0) d a (6 9) Memorandumitem ExternalBudget(USS million) 503 1,955 2,182 3,448 2,503 3,180 ExchangeRate ( M A perUSD) 45 3 45.3 49 0 49 0 47 8 47.7 48 5 GDP (USSmillion) 4,084 4,084 4,585 4,585 5,975 5,975 7,130 ai Ple1lrmnary annd Itatemenll w Flnanclngdata basedan &IF r e p a(ARTF for 2W410Se s l m t e and ZWS106 budget) cl For 200203 10ZW4105, includesM mmleof ur-iundfmd w e n 10employees afthe M m qof hmoi, Defenre,andNationalSecunry(WEamount IS show under "pwchaseOf gods and renices" !n the Govemmenl'r budget) d/ me 2004105 budget include9 USE304rmllionsham as "unallwated"ur lhe o n m l buds1 Source: MoF documents, IMF 6 1.22 These achievements have been supported by institutional improvements and capacity development,notably in MoF. So far, improvements have been to a large extent dependent on external capacity brought in on an "emergency" basis to get things going. Examples include budget formulation, payments and reporting, procurement, audit, and support to central banlung functions. However, some progress has also been made inbuildingnational capacity in the fiscal realm on a more sustainable basis, for example in the Treasury and certain other departments o f MoF. The shift from reliance on external capacity to sustainable core Government capacity for key public finance functions has only just started for the most part, and there are also very serious weaknesses in public financial management capacity in the line ministries (Chapter 9). C.Key PublicFinanceIssues 1.23 Despite the real progress made inmany areas, fiscal challenges remain daunting. Among the key o fpublic finance issues for the coming years are the following. 1.24 Extremely low domestic revenue. Afghanistan's revenue to GDP ratio (below 5%) i s one o f the lowest in the world (it would be even lower if the opium economy were included in GDP), well below half the level achieved by most poor countries. Only around 8% o f total budgetary spending i s covered by domestic revenue. This constitutes a major constraint and liability for the country, necessitating substantial reliance on external assistance for the operating budget, which donors generally are reluctant to provide in most country contexts. Figure 1.3 highlightsthat domestic revenues are much smaller than external assistance, and that the gap between the two has increased sharply inrecent years due to the rapid growth o f aid. Thus sustained rapid growth o frevenue and a risingrevenue-to-GDP ratio are imperative. Figure 1.3: Domestic Revenues and External Assistance, $ million 1,500 1,250 1,000 750 500 250 1 2002103 2003104 2004/05 2005106 -+Domestic Revenues +Donor Assistance Grants to Core Budget +Total ExternalAssistance to Core Budget Note: Original budgetfor 200Y06. Source: MoF,IMF. 1.25 A challenging expenditure structure. Strikmgly, public expenditure in Afghanistan not only i s very large (57% o f GDP in 2004/05) but i s dominated by external financing. Only a quarter i s channeled through national Treasury and budgetary procedures (the "Core Budget"): three quarters i s executed by donors through implementing agencies contracted directly by them (the "External Budget", see Figure 1.4). This constitutes a major challenge - for both Government and donors - inbudget management. In thisregard, Box 1.3 highlightskey issues related to aideffectiveness. 1.26 The need tofurther improve the budgetprocess. Inorder to make the nationalbudget the central instrument o f policy and reform it must become more strategic (with expenditures linked to the Afghanistan National Development Strategy), multi-year in its orientation (anchored in a Medium-Term Fiscal Framework), integrated (as between the recurrent and development budgets), and with strong political buy-inand ownership (Cabinet, Parliament, public). 1.27 Sustainability of expenditures. Massive expenditures are occurring in areas like security, counter-narcotics, and highway rehabilitation and road construction, mostly through the External Budget. 7 Inaddition, social services like education and health are being sharply expanded. These investments and programs are creating substantial expenditure liabilities for the future - roads will need to be maintained, teachers paid, and the sustaining costs o f the Afghan National Army (ANA) and other security services covered. The same will be true o f investmentprograms in sectors like electric power and irrigation. The challenge inthis regard i s two-fold: (i) public spending -both aggregate and at the sector level -needs to be increasingly based on a medium-term vision o f what the public sector can and will deliver, within a sustainable fiscal resource envelope, and (ii) a variety o f funding sources for operation and maintenance (O&M) will need to be developed as appropriate in the sector and program context, getting away from relying primarily on budgetary funds. Figure 1.4: ExpenditureStructureand Financing,2004/05 Estimates 1.3a. Expenditure 1.3b. Financing Operating Domestic Ex~endtures Rewnues ExternalBudget De\elopment Other Direct BudgetSupport ti7 \ - CoreBuCet Project assistance I /&e Budget External Financing Donor executedassistance (USS3,109 million) to the externalbudget Source: MoFpreliminary financial report; Staff estimates. 1.28 Problematic salary structure and weak public administration. A key driver o f spending i s Government staffing and salaries. Afghanistan's civil service i s not large by international standards, but i t ' i s not very effective, and many staff lack required skills and qualifications. Severely eroded by a decade o f hyperinflation and sharply compressed by across-the-board cash allowances, the salary structure i s dysfunctional and inadequate to attract and retain qualified people for managerial and technical positions. The Priority Restructuring and Reform(PRR) program i s intended to provide a short- run remedy by allowing some qualified staff to be paid more on a temporary basis. Key challenges include: (i)implementing an appropriate (and affordable) pay and grading structure across the public sector; (ii)fully implementing merit-based recruitment and promotion practices; (iii)handling appropriately the many civil servants whose shlls and qualifications are inadequate for a reformed public administration; and (iv) restructuringorganizations for greater effectiveness. 1.29 Need to further strengthen fiduciary systems on a sustainable basis. While considerable progress has been made in developing the legal framework and in installing adequate fiduciary systems, these have relied largely on external capacity provided by firms contracted by the Government. Although understandable, even essential as short-term measures, these mechanisms are not sustainable and will have to progressively give way to national fiduciary systems based on core Government capacity. Getting procurement right i s of particular importance in view o f the downside risks associated with weak procurement practices (Chapter 5). Financial management in line ministries i s another area o f weakness where major improvements will be needed. 1.30 Poor sewice delivery. While much has been done to improve public finances in Afghanistan, it will not mean anything ifit does not translate into delivery o f essential public services to the Afghan people, with widespread access and adequate quality (Chapter 8). There has been progress in some areas (e.g. telecommunications, health, and education - based on very different models), but many public services remain grossly inadequate, giving rise to widespread public complaints. The challenge o f 8 improving service delivery is multi-fold and will require capacity building (see ,below), good financing arrangements, sound institutions, and appropriate incentives and accountability frameworks. Service delivery mechanisms need to vary in accordance with the characteristics o f the service concerned; the market structure, ability to introduce competition, etc.; and the appropriate role o f the state. 1.31 Lack of sustainable Government core capacity. The progress.achievedinmost aspects o f public finance management so far i s fragile and excessively dependent on unsustainable external capacity. Thus buildingcore Government capacity that will outstay the current large aid inflows and associated injection o f external capacity will be essential (Chapter 9). Capacity development i s at the heart o f public administration reforms but also involves related aspects like making the best possible use o f external technical assistance and ensuring that it helps build sustainable Government capacity as opposed to substitutingfor or detracting from it. 1.32 Addressing these issues will be discussed in more concrete detail in subsequent chapters, following an overview of the performance o f Afghanistan's PFM system in Chapter 2. Box 1.3: The Challenges of Increasing Aid Effectiveness Given its low revenue base (Chapter 4) and the enormous needs for reconstruction (Chapter 5), Afghanistan very much needs external assistance (Chapter 3). With external assistance above 40-50% o f GDP (Figure 6.1) and above 90% o f public expenditures, the importance o f increasing aid effectiveness cannot be overstated. In 2003, Afghanistan was among only seven countries in the world with an aid to GDP ratio above 30% (with Burundi, Congo, Eritrea, Guinea- Bissau, Sierra-Leone, and Timor-Leste). However, in other post-conflict countries aid has often declined after a few years even though such declines correspond with the period o f time when absorptive capacity and aid effectiveness increase (World Bank, 2003a). This box reviews some o f the critical challenges o f aid effectiveness. First, at the macroeconomic level, these massive inflows of foreign exchange - compounded by very large flows resulting from opium exports - challenge the management o f monetary and exchange rate policies. These inflows increase demand for goods and services. Whereas demand for tradables can be met by an increase in imports, increased demand for nontradables may encounter production bottlenecks, resulting in price increases, pushing up the real exchange rate. There i s considerable uncertainly about this effect, known as "Dutch Disease", but the magnitude o f foreign exchange inflows and capacity constraints in Afghanistan require that this risk be monitored and managed. Certainly, the impact on the skilled labor market inAfghanistan i s marked (Chapter 9). Second, also at the macroeconomic level, i s the issue o f aid predictability. International experience indicates that aid can be more volatile than domestic revenues. This i s challenging in terms o f maintaining a stable macroeconomic environment. Uncertainty about available resources also makes budget preparation more complex. This requires donors to provide as much predictability as possible to the Government, as well as reducing the constraints attached to their support (notably in the form o f project earmarking). On the Government's side, an adequate budget formulation process is required, as well as a cautionary approach to programming pledges from donors (to make the budget realistic, some countries systematically discount pledges by as much as 30%). Chapter 6 explores these issues. The third, related issue i s fiscal sustainability, including the risk o f "aid dependency." Afghanistan's severe fiscal constraint is o f course eased by significant aid inflows - but these inflows are difficult to manage. First, for a number o f decisions it has to make (e.g. pay and grading reform; investment decisions with downstream maintenance costs), the Government needs clarity from donors on their medium-term support. Second, there i s a risk that incentives to raise more domestic revenues could be reduced by massive external assistance, also jeopardizing medium-term fiscal sustainability. This requires the Government to be upfront in preparing a Medium-Term Fiscal Framework (MTFF) and donors to establish a good dialogue around such a framework (Chapter 3). Fourth, massive aid inflows have implications for strategic sector allocations. Chapter 5 reviews these issues, stressing the role o f a Government-led strategic plan to guide budget allocations. This i s especially important for building the State, where the Government can demonstrate its stewardship o f the reconstruction process. This issue is related to the previous one: aid not only could be ineffective if not well spent, it could make Government's budget ineffective if aid forces the Government to use its resources to operate and maintain low-priority or low-quality investments. Fifth, makingaid effective requires adequate delivery and managementmechanisms. At the very minimum, donors should follow the Government's strategic leadership and report to the Government on expenditures. More systematic use o f Government systems can increase aid effectiveness, reducing the burden on the Government's limited capacity (e.g. to prepare dedicated reports, see Chapters 6 and 7) and clarifying accountability. However, the Government does not need to create the capacity to itself deliver a large development program. The modalities of aid delivery (including parallel mechanisms, Chapter 8, and technical assistance, Chapter 9) require attention. Finally, the implications of large amounts o f external assistancefor vulnerability to corruption needto be monitored (Chapter 7). 9 CHAPTER 2. ASSESSING PERFORMANCE INPUBLIC FINANCIALMANAGEMENT 2.1 The challenges outlined at the end o f Chapter 1 need to be addressed in a systematic way that transforms Afghanistan's public finance management (PFM) system into a highly effective enabler o f national development through its beneficial impact on fiscal and budgetary outcomes. Assessing the current and future performance o f the PFM system provides a basis for evaluating progress and identifying areas o f weakness for priority actions. A PFM performance rating system has been recently developed by the Public Expenditure and Financial Accountability (PEFA) multi-agency partnership program, which i s intended to provide an objective, internationally comparable framework for assessing the performance o f a country's P F M system (see PEFA, 2005). This framework i s outlined below, followed by a summary o f the performance o f Afghanistan's PFMsystem against the PEFA indicators. A. A Framework for Analysis andAction 2.2 The PFM system and the PEFA indicators do not attempt to measure fiscal outcomes, the substantive appropriateness o f public expenditure policies and decisions, or the actual impacts and value for money achieved through public expenditures. The PFM system instead should be seen as a crucial enabler for achieving broader development goals and substantive outcomes (including in the public finance sphere), which depend on sound Government strategies, policies, and institutions. The discussion below starts from outcomes and moves to processes and the linkages between them. 2.3 Three levels of budgetary outcomes. The outcomes o f a country's PFM system can be assessed interms ofthree levels, allo fwhichhavewider implications for nationaldevelopment (Figure 1.2): e Budgetary aggregates and overall fiscal discipline (including fiscal sustainability from a medium-term perspective). Desirable outcomes include macro-fiscal balance and low inflation in the short run, and fiscal sustainability or (in the case o f Afghanistan) sustained progress toward fiscal sustainability over the medium term. e Strategic allocation of resources across sectors and programs in accordance with Government priorities to implement national objectives. Desirable outcomes include allocation o f public sector resources across programs based on an agreed development strategy, allocation o f resources within programs based on sound sector strategies, and allocation o f expenditures over time in line with appropriate prioritization (e.g. tahng into account recurrent cost implications of public investments and staffing decisions). e Managing the use of budgetary resources in the interest o f efficient service delivery and value for money. Desirable outcomes include cost-effective service delivery, avoidance o f waste or cormption, proper accountability for use o fresources, etc. 2.4 Critical dimensions of PFM performance. Turning from broad budgetary outcomes to the performance o f the PFM system itself (as "enabler"), PEFA has identified six critical dimensions o f a well-functioning P F M system. Four o f these comprise important parts o f the budget cycle (policy-based budgeting; predictability and control in budget execution; accounting, recording, and reporting; and external scrutiny and audit - see Figure 2.1). Another relates to key cross-cutting features o f the budget process, i.e. comprehensiveness and transparency, and the last covers out-turns o f the PFM system, namely the credibility o f the budget. Figure 2.1 also signals that, given donors' importance inproviding financial support to many developing countries, donor practices have a significant impact on the performance o f the PFMsystem. 2.5 PFM indicator set. Inorder to operationalize the PFMperformance assessment framework, 28 high-levelPFMperformance indicatorshave beendeveloped, each o f themmeasuringperformanceinone 10 o f the six critical dimensions. Three additional indicators o f donor performance have also been developed. As indicated earlier, the P F M performance indicators measure how key parts o f the budget process are worlung, cross-cutting features o f the budget, and whether the budget i s credible (as demonstrated by its out-turn). These are o f critical importance as enabling factors for achieving national ' development and fiscal objectives. Figure 2.1: LinkagesAmong the Six CoreDimensionsof PFMPerformance C. Budnet Cycle D. Donor Practices Policy-based budgeting Predictabilityand A. PFM Out-Turn External Scrutiny Control in Budget and Audit Execution Budget Credibility Recording, and Reporting Source: PEFA (2005, p. 4) 2.6 Overall structure of the PFMperformance measurementframework. Assessment o f the high- level P F M performance indicators provides an evaluation of performance in terms of the six core dimensions o f the PFM system, which in tum leads to an assessment o f the performance o f the PFM system in achieving key desired outcomes (aggregate fiscal discipline, strategic allocation o f resources, and efficient service delivery). A model for a PFMperformance assessment i s presented inPEFA (2005) and applied to Afghanistan inVolume 11,Part 1. B. The PerformanceofAfghanistan's PFMSystem 2.7 The full list o f indicators along with the ratings for Afghanistan, which reflect an assessment o f the PFM system's performance as of June 2005, i s shown in Table 2.1. The assessment i s discussed in detail, rating by rating, inVolume 11,Part I; i s summarized below under the six critical dimensions o f a it PFMsystem. The situation inAfghanistanhas some special features which needto be kept inmindwhen interpreting the ratings. First, some ratings reflect temporary factors (for instance the absence - at the time o fthe assessment-o fa Parliament). Second, some ratings are based on changes made recently, and it is still uncertain whether these changes will be sustained. Third, the ratings assess the current situation in which significant external support (both advisory and operational) is being provided to the Government; in several areas, this raises sustainability issues as the external support will decline over time, especially inrelation to financial management operations. 2.8 Credibility of the budget (Performance Indicators 1-4). Progress has been made toward a credible budget, based on fiscal discipline and improvements in the budget process in recent years. However, the credibility o f the budget i s hampered by optimistic budget projections, attributable to a combination o f lack o f realism at the budget formulation stage and limited capacity to implement the budget, and large deviations between budgeted amounts and actual out-turns. This is in particular the case for the External Budget (accounting for as much as three-quarters o f total public spending), which remains effectively outside the Government's control, further reducing the credibility o f the budget. 11 D-1 Predictability o f Direct Budget Support D-2 Financial information provided by donors for budgeting and reporting on I+ project andprogram aid D-3 Proportion o f aid that is managedby use o fnational procedures Note. I n the rating system 4 is the top performance rating, I the lowest. The ratings measured performance as of June 2005. Broadly speaking, most indicators would have been rated "I" rn 2002 Source Staffestimates based on PEFAframework (PEFA,2005, Annex 1, and table on p 9) 2.9 Comprehensiveness and transparency of information (Performance Indicators 5-10). The Government i s committed to share budgetary information transparently. The annual budget for the national Government i s impressive, with coverage o f most public spending o f the general Government sector, whether implemented by the Government or directly by donors. It i s a useful public document 12 relating the expenditures on specific projects to National Programs. Fiscal, revenue, and expenditure records for the Core Budget are maintained through the budget implementation system based in AFMIS (the Afghanistan Financial Management Information System). Comprehensive reports on budgetary aggregates are produced monthly and are now available on MoF's website. This information i s generally reliable but i s affected by lags in recording provincial operations in AFMIS. Improvements are needed with respect to external audit reports. There are also uncertainties about the quality o f information on the External Budget. Inaddition, fiscal risks which could arise from activities in state-owned enterprises and municipalities are not effectively monitored because the financial information i s unverified and too limited for fiscal risk management inthese areas. 2.10 Policy-based budgeting (Performance Indicators 11-12). Some progress has been made toward the Government's key objective o f making the national budget the central instrument o f policy and reform, but there are still important constraints. The full implications o f policy decisions, including for fiscal sustainability, are not taken into account due to lack o f capacity in Government to develop policy, agree on trade-offs between investments and recurrent costs, and match resources with policies. Multi- year planning is also constrained by the complexity o f aid coordination. Budgeting for recurrent and investment expenditures i s carried out separately in parallel, exacerbating fiscal sustainability risks from recurrent expenditure implications o f investment operations. The budget process is orderly and well understood within the Government, but it needs to be linked more closely with the Government's strategy; more widely owned politically including by the Cabinet and the Parliament; and more reflective o f sector issues and with more engagement by sector agencies, provinces, and civil society. 2.1 1 Predictability and control in budget execution (Performance Indicators 13-21). Muchprogress has been made in improving the implementation o f the budget (inparticular with respect to the allotment process and cash management), but there i s a need to enhance the predictability of funding for service delivery units. Line departments at the provincial level often cannot predict with accuracy the funds that will be available to them because o f delays incommunicating with the Mustoufiats (provincial offices o f MoF) and the complexity o f the allotment process; this has led to cash rationing at the discretion o f the Treasury and the Mustoufiats. Many adjustments are made during the year, in particular through allotment transfers and the use o f contingencies. N o information i s available at the local level (e.g. primary schools) on actual resources available. 2.12 A particular weakness inthe budget execution cycle i s related to revenue collection. The system i s not very effective, as demonstrated by the large gap between revenues that could be collected and actual collections. In addition, despite some progress with taxpayer registration, the control framework on the revenue side remains weak. Nevertheless, clarification o f the legislation and strengthening o f financial management (including banlung arrangements to ensure a steady flow o f collected revenues to the Treasury Single Account, TSA) have generated some improvements inrevenue performance. 2.13 Much progress has been made in implementing the control framework, as reflected by a significant increase in the eligibility rate measured by ARTF. There i s a division of duties between the line ministries, which approve expenditure, and MoF, which makes payments. The controls also include effective cash management, allotment control on budget uses, and reconciliations of records o f the Treasury and the line ministries. Significant weaknesses remain, however, compounded by the lack o f internal audit to provide feedback on performance. Further improvements in the internal control framework therefore are needed both for payroll and non-payroll expenditures and for revenues. These arrangements are also fragile since they rely on external operational support in Treasury and the review by the ARTF Monitoring Agent (Box 6.2). The recent approval o f a new Public Finance and Expenditure Management (PFEM) Law will helpin further strengthening the control framework. 2.14 Most procurement at present i s carried out under standards agreed with donors. The procurement carried out under national rules has been subject to an outdated regulatory framework which lacked adequate institutional support for effective enforcement. There has been insufficient competition as the basis for contract awards and no guarantee of value for money. The Procurement L a w recently approved provides the basis for modernization o fthe procurementframework. 13 2.15 Accounting, recording, and reporting (Performance Indicators 22-25). Considerable improvements have been made in this area. The recording o f expenditures and revenues has been strengthened by implementation o f a computerized system (AFMIS) in MoF. Accountability o f line ministries and M o F to the Cabinet is supported by frequent reporting on budget implementation, but more reporting and accounting capacity in line ministries i s required. Significant progress has been made toward consolidation o f bank accounts under the TSA system. However, much remains to be done to improve standards o f accounting, frequency o f bank reconciliation, and usefulness o f reporting for line ministries' management. 2.16 External scrutiny and audit (Performance Indicators 26-28). Government accountability generally is sought through the publication o f audited financial statements on the state budget and specific donor funds. The Auditor General i s worlung to international auditing standards with operational support from an audit advisor. However, the final audit o f the 2003/04 annual statements has not yet been released. In addition, the absence o f a Parliament (until recently) has limited the extent o f external scrutiny o f the budget and its performance. 2.17 Donor practices (Donor Performance Indicators 1-3). Donors have mobilized a considerable amount o f support for Afghanistan's budget, mainly through the ARTF. In particular, support to the Government's recurrent budget through ARTF has been large and predictable. However, direct budget support still accounts for only a very small percentage o f total external assistance to Afghanistan, and the proportion o f aid that i s managed by use o f national procedures remains low, with most aid executed by donors through their contractors. Financial information provided by donors on donor-executed assistance activities, while reasonably comprehensive, i s subject to delays, usually relates to disbursements rather than actual expenditures, and may suffer from inaccuracies and inconsistencies 2.18 Concluding summary. Afghanistan's ratings against the PFM performance indicators generally portray a public sector where financial resources are, by and large, being used for their intended purposes as authorized by a budget which i s processed with transparency and has contributed to aggregate fiscal discipline. The expenditure and financial position o fthe resources under the authority o f the Government are reported reliably in an understandable format, although there i s some uncertainty with respect to revenue reporting. Performance regarding the allocative efficiency o f spending across programs and the efficiency o f operations i s not as good, however. Given that performance most likely would have been rated "1" on all dimensions in 2001 or 2002, this assessment highlights the significant achievements o f the last four years. Inmost dimensions, the ratings are now comparable to other low-income developing countries (for instance A f i x a n countries). Areas o f relative weakness include tax collections (indicators 13-15), as well as legislative oversight (27-28) due to the absence o f a Parliament hitherto. Payroll control (18) and predictability o fbudget-support (Dl) are areas o frelative strength. 2.19 Using the ratings o f this report as a baseline, future progress inimproving the performance o f the PFM system can be monitored. However, the nature o f the indicators and ratings means that measurable progress in terms o f changes in ratings i s more likely to occur over periods o f a year or longer than on a short-term basis. Moreover, since the ratings assess the system's performance with significant external support (both advisory and operational), questions arise about the sustainability o f current levels o f PFM performance. This may mean that further improvements will be manifested in maintenance o f current ratings based on sustainable national capacity and with declining levels o f external support. 2.20 The ratings against the PFM performance indicators also provide an assessment for the donor community o f the potential capacity o f Government systems (currently supported by external assistance) to implement the operations which are now occurring outside Government systems. 2.21 Chapters 3-7 o f this report discuss inmore detail key development and budgetary outcomes and aspects o f the PFMsystem, frequently referring back to the PFMperformance indicators. After Chapter 8 (on the organization o f public service delivery) and Chapter 9 (on institutional reforms and capacity development), Chapter 10 returns to the PFM performance measurement framework and puts forward a roadmap for making improvements (see Table 10.1). 14 CHAPTER 3. PROGRESSINGTOWARD OVERALL FISCAL SUSTAINABILITY 3.1 A key outcome o f a well-performing PFM system i s fiscal discipline: the Government spends only what it can afford. Affordability inthe short runrelates to the budget balance, i.e. keeping the fiscal deficit within available normal financing. Inthe medium-term, fiscal sustainability i s critical - spending decisions and the time-path o f expenditures need to be affordable from a multi-year perspective. Failing to keep public expenditures within affordable limits has dire consequences (Box 1.1). Risks are especially great inpost-conflict situations like that faced by Afghanistan. A large influx o f aid, combined with lack o f domestic capacity and low domestic revenues, can result in an unaffordable pattern o f expenditures if not well-managed. While international support gives Afghanistan an opportunity to temporarily soften its national budget constraint, a visionary medium-term framework i s needed to ensure that the country moves toward a sustainable fiscal position. This chapter first reviews the structure o f Afghanistan's budget and recent fiscal trends. It defines fiscal sustainability and puts forward the concept o f a Medium-Term Fiscal Framework (MTFF). The main expenditure drivers related to fiscal sustainability are then analyzed - mobilizing domestic revenues i s discussed in Chapter 4. The final sections look at financing options and management o f fiscal risks, and present several illustrative medium-term fiscal scenarios. A. Fiscal Structure and Trends BUDGET STRUCTUREAND RECENTTRENDS 3.2 Fiscal Structure. Afghanistan has an unusual fiscal structure reflecting its historical legacy (e.g. traditionally low revenue mobilization), the longperiod o f conflict (e.g. civil service wages greatly eroded by hyperinflation and severely compressed by fixed allowances), and current reconstruction activities (very high spending) and priorities (e.g. massive expenditures on security). More specifically, as shown inTable 1.1andinthe Statistical Appendix: Public expenditures are extraordinarily high. Total budgetary expenditures in 2004105 were equivalent to 57% o f GDP (excluding the drug economy from the denominator). This is accountedfor by extremely high development spending. Unlike in most countries, the development budget ($2.8 billion in2004105) dwarfs the operating budget ($0.6 billion). Operating expenditures are in line with international patterns. At 9% o f GDP in 2004105, the operating budget was split roughly 70-30 between wage and non-wage spending. There are however large recurrent expenditures in the development budget. These include health services, technical assistance, salary payments, and grants, among others (Figure 6.2). There is underspending on non-wage operation and maintenance requirements. Ineducation, for instance, there i s almost no spending at the school level other than teacher salaries (textbooks, chairs, etc. are provided in-kind,for the most part directly by donors). Most spending occurs outside Government channels. Around three-quarters o f expenditure in 2004105 was donor-executed, inthe external budget, with very limitedGovernment oversight. The security sector is a major driver of overall spending. Security spending has been growing rapidly and accounts for 39% o ftotal expenditure A disproportionate share of public spending occurs in Kabul. Most notably, only 30% o f non-wage O&M expenditures are made outside Kabul (see Figure 7.1). Domestic revenues are very low. They comprise only about 4.5% o f GDP (likely the lowest ratio inthe world for a sizable country) and pay for only about 8% o f total expenditures. 15 0 Thefiscal deJicit (before grants) is extremely high and entirely aid-financed, mostly through grants. The non-resort to domestic financing o f the deficit i s salutary and shows strong fiscal discipline, but this level o fbudget deficit and foreign financing clearly are unsustainable. 3.3 Fiscal trends. Assessing fiscal trends based on only three years o f data since major conflict ended in late 2001 i s not easy, and there have been both significant volatility and step-changes in public expenditures. Nevertheless some important developments have occurred: 0 Budgetary expenditures have increased sharply during 2002103 to 2004105, with operating expenditures almost doubling and development spending growing even faster Revenues also have risen rapidly, more than doubling duringthe same period. 0 Since revenues started out from a much lower base, although their growth was more rapid than that o f recurrent expenditures in percentage terms, the absolute gap between recurrent spending and revenues more than tripled. 0 This means that externalfinancing of recurrent expenditures has been one of the fastest growing components of the budget (Figure 1.3). 0 Thefiscal deficit (before grants) has increased sharply, reflecting large inflows of external assistance. Although aid predominantly has consisted o f grants and thus does not have implications for debt sustainability, it will eventually decline and shift more toward a mix o f grants and loans, with implications for the fiscal deficit and external debt. 3.4 Actual development expenditures (in both Core and External Budgets) have fallen far short o f budget targets, by margins of around half in recent years (see Figure 3.1 for 2004/05). These shortfalls are attributable to several factors, including the existence o f unfundedprojects inthe development budget (a practice now changed for the Core Budget), confusion sometimes between commitments and disbursements, optimistic projections o f disbursements, insufficient project screening during budget preparation, large contingencies in the budget, and implementation constraints (security issues, lack of capacity to prepare sound budget proposals, excessive optimism about implementation delays, etc.). Figure3. 1: BudgetImplementation (2004/05 actualsas YOof bud! !t) - 10 20 30 40 50 60 70 80 90 100 Source: MoF budget document, Staff estimates. IMPLICATIONS 3.5 The first implication of the fiscal structure and trends outlined above i s that there is a great and urgent need to mobilize more domestic revenue. Only an intensive effort to improve tax collection - with sound policy measures and administrative improvements leading to higher compliance - will result in sustained rapid revenue growth and a sustainable fiscal balance over the medium term (Chapter 4). Robust revenue growth will also give donors greater confidence to continue providing external assistance 16 to finance the operating budget. However, taxation efforts need to be non-distortionary and should not impose an excessive burden on the private sector which would adversely affect economic growth. 3.6 A second implication is that all spending decisions have to be considered in the light of their multi-year implications. Even if substantial operating costs can be paid for by external assistance inthe short run, they will become a claim on the Government's own resources later. 3.7 A third implication is that significant external assistance will be required for Afghanistan's operating budget for some time, and for the development budget over the longer term. Projecting budgetary aggregates forward, taking into account likely growth o f expenditures and even with aggressive measures to promote rapid growth o f domestic revenues, there will continue to be a gap between the Government wage bill and total domestic revenues for a few years to come, and for some years longer a gap between total recurrent expenditures and revenues. 3.8 A fourth implication is that the budget needs to be realistic. Overspending leads to budget cuts during the year and can harm fiscal and program management, undermining development effectiveness. Underspending, particularly what has been seen in relation to the civilian development budget in recent years, can undermine public confidence and discourage donors from providing funding. The Government can aggressively hold donors to their pledges/commitments o f external financing, and "stretch" targets can be set for accelerated domestic revenue mobilization (as has been the case inthe IMF Staff Monitored Programfor the current year), but the Government should plan its budget on a conservative basis. 3.9 Fifth, key drivers of public expenditures must be carefully planned and monitored. As discussed in Section C, these include notably the Government wage bill, security spending, public investments, expansion o f social services, andrecurrent costs inthe development budget. 3.10 Finally, there i s a need to increase expenditures outside Kabul. The Kabul-centric pattern o f spending and small non-salary budgets inthe provinces and districts, as well as difficulties in accessing and spending non-salary allotments, negatively impact on service delivery (Chapter 8). B.A Frameworkfor AchievingFiscal Sustainability FISCAL SUSTAINABILITY INTHE CONTEXT OFAFGHANISTAN 3.11 What is meant by fiscal sustainability? The standard definition o f fiscal sustainability is that a country's fiscal balance and underlying trends are such that, in a steady state, the ratio o f total public debt and debt servicing to macroeconomic aggregates like GDP i s not increasing over time.' Key parameters determining fiscal sustainability under this definition include the projected economic growth rate, the revenue growth rate and its elasticity with respect to GDP growth, expenditure growth, and the interest rate at which the Government can borrow to finance its fiscal deficit (because the cost o f financing the deficit forms a component of budgetary expenditures). Of particular interest i s the level o fprimary deficit (i.e. the fiscal deficit excluding interest payments) that i s sustainable over the medium term. A number o f countries have adopted laws with specific rules related to fiscal sustainability (based for example on a targeted debt level, overall deficit, or primary deficit). 3.12 Afghanistan does not fit neatly in this conceptual framework. The country i s nowhere near a "steady state" in terms o f fiscal aggregates. Financing costs are negligible since the bulk o f the fiscal deficit i s grant-financed, and the Government has adopted a conservative stance that forbids overdrafts or borrowing from the Central Bank. On the other hand, as seen earlier the fiscal deficit (before grants) i s enormous as a ratio to GDP, and the levels o f external grant assistance seen inrecent years will eventually decline over time. Domestic revenue and its growth are far too l o w to support fiscal sustainability as defined above inthe absence of massive grant financing. Under these circumstances the size o f the fiscal deficit in relation to GDP would be a better (negative) indicator o f fiscal sustainability, but it also i s highly imperfect for this purpose and would vary greatly with the level o f external grant financing. 'Seeforinstance Heller (2005). 17 3.13 A practical definition of fiscal sustainabilityfor Afghanistan. This report analyzes a simpler and more practical fiscal sustainability objective for Afghanistan: whether and at whatpoint in time the country will be able to cover its recurrent expendituresfrom domestic revenues, with an intermediate target o f when domestic revenues will be able to cover the government wage bill, a primary component o f recurrent spending. The logic i s that the recurrent budget to a large extent represents the core costs of maintaining government functionality and hence needs to be covered by reliable domestic revenues. Issues related to the sustainable level o f borrowing and primary deficit will become more relevant over the longer term, after this critically important objective i s achieved. This definition o f fiscal sustainability does not imply dual budgetingas between operating and development budgets.' On the contrary, there i s a strong argument for moving toward fully integrated budgeting o f recurrent and development expenditures, recognizing that, if the existing grant financing o f the development program dries up, fundingneeds for thispurpose would still remain, puttingheavypressure onthe budget. 3.14 By this definition, Afghanistan falls far short o fbeing ina sustainable fiscal position. In2004/05 domestic revenues were equivalent to only 48% o f recurrent expenditures, and the budget for 2005/06 has a similar target (49%). The simple mathematics o f growth from different bases means that domestic revenue will need to grow much more rapidly than total recurrent expenditures for a number of years in order to "catch up". MEDIUM-TERMFISCAL FRAMEWORK: ITMEANS,WI TWORKS WHAT H O 3.15 The concept o f an MTFF i s straightforward. An MTFF sets forth a multi-year fiscal path that reaches, or at least makes progress toward, a sustainable fiscal position for the country. In addition to projections (typically over 3-5 years) o f fiscal aggregates including revenue, expenditure, deficit, and key expenditure components (e.g. the wage bill), the MTFF lays out the assumptions behind the projections, including the fiscal strategy and set o f policy actions envisaged to achieve the fiscal objectives embodied inthe MTFFtargets. Projections inthe absence ofpolicies and actions to achieve the forecasts would be an empty exercise. Both quantitative projections and the policy package would be reviewed and updated annually. Possible elements o f an MTFFfor Afghanistan are outlined inBox 3.1. 3.16 It should be noted that the MTFF is more rudimentary than other medium-term frameworks that are commonly talked about and sometimes applied. A Medium-Term Budget Framework (MTBF) introduces a sectoral perspective to the multi-year framework. At a more advanced stage, a Medium- Term Expenditure Framework (MTEF) would contain the same aggregate and sectoral projections but would also seek to fully cost sectoral strategies and reconcile projected costs with sectoral resource envelopes. While this provides more detailed guidance for the allocation o f present and future spending across sectors (see also Chapter 5), the technical demands o f an MTEF are much greater than those for an MTFF. Thus anMTEFis amore distant prospectthan anMTFFwhich canbeprepared fairly quickly. 3.17 While the technical analysis that underpins the MTFF i s important, political buy-in and a meaningful linkage to the annual Budget are essential for an MTFF to be effective. The former i s facilitated by Cabinet review and approval o f the MTFF, along with presentation o f the MTFF to Parliament, either in conjunction with the annual budget or as part o f a national strategy document (e.g. the ANDs). Chapter 6 reviews inmore detail the budget formulation process. 3.18 M o F has prepared an initial MTFF for Afghanistan which, together with the 2005/06 mid-year budget review, was discussed and approved by the Cabinet inOctober 2005. This initial step represents a good start in developing a medium-term fiscal vision. The key to success will be steadily improving the MTFFover time and ensuring that it is closely linked with boththe national development strategy and the annual budget process, which i s clearly the Government's intention. This will in particular require a gradual improvement inthe capacity to cost policies (Box 3.1). *For example, the UK has a policy rule of limiting capital expenditures to the level of borrowing, but this is a fiscal discipline tool and doesnot detract from unifiedbudgetingwhich the UKpractices(Sarraf, 2005, p.4). 18 Box 3.1: Possible Elements of a Medium-Term Fiscal Framework for Afghanistan An MTFF couldhavetwo main components: (i)fiscal tablewiththree-orfive-yearprojectionsofdomesticrevenues,expenditures(bothaggregateandkey a itemslike the wage bill), the fiscal deficit, and sourcesof financing, and (ii)aconcisedescription of thepolicyactionswhichsupporttheprojections (i.e.whichwillcontributetoachieving the fiscal targets). Within this broad structure, the MTFF for Afghanistancould seek to addressquestionslikethe following. Domestic revenues: What are the sources of expectedincreasesin domesticrevenues? These could include: (i)Expansion ofthe base on which domesticrevenuesare collected. Expansionof the tax base i s mostly driven by (non-agricultural) economic growth, so the MTFF would needto specify assumptionsregardingeconomic growth. (ii) An increase in revenue collection efficiency. Important measures that have been introduced to increase tax collection efficiency include the creation o f the Large Taxpayer Office (LTO) and implementation of the Customs reform. What are the medium-termcollection objectives from these measures? Does the Government intend to implementadditionalmeasuresto increaserevenuecollectionefficiency? Ifso, what measures? (iii) measures to increase domestic revenues. Doesthe Governmentplanto introducenewmeasuresto New increasedomesticrevenues? Which ones? What are the expectedimpactsinterms of incrementalrevenue? Civilianwage bill: What policy actions explainthe projectedcivilianwage bill. The civilian wage bill depends on: (i)Thenumber ofcivil servants. WhatistheGovernment'sstrategyregardingthesizeofthecivilservice. What public services does the Government plan to deliver itself? For example, has the Government endorsed the purchaser-providermodel in health? Has the Government decided that the overall size of the civil service should notbe increasedexcept for teachers? (ii)Wagestructure. WhatsalarystructuredoestheGovernmentplantoimplementinthecoming3-5 years? Isit the salary scale which is going to be derived from the on-going Pay and Grading Review? What are the Government's intentionswith respectto other salary increases (annual, ad-hoc, other)? (iii)Short-term programs. How many people are expectedto be hired under short-termprograms created to attract skilled Afghans at higherrates on short-termcontractsand the PRRandPRR superscaleprogram? At what cost? Absorption of non-civilian expenditures into the Budget: Does the Government plan to absorb non-civilian expenditures (including on the new Afghan National Army) into the Budget? How will these expenditures be financed? What trends are expectedinthe security sector wage bill inparticular,and its key components? Developmentexpenditures: The Government is preparingits interim AfghanistanNational Development Strategy (ANDs). How does the Governmentplanto ensure that the ANDs is translatedinto annualnationalbudgets? Fiscal Deficit: Domestic revenues are expected to fall short of total (ordinary and development) expenditures, leadingto afiscal deficit. How is the Governmentplanningto finance fiscal deficits? How muchfinancing does the Government expect from the recurrent andinvestment windows o f the ARTF? How muchfrom other sources? C.ManagingExpendituresfrom a Medium-termPerspective 3.19 Inaddition to domestic revenue (discussed inChapter 4), total expenditure and key expenditure components comprise the cwe o f an MTFF. The main expenditure drivers for Afghanistan include the Government civilian wage bill, non-wage operating costs, development expenditures, and security spending (Figure 3.2). Medium-term issues around each o f these expenditure categories are summarized below. Underlying these issues are key questions about the role o f the state - which services it should deliver and finance itself, which it should deliver but finance only partly (relying in part on cost recovery), which it should finance but deliver through the private sector, and for which services both financing and delivery should be left to the private sector, with Government regulation as appropriate. These strategic issues are also discussed inChapters 5 and 8. THEGOVERNMENT NON-SECURITY PAYROLL ($187 MILLIONIN2004/05Q 3.20 Managing and containing the growth o f the Government payroll, while ensuring that the salary structure i s adequate to attract and retain qualified staff (Chapter 9), is a major challenge for Afghanistan. Inmany countries, the Government's wage billtends to be a source ofupwardpressure on expenditures in both the short run and over the medium-term, since there are often political pressures to increase recruitment into the civil service as well as to raise pay (at least to keep up with inflation). Afghanistan has maintained a remarkable record o f fiscal discipline during the past several years, with limited Excludes police, army, and justice. The total wage bill including security was $377 million in 2004105 (Table 1.1). See Volume 111, Chapter2. 19 recruitment into the Government except for teachers and security forces. However, there have already been pressures for ad-hoc pay increases, and pressures for higher pay and recruitment can only be expected to increase with the transition toward more "normal" politics and the formation o f Parliament. Another fiscal risk factor i s pensions, whose cost to the budget i s still relatively low despite a recent increase, but which can be expected to rise over time particularly with increasing retirements. Figure 3.2: Key Drivers of Afghanistan's PublicExpenditures,2004/05 (YOof total budget) //cly~6;oage Development- recurrent, 20% --- investment. 32% Source: MoF reports; staff estimatesfor development budget classification into "recurrent" and "investment". 3.21 With average salaries equivalent to 3-3.5 times average per-capita GDP (Table 3.1), current pay levels in Afghanistan's civil service are not particularly highcompared to other countries. The challenge i s to decompress the salary structure to ensure that pay at higher levels i s adequate to attract and retain qualified staff, while keeping the overall wage bill aff~rdable.~Hyperinflation in Afghanistan during the 1990s almost completely eroded the real value o fbase pay, which was offset (but only inpart) by giving a cash allowance that was the same for everyone and comprised the bulk o f overall compensation. As a result, in 2001Afghanistan inherited what was probably the most compressed civil service compensation structure inthe world, with total pay varying by less than 20% betweenthe top and bottom grades. A pay increase in 2003 somewhat decompressed the salary structure, increasing the variation to 55%, still much lower than inother countries. Inaddition, various interimreform schemes provide higher pay for selected positions (Figure 3.3), butthese affect only a small part o f the civil service. Note: for comparison purposes, this table refers to the civil service, including justice and police (but excluding the army). Source: Afghanistan: staff estimates for 2004 (a/ unweighted average pay scale for general civil service (3.1) or 2004105 budgeted wage bill (3.5) divided by the non-drug GDP); other countries: Schiavo-Campo et a1 (1997)for variousyears in the 1990s. There i s a risk, seen in many countries, that political pressures (including from civil servants) result in a compressed pay structure in the interest o f "equity" across employees, while fiscal constraints keep average and top salary levels relatively low, leaving the Government unable to recruit and retain qualified staff inmanagement and senior technical positions. 20 3.22 Afghanistan's civil service size i s not large by international standards, as can be seen from Table 3.1, even talung into account that basic health services have been almost entirely contracted out (Chapter 8). But its composition raises a number o f challenges (Chapter 9). First, Afghanistan's civil service has not been very effective, in running itself or in delivering services to the Afghan people. Thus the developmental returns on the Government wage bill may be relatively low. This means that some retrenchment - possibly limited to early retirement or natural attrition - may be unavoidable in the medium-term. Inparallel, recruitment needs to be managed to focus on genuine needs for qualified staff, e.g. teachers. Finally, there i s excessive concentration o f staff inKabul, where a third o f civil servants are located, and correspondingly relatively low staffing levels in the provinces and especially in rural areas. This pattern is problematic for service delivery, although it is gradually changing in particular as more teachers are being recruited. Figure 3.3: Pay Scales ($/month) Lower and upper bounds of the scales. This includes base salary and an estimate of cash allowances See VolumeIII,Chapter 2. 3.23 Thus a modest increase inthe size o f the civil service may be expected inthe future, as well as pay increases at least inline with growth o f per-capita GDP. As a result the ratio o f the wage bill to GDP i s likely to moderately increase in coming years. This trend will be accentuated by the need for the operating budget to absorb substantial salary costs currently inthe development budget (paras 3.30-3.32)? 3.24 These considerations underline the need for careful management o f the Government payroll. First, there needs to be a strategy for managing the payroll from a medium-term perspective, as opposed to merely responding on an ad-hoc basis to pressures as they arise. Second, the Government will need to implement its strategy carefully. Even if the wage bill grows in a programmed manner as envisaged in the Securing Afghanistan 's Future (SAF) projections, this must not be perceived to soften the hard fiscal constraints within which the Government operates. All sources o f financing for the payroll - domestic revenues, ARTF contributions, other budget support - are scarce, precious resources for Afghanistan's reconstruction which have a high opportunity cost in terms o f alternative uses. Third, maintaining strict control over staffing levels will be critical, even as disciplined recruitment o f needed slulls proceeds. Fourth, transfer to the operating budget o f salary costs currently covered by donors inthe external budget needs to be handledcarefully, with transitional mechanisms to avoid unmanageable fiscal outcomes. And finally, the Government needs to provide full information to donors to secure their continued support. The SAF projections were well-received when presented to the international community in April 2004, but there needs to be a clear statement o fGovernmentpolicy onthe wage billandupdatedprojections. Inadditionto these recurrent costs, reformingpublic administrationwill haveadditionalcosts. The Government, initsSecuring Afghanistan's Future report (2004), estimated these costs at about $200 million over 2004-2010, including a large amount for public buildings, systems, and equipment; a small component for retrenchment and retraining; and a component for technical assistance(to restructureministries, developpolicies, etc.). 21 NON-WAGE OPERATING EXPENDITURES MILLION IN 2004/05, EXCLUDING SECURITY) ($96 3.25 In contrast to the Government payroll, non-salary expenditures tend to be unduly low in many countries, reflecting a variety o f factors including a political preference for investments and salaries and the Government's response to short-run fiscal constraints -non-wage O&M is much easier to cut inthe short run than salaries. More fundamentally, underspending on non-wage O&M may reflect previous decisions with respect to public investment, public sector service delivery institutions, and staffing that turnout to beunaffordable. This results inchronic shortfalls inO&M spending, which means that public assets are deteriorating due to lack o f maintenance (e.g. roads); they cannot fully function due to shortages o f fuel or parts (e.g. power stations); service delivery networks' ability to deliver services is impaired due to lack o f needed inputs (e.g. teaching materials for schools, medicines for basic health facilities); and Government employees lack the resources to do their jobs (e.g. fuel for vehicles). 3.26 Also constraining non-salary recurrent spending at the local level in Afghanistan i s the low allocations for such expenditures provided to provinces and districts, and the difficulties lower levels o f administration face in accessing and utilizing their non-salary allocations (Figure 7.1). This appears to reflect Ministries protecting non-salary expenditures at the central level, as well as various constraints affecting allotments and payments (Chapter 8). As the amounts involved are relatively small, such imbalances between Kabul and the provinces/districts can and should be rectified relatively easily, with monitoring through the budget execution process. 3.27 Turning to the future, fiscal challenges in meeting non-wage O&M requirements will be daunting. A simple rule o f thumb would suggest that every dollar o f public investment leads to at least ten cents o f additional recurrent costs on an annual basis. This suggests that the public investment o f 2004/05 would generate more than $100million o f additional O&M costs eachyear. The highway sector, with a current rehabilitation investment program o f more than $1 billion, is an important example. Annual maintenance costs, not including those o f future highway investments envisaged by the Government, are expected to be in the range o f $40 million per year (as a first step, the 2005/06 budget includes $10 million for road maintenance). Another important example i s electric power, where access to the grid i s extremely low (estimated at only 10%) and large increases in capacity will be required to support private sector development and home consumption, yet even now the mainthermal power plant in Kabul requires $40 million worth o f fuel per year, currently being provided by a donor in-lundoutside the budget. 3.28 Although the examples cited above point to cost recovery as an essential way forward (Chapter 8, Section B), non-wage O&M will nevertheless be a rapidly growing claim on budgetary resources that will need to be well-managed. Clearly, inadequately funding O&M requirements i s damaging to service delivery outcomes and development objectives, but the root o f the problem usually lies in public investment decisions made years earlier. This highlightsthe need for such decisions to take into account their future consequences for the budget(Chapter 5). THEDEVELOPMENT PROGRAM ($300 MILLIONCORE, $1,467 MILLIONEXTERNAL 2004/05) IN 3.29 Public investment. At some 30% o f GDP, Afghanistan's non-security investment program i s very large, reflecting massive requirements for reconstruction and availability o f large amounts o f external financing. These requirements had been reviewed by the Government in its S A F report, presented to donors in Berlin in 2004. This report projected a capital investment of some $26 billion between2004 and 2010. As the development budget i s fully funded by external assistance and most o f it i s donor executed (only about one-sixth o f civilian public investment spending flows through the Treasury), it i s easy to make the mistake o f thinking that it does not have fiscal consequences for Afghanistan. However, the very large O&M requirements for future years generated by current public investments (inboth core and external budgets) cannot be ignored (Chapter 5). 3.30 Recurrent costs in the development program. As noted earlier, Afghanistan's development program does not consist entirely or even predominantly o f public investment - there i s a large 22 component o f recurrent costs estimated very roughly at 38% (and more than 75% in the Core DevelopmentBudget, see Figure 6.2). These include: 0 Basic health services, which are contracted out to NGOs and funded by both Core and External development budgets. The estimated cost in 2004105 exceeded $100 million, which includes salaries and other recurrent costs (e.g. vaccines) as a large component. 0 Technical assistance, which predominantly consists o f recurrent costs (payments to consultants etc.). This i s truly large: OECD data estimate TA flows at $400 million in 2003, but only a fraction is on budget (the 2004/05 budget identifies around $200 million as stand- alone capacity-building, technical assistance, or feasibility studies projects). 0 Salalypayments and top-ups paidby donors through the External Budget in several sectors. 0 Transfer schemes, such as the National Solidarity Program (NSP, $130 million in 2004/05) which provides block grants to communities; most o f these are used for small investments but likely include some recurrent costs as well. 0 Teacher training, textbook printing,and other smaller items o frecurrent costs. 3.3 1 Since the funding of these recurrent costs by donors cannot be expected to be permanent, and in many cases i s committed by donors only on a yearly basis, such expenditures will sooner or later become a charge on the Government's core fiscal resources. The magnitudes o f some o f these recurrent costs are sufficient that they pose a large, even unmanageable fiscal risk over the mediumterm. Moreover, as they are usually funded and executed by donors on a fragmented basis with a focus on implementing the activities themselves and often without competitive procurement, there i s a tendency for costs to be higher than underlying conditions in Afghanistan, let alone fiscal affordability considerations, would dictate. This aggravates further the fiscal risks involved and could inflate costs more generally. 3.32 There are no easy answers to the dilemmas posed by recurrent costs in the development budget, as for the most part they are responding to urgent and high-priority reconstruction needs. At a minimum such recurrent costs need to be fully understood and their fiscal implications taken into account by the Government and donors. This requires accurate reporting by donors and a dialogue with the Government on the levels o f activities and costs involved, as well as the time horizon o f the donors concerned in providing assistance. It i s also quite legitimate for the Government to raise concerns about the unit costs and procurement methods used, as keeping unit costs within reasonable limits will help ensure fiscal sustainability o f the activities concerned (see for instance Chapter 8 on the cost o fhealth services). SECURITY SECTORSPENDING($1,328 MILLIONIN2004/05, OFWHICH $292 MILLION CORE) 3.33 The security sector i s a major driver o f expenditures with a large impact on Afghanistan's fiscal position over the medium term. Therefore it needs to be fully incorporated inthe MTFF and analysis o f the national budget. In view of the critically important and complex security problems that Afghanistan faces, it i s not surprisingthat massive resources (including Afghanistan's own domestic revenues as well as donor funds) are being spent on security. However, with the bulk o f security sector expenditures occurring through the External Budget, decisions on funding and spending have occurred in a fragmented way, often somewhat in isolation from the requirements and strategy for the security sector as a whole as well as from medium-term fiscal affordability issues. The lack inthe past o f an agreed national security strategy has made it more difficult for the Government to provide oversight and coherence across the various security subsectors and activities. There i s also a risk that pay levels established in the security services will exert upwardpressure on general pay levels inthe civil service (Figure 3.3). 3.34 Some salient characteristics o f Afghanistan's security sector are summarized in Box 3.2. The analysis o f security sector expenditures conducted as part o f the PFM Review (discussed in detail in Volume V) has identified the following key issues from a PFMand development perspective: 0 The need for an integrated security strategy and policy framework, without which sound decisions on programs - e.g. force sizes, equipment - and expenditures cannot be made. The 23 national security strategy, preparation o f which i s at an advanced stage, will facilitate the necessary strategic coordination and decisionmaking including on public expenditures. 0 Very high donor-executed spending through the External Budget (Figure 3.4), which renders achieving coherence o f spending across and within subsectors much more difficult, and also raises questions about the fiscal sustainability o f the security sector inthe medium term when donor fundingmay decline. 0 Dfficulties in coordinating and prioritizing security sector expenditures and actions, reflecting the above problems and fragmentation o f decision-malung across donors. These factors make it more difficult for the National Security Council (NSC) to fully carry out its mandated strategic, leadership, and coordination functions, although the situation in this regardhas improved in2004 and 2005. 0 Concerns about the growing security sector wage bill and pressuresforfiscally unaffordable salaly increases, as well as the fiscal sustainability of security sector staffing levels. The ANA salary structure, determined apparently without reference to fiscal constraints or pay elsewhere inthe civil service, has set a precedent which the police and other sectors aspire to and which will be fiscally costly (Figure 3.3). There are also doubts about whether recruitment practices are beingregulated by staffing establishment size limits. 0 Slow implementation of priority public administration reforms and capacity building in security sector management and oversight institutions, which have lagged behind the development o f security forces like the ANA. This imbalance, and also the lack o f progress insome security subsectors (e.g.justice), carry avariety ofrisks for the future. 0 The need to develop good governance and sound and sustainable financial management practices in the security sector, fully in line with and integrated with national budget processes and fiduciary provisions. Figure3.4: ForecastSecurity Sector SpendingAgainst ProjectedRevenues 1,800 .II totalbudget Secunty- +Revenues - BaseCase -A- Revenues Low Case - Source Volume V, Chapter4. 3.35 The most important message emerging from this analysis o f security sector expenditures i s that the security sector needs to be appropriately integrated into all aspects of Afghanistan's PFM system: medium-term fiscal programming (MTFF), annual budget formulation, staffing and payroll management, procurement, financial management and controls, and external scrutiny, among others. There i s no justification for treating the security sector as separate or sacrosanct, and not subjecting it to budgetary and fiduciary processes. Specific recommendations include the following. 3.36 Review the Fiscal Implications of Ongoing Security Sector Reforms: Quickly enhance the information base on security sector expenditures, and conduct an integrated security sector-wide fiscal 24 review to forecast fiscal implications o f security sector development and reforms over the medium- and longer-term. This would serve as a basis for revisiting security sector force levels and for discussions with donors on longer-term financing. It would also facilitate policy- and program-based budgeting o f security sector expenditures with a medium-termperspective. Box 3.2: Salient Features of Afghanistan's Security Sector As discussed in detail in Volume V, Afghanistan's security sector has some important characteristics with fiscal and development implications. First, the historical legacy of conflict wasfragmentation into regional and local militias, capture of policing and military functions by non-legitimate actors, and erosion of both formal and traditional justice systems, also with capture by non- legitimate authorities. Building the state since 2001 therefore has entailed reconstituting and reforming the security forces under legitimate oversight by the civil authorities. Second, the immediate security threats faced by Afghanistan are variegated and disputed by different stakeholders. Some observers see the continuing insurgency in the south and east and conflict with Taliban/Al Qaeda forces as the most important security problem. Others are most concerned about criminality and lack o f rule o f law at the local level, with associated capture of policing and justice functions by warlords and commanders. Still others are concerned about the limited management capacity o f civil authorities and inadequate oversightlaccountability o f security forces. The drug industry is widely seen as a major cause and consequence o f insecurity. An agreed national security strategy, currently at an advanced stage o f preparation, will facilitate reaching consensus on different security threats and their relative priority. Third, the overall size of the security sector does not seem unreasonably high by international standards. This includes a planned size o f the ANA o f 70,000 (not all o f them combat troops) and a planned size o f the Afghan National Police (ANP) o f 62,000. However, the appropriateness o f the size and balance between different security forces can be questioned in the context o f Afghanistan's particular situation and security needs. For example, the role o f the ANA in meeting different external and internal threats would needto be clarified inorder to assess the appropriateness o f its size. Fourth, AfghanistanS security sector nevertheless is costly and may be unaffordable. The lack o f short-run budget discipline (and lack of policies and programs to form the basis for sound budgeting), and the involvement o f different "lead donors" and line ministries in decision-making on force levels and expenditures, have resulted in limited efforts to control costs. Over the medium term, it is extremely doubtful whether the sustaining costs of Afghanistan's security sector at planned force levels can be absorbed by Afghanistan's national budget (Figure 3.4). This implies that difficult trade-offs will need to be made inprioritizing expenditures, means o f achieving efficiencies in security-related expenditures will need to be identified, and the possibility o f multi-year donor financing commitments for the security sector should be explored. Fifth, the progress in developing different parts of the security sector has been uneven. After a slow start initially, the formation and expansion o f the ANA have moved forward rapidly. Progress in forming the new ANP has been much slower, and reform and capacity building in the justice sector have lagged far behind, with adverse implications for legal reforms and for a pro-private sector business climate. Moreover, within sectors there has been a disproportionate focus on building up security forces and less attention to strengthening the key management and oversight institutions, including the MinistryofDefense and especiallythe MinistryofInterior. Sixth, security andfiscal issues are closely interlinked. For example, low domestic revenue (discussed in Chapter 4) i s both a consequence and a cause o f insecurity. Reconstruction costs tend to be higher due to insecurity, while on the other hand 3.37 Further Develop Policy Making Capacity in the NSC and Finalize the National Security Strategy: The NSC and National Security Advisor are tasked with coordinating policy development and overseeing integrated strategic planning across all security institutions. Adequate capacity i s required for these purposes. An integrated security sector strategy and policy i s needed to determine acceptable longer-term costs of security institutions and to appropriately prioritize and sequence spending across and within subsectors (e.g. justice and police versus ANA). As in the case o f the ANDs as a whole (discussed below), the National Security Strategy i s being developed by Afghan stakeholders, with full Government ownership (across the ministries and agencies involved), and will need to gain wide buy-in within and outside Government. This will require a meaningful and substantive consultationprocess. 3.38 Treat Security as an Integral Part of the National Development Strategy currently under Preparation: The ANDs can be used as an opportunity to ensure that the security sector both is accorded its due importance as a development issue for Afghanistan and that the broader developmental and fiscal perspective i s injected into the security sector strategy. 3.39 Implement Priority Administrative Reforms and Develop Security Sector Management and Oversight Capacity: Support reforms enhancing human resource capabilities in the Ministries o f 25 Defense, Interior, Justice, Counter Narcotics, and National Security Directorate. A particular priority i s developing solid financial management capacity and practices in these line ministries. Strengthening the control framework, notably with the development o f an internal audit function, i s a pri~rity.~ 3.40 Rigorously Enforce Staffing Size, and Set Salary Increases Based on Fiscal Absorption Capacity: It i s vital that the agreed formal staffing establishment sizes be enforced. Moreover salaries, like non salary costs, must be subject to the normal rigors o f public expenditure management and fiscal discipline -the security sector wage bill mustremain affordable. 3.41 Role of donors. Donors play an even greater role inAfghanistan's security sector than they do in other sectors. Specifically, donors can contribute to improving public finance management inthe security sector by: (i) supporting development o f an agreed national security strategy and corresponding strategies and policies for the individual sectors (defense, police, justice, etc.); (ii) reassessing their views about force sizes and cost requirements in the light o f fiscal constraints; (iii)helping enhance the capacity o f oversight actors (such as the Auditor General, legislature) and economic managers (Ministry o f Finance Budget Office) to address issues relating to financial management in the security sector; (iv) channeling more o f their assistance to the security sector through the Core Budget; (v) for resources not channeled through the Core Budget, ensuring that timely and accurate information i s made available to the budget authorities, and maximizing use o f internationally acceptable procedures e.g. for procurement; (v) helping key security ministries strengthen their budget formulation and budget execution processes; and (vi) discussing medium- and longer-term availability o f external resources for the security sector. 3.42 Implementing these recommendations and other initiatives to improve security in Afghanistan, which i s a necessary condition for the country's development, undoubtedly will take time. However, the need to get a handle on the fiscal sustainability issues inthe security sector i s urgent, as there is a risk that decisions on force levels, pay, non-salary spending, equipment, construction, etc. that are being made or implemented today will be unaffordable for Afghanistan in the future. Dialogue with donors, including on possible medium-term financial support to Afghanistan's security sector, also i s urgently needed. D.FinancingReconstructionand Managing FiscalRisks MANAGING BUDGETFINANCINGOVER THE MEDIUM TERM 3.43 Afghanistan has been relying entirely on external assistance to finance its fiscal deficit, including a large operating deficit as well as the entire development budget. This approach has made a great deal o f sense in a context where reconstruction needs are enormous, revenues are extremely low, and large amounts o f external assistance have been available (mostly grants), including substantial amounts for financing recurrent expenditures. The "no overdraft rule" has been an effective tool for inflation control and macroeconomic stability. While substantial external assistance can be expected to continue over the medium-term, it may decline inmagnitude, and inparticular external financing for the recurrentbudget i s expected to be phased out over time. Thus financing o f the fiscal deficit will need to become more diversified, probably includingprudent levels o f domestic borrowing inthe mediumterm. 3.44 Looking forward, risks to budget financing, external financing in particular, may be exacerbatedbypoor performance in revenue mobilization or expenditure management. Future external assistance, particularly the part going through the Core Budget, i s likely to be positively correlated with domestic revenue effort and improved expenditure management. This suggests that mutually-reinforcing trends may lead to a high scenario whereby good revenue performance and expenditure management (money gets well spent) encourage high inflows of external assistance, further enlarging the resource envelope, or a lower scenario whereby poor revenue performance and weak expenditure management One option would be to focus the newly developingInternal Audit Department of MoF initially on the security sector. The vocation of this department, as per the new Public Finance and Expenditure Management Law, is to cover the whole of Government. Given that a mechanism (the ARTF Monitoring Agent) is temporarily in place for most o f the civilian recurrent budget, this departmentcould initiallyprioritize its work programtowardthe security sector. 26 (including inability to fully spend resources) result in flagging external assistance, severely constraining the resource envelope. Thus progress toward fiscal sustainability could occur under a highly constrained situation or under a developmentally very positive scenario, enabled by rapid revenue growth and more generous external assistance. Achieving the positive scenario requires careful management and further underscores the need for strong revenue mobilization and effective utilization o f resources. 3.45 Within the supply of external assistance, Afghanistan needs to carefully assess what level of external debt it can prudently incur, to serve as the basis for a sound debt management strategy. The International Monetary Fund (IMF, 2005) has undertaken a debt sustainability analysis for Afghanistan, although handicapped by lack o f data and uncertainties about initial conditions (most notably about the disposition o f Soviet-era Russian claims on the order o f $10.8 billion - equivalent to about 190% o f 2004/05 GDP excluding the opium economy). The main findings o f the analysis are as follows: 0 Assuming forgiveness of all pre-existing bilateral claims (including Soviet-era debt), strong economic and revenue performance, and a very high degree o f concessionality in new debt (Le. along the lines o f IDA concessionality), Afghanistan can engage in modest new borrowing while keepingits debt sustainability indicators within safe limits. 0 However, it i s clear that the bulk o f external assistance for some time to come will need to be inthe form ofgrants. 0 Any combination o f weaker economic or revenue performance, failure to secure full forgiveness o f all bilateral claims, a substantial share o f loans intotal external assistance, or a lower degree o f concessionality o f new borrowing could lead to a situation where Afghanistan's debt and debt service burdenbecome unsustainable. 0 Given the uncertainties, a very prudent external borrowing strategy i s called for, relying only on modest amounts o f highly concessional loans and ruling out commercial or export credits for the foreseeable future. STRENGTHENING MUNICIPAL FINANCES 3.46 Afghanistan's municipalities constitute the only semi-autonomous level o f government and should play a very important role inurban service delivery (Chapter 8). They have their own budgets, but levels o f municipal taxes and fees are determined by the central Government, which also approves municipal budgets. Since municipalities are supposed to live within their means (i.e. expenditures are constrained by resources), they do not directly contribute to fiscal deficits. However, there are important issues related to future municipal expenditure liabilities associated with public investments (which often are decided on and executed by central ministries rather than by municipalities themselves). Similar potential issues arise with respect to staffing, although there i s no indicationthat municipalities have gone on hiringsprees, probably due to their relatively hard budget constraints. There are serious weaknesses in the basics o fpublic financial management at the municipal level -budget formulation, budget execution, accounting, and auditing. And there are widespread perceptions o f corruption inmunicipalities. 3.47 Managing the fiscal risks associated with municipalities will require: (i) much better information on municipal revenues and expenditures; (ii) clear delineation o f their roles and responsibilities (in particular avoiding confusion and overlaps vis-a-vis central ministries); (iii) assessment o f the future an municipal expenditure liabilities that are being created by public investments, including those by the MinistryofUrban Developmentand other line ministries; (iv) pro-activerevenue mobilization efforts by municipalities (focused on a small number o f tax sources that have been widely demonstrated by international experience to be appropriate municipal revenue sources); (v) strengthening municipalities' core PFM capacity; (vi) integrating their operating and development budgets; and (vii) improving governance. Progress inthese areas will require significant capacity development (see Chapter 9). 3.48 Kabul Municipality i s much larger than all other cities in Afghanistan (accounting for something like 15% o f the total national population) and currently has a special status (more-or-less akm to that o f a ministry) inthe overall structure o fthe Government. Thus it isparticularly important that budget data for 27 Kabul Municipality be made available on a timely and transparent basis, and that the Municipality's PFM system - including the budget process -i s improved (Volume 111, Chapter 5). MANAGING NATIONAL ASSETS INCLUDINGSTATE-OWNEDENTERPRISES' 3.49 Public investments create state assets, whose sound management i s very important for the effectiveness o f the PFM system. Often (though not always) public assets are managed by State-Owned Enterprises (SOEs) involved in delivery o f public services or other commercially-oriented activities. Government departments also are custodians o f various types o f public assets. While many SOEs in Afghanistan are defunct and were involved in activities that would be more appropriately handled by the private sector, there are some activities (e.g. public utilities like electricity and water supply) that are likely to remain primarily in the public sector for the foreseeable future. SOEs and other commercially- oriented activities carried out by the state also entail fiscal risks which need to be managed. In many countries the financial losses, wage costs, and other liabilities o f SOEs constitute a serious risk to fiscal and macroeconomic stability. 3.50 The SOE sector in Afghanistan does not appear to be very large, and many SOEs are defunct, with their main asset being land. Information compiled by the SOE Department o f MoF indicates that there are about 70 SOEs, with total employment o f 20,000. Although financial reporting by SOEs and their supervisory line ministries i s poor and incomplete, it appears that, particularly in 2004/05 when controls on SOE spending were tightened up, budgetary transfers to SOEs (subsidies and payments o f salaries to SOE workers) have been small. Transfers from SOEs to the budget also appear to have been very small. However, a very rough estimate o f public investments in SOEs through the (core and external) development budget since late 2001 i s in the range o f $150 million, o f which 7040% has gone to the electricity and water utilities. Analysis by the SOE Department suggests that 10 SOEs would appropriately stay under Government ownership, 20 could be liquidated, and 40-plus privatized. 3.51 The figures on SOEs are far from complete, however, leaving out some very important enterprises and their assets. Certainjoint venture enterprises are excluded from the list even though their total Government ownership i s well above 50% (e.g. Ariana Airlines, and some public sector banks). In other cases, activities and associated assets are under Government departments without formal creation o f an SOE. Moreover, both SOEs and their supervising line ministries generally lack financial management capacity, and as in the case o f the budget in general, their sources o f financing tend to be fragmented, including investments sponsored by line ministries and also those channeled through the External Budget. 3.52 It is very important to ensure that SOEs do not become a significant fiscal risk factor for Afghanistan and that those which are defunct or which are engaged in activities which should be in the private sector are appropriately disposed o f in a manner that encourages private sector development, with reasonable provisions for the affected SOE employees. Key priorities for monitoring and managing the fiscal risks associated with SOEs include: Improving the legalframework and information base on SOEs and other state assets. It i s strongly recommended that the existing legislative framework within which SOEs operate be reviewed. It may be better to repeal the SOE Law and corporatize all SOEs so that they can be governed under the same legislativeprovisions as other companies. An inventory of state assets should be compiled, including SOEs and other public sector activities. Regular reporting of keyjkancial variables by SOEs (including those with less than 100% Government ownership), and their compilation for use in monitoring and policy decisions. This will require capacity development inSOEs, supervisingline ministries, and MoF. Careful policies on and monitoring of SOE investments (including avoiding creation of new SOEs), which should include SOE investments occurring outside the Core Budget (Box 3.3 suggestspractical guidelines for responding to fundingproposals related to SOEs). 'Thisdiscussionon SOEs is based on Volume 111, Chapter 4, and also on Volume IV, Chapter 4 on the mining sector, which looks at SOEs inthis sector. 28 Ensuring transparent financial transfers between SOEs and the Government budget (tax payments and subsidies). e A sound legal andpolicyframework for privatizationhestructuring of SOEs. The Cabinet has recently adopted a privatizationpolicy and has amended the SOE Law to facilitate divestment o f SOEs. The privatization process will need to be clear and transparent; investments in SOEs to be divested as well as creation o fnew SOEs should be avoided. Box 3.3: RecommendedGuidelinesfor Respondingto BudgetProposalsfor SOEs First, use of public funds (whether domestic resources or external assistance) should be limited to existing SOEs that will remain under public ownership for some time. MoF should first confirm whether the Government has full ownership o f the firm concerned, to avoid potential claims by private parties. MoF should then make use o f the SOE Department's prioritized list, inwhich the SOEs that will remain under public ownership are identified. Specifically: . Fresh investments in utilities could be considered, as utilities are likely to remain under public ownership for . some time, so long as these investments are accompanied by necessary reforms and improvements in financial management in the public utilities concerned, including a sound business plan. No investment should be made in SOEs on the privatization list. Labor or financial restructuring could be considered, but other investments (e.g. in new equipment or technology) are not appropriate as international . experience demonstrates that Governments usually do not recover such investments through a higher sale price. No investment should be made inSOEsthat are to be liquidated. No investment should be made to create new SOEs because this would send an adverse signal to the private sector, and creation o fnew SOEs would generate fiscal risks and potential liabilities for the Government. 1 Finally, some rehabilitation investments may be considered for SOEs that will stay inthe public sector. Second, SOEs should have an acceptable minimum level of financial oversight. A certain level o f administrative reform in the parent ministry o f the SOE is required, so it has the capacity to supervise the SOE's operations. There also needs to be a minimumlevel o f financial management capacity (accounting and reporting in particular) in the SOE itself. Beingable to regularly assess the cash flow o f the SOE is important, and ideally, audited accounts should be available. Ifthese conditions are not met, major investments should be heldback until improvements are made. Third, MoF in general should not respond positively to proposals for budget funding of SOEs operating costs. In exceptional cases the budget may have to cover operating costs o f certain SOEs in the short run, but such funding when required should be tied to improvements incash flow, e.g.through increasedcost recovery. Finally, in reviewing SOE funding proposals, MoF should consider the valuefor money within a limited resource envelope, and also the multi-yearfiscal implications. The benefits and value o f SOE proposals can be compared with other budget proposals, and the estimated economic rates o f return should be adequate. Proposals with substantial downstream fiscal implications needto be turned down untilthere are sustainable financial arrangements. E.Medium-termFiscal Scenarios 3.53 InOctober 2005, the Government adopteditsinitialMTFF. As indicatedinChapter 6,anMTFF should be an evolving document, updated at the beginning o f each budget cycle. Most important, the MTFF needs political buy-in, widespread ownership within the Government, and strong linkages to the development strategy and the annual budget process. Three illustrative medium-term fiscal scenarios are presentedinthis section, based on a simple spreadsheet model, to show some o f the trade-offs involved in medium-term fiscal decision-making and the implications o f some o f the key issues discussed earlier in this chapter. For simplicity, the macroeconomic framework is the same inall three scenarios.8 3.54 Inthe base case, itis assumedthat a combinationofprudent expenditure management, effective efforts to raise revenues, and sustained donor assistance allows Afghanistan to achieve a number o f its policy objectives in the next decade: implement a pay reform, finance the wage bill by 2007, bring the ANA on budget (by 2010 in this scenario), increase O&M expenditures to operate and maintain new investments, and close the operating fiscal deficit before grants by 2012 (Figure 3.5a and Table 3.2). Revenue collection i s expected to increase by one-half percentage point o f GDP per year as a result o f The key assumptions are: (i)dynamic growth and stable inflation (based on IMF projections, with the exchange rate assumed constant); (ii)an increase inthe wage billdriven by recruitment o f teachers, pay reform, and progressive inclusion o f ANA salary costs on budget; (iii)an increase in O&M expenditures to operate new investments (based on data from the SAF report); (iv) an increase in investment expenditures, notably in the short run (based on data from the SAF report); and (v) a gradual move o f development expenditures from the External to the Core Budget. More details are given inthe Statistical Appendix (Table A.6). 29 forceful implementation of the customs and tax administrative reform program and some expansion o f the tax base through the implementation o ftax policy measures already decided. Even though revenue would only rise to 8.3% of GDP by 2012113 (still very low by international standards), in absolute terms it would quintuple from less than $300 million in 2004/05 to more than $1,500 million in 2012/13 (Chapter 4). The wage bill is projected to remain largely stable as a proportion o f GDP. Iti s assumed that external assistance, primarily grants, could finance the gap between revenues and operating expenditures (more than $2 billion over seven years - compared to $0.7 billion duringthe first three years o f reconstruction). Figure 3.5: Three Illustrative Fiscal Scenarios (% of GDP) a. Base Case +Domestic Revmues +Total Operating Evenditures wages and Salaries 12 10 8 6 4 2 02/03 03/04 04/05 05/06 06/07 07/08 08/09 09/10 10/11 11/12 12/13 b. High Case 12 10 8 6 4 1 - '02/03' 03/04' 04/05' 05/06 '06/07' 07/08' 08/09,09/10, 10111'11/12 '12/13' c. Low Case 2 I 02/03 03/04 04/05 05/06 06/07 07/08 08/09 09/10 10/11 11/12 12/13 Source: MoF estimates; Staff projections. 3.55 However, some recurrent expenditures currently included inthe Core Development Budget - e.g. health services - and inthe External Budget - e.g. costs of runningelections - would still not be financed by domestic revenues. Based on simple assumptions, it is estimated that in 2004/05 revenues covered 48% o f the Government's Core Recurrent Budget but only 28% o f all recurrent costs. With similar assumptions, inthe base case revenues would cover only three quarters o f all recurrent costs by 2012/13. 30 3.56 Turning to the high-case scenario, with additional effort collection o f domestic revenue could reach 11% o f GDP toward the end o f the projection horizon (2011/12), i.e. at the lower end o f the range observed in other low-income countries (Figure 3.5b and Chapter 4). This could enable the Government to speed up implementation o f its civil service reform program, with a positive impact on budget implementation inthe outer years. Inthis scenario, the Government could almost pay for its wage bill in 2006/07 and would be in a position to cover all operating expenditures by 2010/11. The external financing required for operating expenditures would be reduced to $1.4 billion over six years, freeing additional external assistance to finance development programs. 3.57 On the other hand, under a low-case scenario, if domestic revenues grow more slowly than inthe base case the Government i s likely to delay the implementation o f pay reform and limit the increase in O&M spending (Figure 3.5~). This would have a negative impact on the Government's capacity and on the effectiveness o f the investment program. Moreover, the Government would not be able to finance its wage bill before 2009/10 and would still run an operating deficit in 2012/13. This scenario probably would prove unsustainable, as external assistance i s unlikelyto finance the considerably larger fiscal gap ($2.7 billion over the next eight years). Consequently, the Government might seek to finance part o f its budget deficit through domestic borrowing, with significant potential adverse implications for macroeconomic stability. The Government might also for example delay the necessary recruitment o f teachers, slowing progress on education. There would also be a risk o f long-term dependency on external fundingfor the ANA, and/or pressures to reduce its size. 3.58 In sum, these projections highlight the importance o f domestic revenue mobilization for Afghanistan to achieve steady progress toward a sustainable fiscal position. They also indicate how more rapid revenue growth can create fiscal space for critical expenditures in support o f reforms and capacity building. Conversely, lower revenue growth can be expected to have adverse development and macroeconomic impacts, which may be exacerbated if lower revenues lead to loss o f donor confidence and financing. This exercise also highlightsthe benefits o f a simple MTFF-to outline the fiscal path that the Government has embarked upon, manage the reform process, review affordability issues, and engage 2002103 12003/04 2004105 1 2005/06 2006/07 2007108 1 I I lion) 2008109 1009110 2010111 2011/12 I 1 12012113 actuals pro& BASE CASE DomesticRevenues 269 379 532 677 853 1,033 1,220 1,393 1,593 CoreExpenditures 874 1,226 2,116 2,360 2,434 2,614 2,316 2,588 2,597 of which OperatingExpenditures 558 678 892 1,094 1,251 1,431 1,535 1,576 1,585 ofwhich wages 282 374 460 562 649 149 845 900 940 948 FiscalDeficit Before Grants 606 047 1,584 1,684 1,581 1,581 1,156 1,195 1,005 of which operating 289 299 361 417 398 398 315 183 (8) HIGHCASE DomesticRevenues 269 393 569 770 1,046 1,315 1,618 1,954 2,225 Core Expenditures 874 1,226 2,l 16 2,398 2,471 2,651 2,407 2,588 2,597 of which O p t i n g Expenditures 558 678 892 1,131 1,288 1,468 1,566 1,576 1,585 ofwhich wages n/a 282 374 460 562 685 785 882 930 940 948 FiscalDeficit Before Grants 213 350 606 833 1,546 1,628 1,425 1,336 789 634 373 of which operating -- 213 241 289 285 323 361 242 154 (52) (378) (640) LOWCASE DomesticRevenues 129 208 269 353 443 527 651 804 970 1,125 1,303 ofwhich CoreExpenditures 342 558 874 1,226 2,046 2,254 2,285 2,406 2,117 2,350 2,431 of which OperatingExpenditures 342 449 558 678 823 988 1,102 1,224 1,276 1,338 1,418 of which wages 374 460 545 631 694 754 794 854 931 Source: MoF estimates; staffprojections. 31 CHAPTER 4. MOBILIZINGDOMESTIC REVENUES 4.1 As seen in Chapter 3, sustained rapid growth of domestic revenue is a central element o f Afghanistan's state-building and reconstruction agenda. International experience suggests that the larger the tax gap (the difference between the taxes actually paid and what should be paid according to existing laws and statutes), the more radical are the changes needed. With a tax gap on the order o f 60%, Afghanistan needs to adopt a comprehensive strategy including revamping o f the tax administration in order to obtain significant improvements in compliance. Based on the more detailed analysis inVolume 111,Chapter 1,this chapter first briefly outlines the key features o fAfghanistan's existing revenuebase. It then reviews progress and summarizes key issues in improving revenue policy and strengthening tax administration. The last section presents some illustrative medium-term revenue projections. A. Revenue Structure 4.2 While in industrialized countries the revenue to GDP ratio i s typically around 45-55%, for the least developed countries it i s closer to 20%. With revenue at 4.5% o f GDP in2004/05, Afghanistan i s an outlier even in this group (Table 4.1). Tax revenues (3.4% o f GDP) are only one-fourth the average for low-income countries. Table 4.1: Central Government Revenues (% of GDP) Countrylregion GDP per capita Total Total tax IUSD) l / revenue revenue Afghanistan (FY2004105 est.) 253 4.5 3.4 Sub-SaharanAfrica 765 19.7 15.9 Asia and Pacific 1,447 16.6 13.2 Low-income countries nia 18.0 14.9 Low- middle income countries n/a 21.8 15.8 Non-OECD average nia n/a 15.2 Selectedc o u n ~ e saverage nia 17.4 14.8 Pakistan 420 16.4 12.4 Iran 1,700 20.5 8.6 India 470 18.5 14.3 Kazakhstan 1,520 22.3 18.4 Kyrgyz Republic 290 19.1 15.0 Azerbaijan 710 20.5 19.4 Uganda 240 11.5 10.7 Rwanda nia 10.4 n/a Sources. Government Finance Statistics(IMF), Keen(2004). Notes: Figuresreferto counny averages bctween 1997-2001,where available. For each revenueclassification, only countries for which dataare availableare included I / Country datarefers to GNIpercapita(World Bank) 4.3 Afghanistan's economy has many features commonly associated with a l o w tax base, including: (i)extremely low level o f development; (ii)large informal sector implying a narrow tax base; (iii) a the dominance o f agnculture which i s hard to tax; and (iv) capacity constraints hindering the ability o f the Government to collect taxes and o f taxpayers to comply with tax regulations. Revenue mobilization i s further complicated by the large opium economy that cannot be taxed directly, the need to consolidate Government control throughout the country, and heavy reliance on aid funds that are exempt from taxation. In addition, Afghanistan historically had very l o w domestic revenue mobilization - in the 1970s the tax to GDP ratio was around 7%, one o f the lowest in the world. Afghanistan's medium-term revenue potentialtherefore i s likely to be toward the lower end o f the 11-14%o f GDP range. 4.4 Domestic revenues have grown significantly over the last two years, starting from an extremely low base. In2002/03, they reached $129 million, equivalent to 3.2% o f GDP. In 2003/04, revenues reached $208 million (4.5% o f GDP), and in2004/05 they rose to an estimated $269 million, also 4.5% o f 32 GDP (Table 4.2). The 2005/06 budget projected a modest increase inrevenues to $333 million (4.7% of projected GDP). Table 4.2: DomesticRevenueCollection 2002/03 2003104 2004/05 2005/06 Est. Est. Est. Budget DomesticRevenue(percentage GDP) 3.2 4.5 4.5 4.7 DomesticRevenue(Afs) 5,864 10,168 12,812 16,150 DomesticRevenue($) 129 208 269 333 DomesticRevenue(percentage shares) 100 100 Tax Revenues 62 75 Taxes on income, profits and capital gains 4 8 Domestic taxes on goods and services 3 9 Taxes on international mde and transactions 53 57 of which: Importduties 52 54 Non Tax Rmenues 38 25 Admin. fees & charges, non-industrial sales 35 20 CentralMinistries(percentage shares) 29 21 Provinces 71 19 of which (inpercent of total revenue), Herat 41 34 Kandahar 7 11 Nangarhar 7 12 Kabul 5 5 Balkh 3 9 Source: MOF, Fundstaff estimates. 4.5 Revenue is concentrated in a small number of taxes (Table 4.2). As in many less developed economies, import duties are the main source of revenue because they are relatively easy to monitor and collect, comprising over 50% of the total (Table AS in the Statistical Annex). Box 4.1 summarizes the major tax instruments currently being utilized inAfghanistan. Box 4.1: Summary of Major Tax InstrumentsinAfghanistan Customs duties constitute the largest source o f revenue in Afghanistan, as in many developing countries, largely because they are relatively simple to administer. While domestic taxes tend to grow in importance as countries develop, these are more difficult to administer and will be a long-termobjective for Afghanistan. In the interim, the average import tariff should be kept relatively low, with a limited dispersion of ratesto reduce arbitrary and excessive effectiverates of protectionas well as to minimize opportunitiesand incentives for corruption. Export tuxes shouldbe avoidedsince they tend to cause an outflow of resourcesfrom the export sector to less efficient uses, compromisinggrowthobjectives. Recenttax reformshave loweredtax rates for exporters. Income Tuxes. Incometaxes ideally shouldbe levied on a global basisandinclude all forms of income. The corporate and top personal tax rates have been aligned at 20%, relatively low by international standards. Accompanying this move, new administrative provisions were added to the law, including introduction of self-assessment principles; provision for tax rulings so that investors have greater certainty about tax rules; and strict secrecyrules to heightenthe accountability of the tax administrationin dealingwith sensitive businessinformation.The plannedrestorationof the wage withholding tax and the simplification of personal tax rates, with only two nonzero personalrates, should help improve compliance. The level of exemption has been set to effectively remove most civil servants from taxation (which should ensure a progressivesystem). Sales tuxes. These taxes should be consolidated as a simple and broadly based tax on final consumption. However, Afghanistan currently has a complex and cascading system o f Business Receipts Tax (BRT) based on turnover and numerouspresumptive(fixed) taxes for smaller traders, which is far from ideal. While a single rate Value-added-tax (VAT), with crediting provisions and zero-rating o f exports, is one of the most efficient taxes, it would be administratively too complex to implement at this time. However, over the mediumterm Afghanistan should expand the BRT toward a more broad-basedconsumptiontax that broadensthe tax base andavoids cascading. Excise tuxes. There are currentlyno excisetaxes in Afghanistan. 4.6 Over 93% of total revenue is collected either by Central Ministries (mainly in Kabul) or injust Jive of Afghanistan's 34 provinces: Herat, Nangarhar, Kandahar, Balkh, and Kabul. Four ministries account for 88% o f total revenue reported by central ministries: MoF, Ministryof Communications which derives revenue from the sale o f telephone services, Ministry of Commerce from commissions on imported goods, and Ministry of Interior from motor vehicle registrations. 33 4.7 I n addition, municipalities collect some revenues. These include, for example in the case o f Kabul, taxes on real estate (a rental tax and a levy on property notionally intended to cover the cost o f sanitation services); various duties for business licenses, contracts, property deeds, markets, etc.; fees for services (e.g. cleaning o f septic tanks o f residential properties); rents; and proceeds from sale o f land etc. The revenue yields o f these taxes and levies are low, however: in the case o f Kabul, total municipal revenues in 2004/05 are roughly estimated at around $6 million - in a center o f economic activity with a populationof around three million. 4.8 There are many small "nuisance" taxes. With around 90 active taxes -half o f them generating less than Afs 1 million ($21,000) during 2004/05 - the administrative costs o f some taxes most likely outweigh the revenue gain. The income tax law provides for a variety o f presumptive taxes on small businesses, usually referred to as fixed taxes, as well as the cascading turnover tax (BRT). The law also identifies 170 tax categories o f business establishments. Removing many o f these small "nuisance" taxes and simplifyingthe tax collection systemwill be cost-effective and helpfulto the private sector. 4.9 There are also illicit taxes and instances of double taxation or ambiguities in tax laws. The business community has noted numerous cases where illicit taxes add to the cost o f business. For example, additional "transport taxes" may be levied on the owner o f the goods, and provincial or municipal authorities may charge other ill-defined taxes. Municipal taxes also need to be streamlined as noted above, and there are cases o f double taxation, for example on rental property income as the central government introduced a new services tax on rental income in 2004 (inmost counties, property income i s taxed at the local level). Businesses have also complained that the legal code i s often unclear; for example with respect to tax-deductible expenses, with varying interpretations. This can lead to disputes and corruption. B. ImprovingRevenuePolicyandTax Administration 4.10 M o F i s responsible for revenue policy and revenue collection, with functions divided between the Afghanistan Customs Department (ACD) and the General Presidency o f Revenue (GPR). The A C D historically has been a much more distinct entity, with functions including collection o f customs duties and management o f a customs service. The GPR has traditionally been in charge o f domestic taxation policy, with actual collection being handledby the Mustoufiats. GENERAL APPROACH 4.11 The ideal tax system raises in a fair manner the required revenues while minimizing distortions. It should (1) minimize interference with individual consumption, saving, and investment decisions; (ii) be relatively simple, transparent, and rules-based to encourage compliance and discourage corruption; and (iii) be stable and predictable to reduce uncertainty. 4.12 The Government'sstrategy shouldfocus on improving compliancefor taxes and locations with the highest revenue yields. This also implies getting rid o f numerous low-yield "nuisance taxes" and curbing illicit levies as well as avoiding double-taxation and confusion. Some taxes may have good potential over the longer run, but large revenue yields in the short- to medium-term should not be expected. Mineral exploitation has good potential and i s likely to generate tax and other revenues once the recently approved Minerals Law i s implemented (see Volume IV, Chapter 4). Similar potentials may also exist in hydrocarbons, for which a Hydrocarbons Law is being prepared. However, these potential revenue sources will take time to develop. 4.13 Agricultural taxes have a long history as a religious tax (zukut). Subsequently taken over by the State and collected at the local level, agricultural tax was once a significant source of revenue. However, it has fallen into disuse over a long period o f time, and any attempt to reinstate it is likely to be severely hampered by the security situation and prevalence o f opium, as well as widespread poverty inrural areas. Nonetheless, given the importance o f amculture to the economy - and to avoid distortions that might 34 result from exempting an entire sector permanently from taxation - it may be advisable to revitalize agricultural and landtaxes over time. 4.14 Tax policy and tax administration reform will be closely linked. Afghanistan's strategy for revenue mobilization needs to take into account the very weak state o f institutions and limited implementation capacity. Tax policy needs to provide a modern framework within which the tax administration can be built. International experience suggests the following guiding principles for tax administration reform: (i)reducing the complexity of the tax system by reducing the number of taxes, harmonizing rates, reducing exemptions, and eliminating illicit charges and double taxation; (ii) encouraging taxpayers' voluntary compliance with simplified self-assessment procedures and better taxpayer education; (iii) differentiating the treatment of taxpayers by their revenuepotential, commonly through the creation o f a Large taxpayer Office and risk management strategies; and (iv) ensuring effective management of tax reforms. 4.15 A major improvement was the establishment in 2004 of the Treasury Single Account (TSA), which integrates Government accounts into an account or set o f linked accounts through which the Government collects revenues and transacts payments. The apex o f the system i s a single account held by the Ministry o f Finance (Treasury) at the Central Bank. The TSA i s designed to ensure that all Government revenues, wherever collected, are transferred to the central Government. 4.16 Civil service reforms are required to develop a motivated and skilled cadre of revenue staff (Chapter 9). The customs and tax administrations face unique challenges in having to collect funds for the Government, which increases the risk o f corruption. Therefore a key component o f the customs and tax reformprograms i s the development o f professional (dedicated) tax and customs services. This entails specialized training programs and institutions, a career development structure within the service, and as and when appropriate, a performance-based remuneration system. These reforms are currently being undertaken as part of MoF's PRR program CUSTOMS ADMINISTRATION REFORM 4.17 Customs administration reform will continue to be critical for success. During 2004, the A C D implemented important measures to: (i) simplify the tariff system (Box 4.2); (ii) introduce simplified customs clearance documentation and procedures; and (iii) establish an effective system o f customs brokers to facilitate a speedier transit process. Ongoing reforms are extending ACD's operations in the provinces and include: (i) a modem customs code; (ii) improvements in the customs infrastructure major throughout the country; (iii) gradual introduction o f a computerized recording and management system (ASYCUDA, a standard international program for customs, Box 7.2); (iv) re-organizing and re-building o f human capacity in the central and regional customs offices; and (v) legislation to establish effective enforcement mechanisms, combined with an effort to reduce illegal charges and consolidate and rationalize the number of charges. 4.18 The Customs administration faces many challenges, including the poor security environment, lack o f experienced managers, poorly trained staff, inadequate facilities and equipment, and outdated and cumbersome operating policies and procedures. Customs regulations reportedly are not applied consistently throughout the country, and the lack o f an automated system undoubtedly impacts negatively on revenue performance. DOMESTICTAX ADMINISTRATION REFORM 4.19 The domestic tax base i s likely to remain relatively narrow for some time, and reforms being implemented will focus on large taxpayers. The current set o f reforms includes: (i) the adoption o f a tax reformpackage in June 2004; (ii) the creation o f a Large Taxpayer Office (LTO) and restructuring o f the headquarters and provincial revenue organizations; (iii) the nationwide roll-out o f a new Taxpayer Identification Number (TIN); and (iv) the consolidation o f tax legislation amendments to address the excessive use o f tax holidays, exemptions, and "special agreements". 35 Box 4.2: Afghanistan's Trade Regime Establishment of an open and transparent trade regime has been a consistent priority of the Government, which has made sweeping tariff and customs administrationreforms, resultingin a significant improvement in transparency and simplicity, leaving Afghanistan as one of the most open economies in the region. Under the previous trade regime there were 25 tariff ,, bands, with rates ranging from 7-150%, and with a simple average of about 43%. However, a highly- 1 appreciated artificial exchange rate was used in calculating payments of customs duties (made in domestic currency), which meant that actual payments ~ were on the order of only 510% of nominal tariff levels. An overhaul of the tariff system brought the ,, number of rates down to six (2.5,4, 5, 8, 10, and 16%), with relatively low dispersion of rates, and shifted to 80 use of the market exchange rate in calculating duties payable. By virtue of this rationalization, the average ' tariff declinedto 4% (weighted by the number of items ineach rate category). Evenwith the additionof arange ~ ~h.-~ C- p.*aan T--" I- Id" Uibckxlsn of "fees", the estimated average appliedtariff is only 5.3% - the lowest inthe region. Afghanistanmaintains import banson only afew products(largely for religiouspurposes), andimposesno seasonal restrictions,quotas, or other non- tariff barriers.Furthermore, licensingrequirements- reformedeffective April 2004 -havebeendrastically simplified. The import license application process, which previously involved 42 steps, 58 signatures, and several weeks of processing, now involves only three steps, six signatures, and two days to process. As a result, Afghanistan's trade regimecurrentlyratesas a"4" interms of the IMF's Trade RestrictivenessIndex, same as the EUand United States. 4.20 A focus on large taxpayers can enhance revenues. In many countries a small number o f taxpayers (perhaps 5%) account for 75% or more o f total tax collections. In Afghanistan, the need to focus on large taxpayers i s heightened by the low level o f compliance and very limited administrative capacity. Afghanistan established an LTO in 2004 within the Revenue Directorate of MoF. Inorder to become fully effective, this newly-created organization needs to focus on selecting and then developing an appropriate relationship with the largest taxpayers, and should not get diverted to collect revenue from smaller clients. Appropriate criteria for large taxpayers have been developed by the Government. Within about two years, the LTO can be expected to be collecting at least 60% o f total (non-customs) domestic tax revenue. 4.2 1 Reform of provincial tax offices will require long-term capacity building efforts. Although the LTOindue course should collect a considerable proportion o ftax receipts, the provincial revenue offices, currently based in the Mustoufiats, will continue to collect substantial amounts, which will be important to ensure a relatively broad base o f taxation. However, the GPR has never been the headquarters o f the tax administration as the provincial Mustoufiats report to the Treasury. To develop a more professional service, the current PRR plan aims to build a new tax administration headquarters and operational tax offices inprovinces reportingto it. MOVING TOWARD A BROAD-BASED CONSUMPTIONTAX 4.22 The current indirect tax system has serious deficiencies, including: (i)narrow tax base leading a to low revenue yield; (ii) serious potential cascading (albeit generally at a low rate); and (iii) disincentives to export. These problems could be addressed by moving toward a broad-based manufacturers sales tax in 3-5 years, which can generate a significant amount o f revenue and will be necessary to ensure fiscal sustainability over the longer term. Value Added Tax (VAT) is the most common form o f consumption tax imposed by central governments, used in more than 100 co~ntries.~However, introduction o f a VAT InaVAT system, all persons or businesses engaged in supply of taxable goods and services are requiredto register for VAT and charge VAT on all taxable sales. Their actual liability, however, is only the net of the VAT charged on their sales (output tax), reducedby the VAT paid or payableby them(input tax) for goods and services purchasedfor use intheir taxable activities. 36 through to the retail level may represent a more radical reform o f the tax system than the Government i s prepared to introduce, or i s needed. 4.23 A single-stage sales tax at manufacturing level may be the best option. Phasedextension o f the recently introduced taxes on selected services could pave the way for a broader-based consumptiontax as development o f administrative capacity permits. A manufacturing sales tax imposes the tax at the point where goods are first sold in the case o f domestically produced goods, and at the point o f customs clearance for importedgoods. A single tax rate for all goods would be simpler to administer and would not distort relative prices. This lundo f sales tax has been employed inmany countries inthe earlier stages o f their tax development, and it can be expanded further to wholesale and eventually retail commerce with a minimumof disruption. TAXATIONAND INVESTMENT INCENTIVES 4.24 Overall, the tax regime in Afghanistan i s rather competitive, with one o f the lowest corporate income tax rates inthe region. In2004, the corporate income tax was reduced to a flat rate o f 20% o f net taxable income. Cross-country comparisons o f tax rates are difficult, but in Palustan corporate tax rates range from 35% to 50%, Iran charges a flat rate o f 25%, in China the rate i s 30%, in India the rate i s 36.75% for domestic companies with a more complex schedule for foreign companies, and inUzbekistan the corporate tax rate ranges from 20-35%. Figure 4.1 highlights Afghanistan's relative position by comparingthe total tax burden (excluding tax on labor) that would face a typical firm inseveral countries. Figure 4.1: Total tax payable (% of gross profits) 0 20 40 60 80 Iran I Afghanistan 1 '~ ~~ Chma ~~ Palustan I I Note: WorldAverage covers 155 countries. Source: Doing Business Indicators, WorldBank. 4.25 The first-best strategy for sustained investment promotion consists o f providing a secure, stable, and transparent legal and regulatoryframework, as well as adequate supporting institutions, including a tax system inline with international norms. On the other hand, available evidence suggests that providing tax incentivesfor investment promotion i s not cost-effective and that the revenue cost i s potentially significant. As a general rule, indirect tax incentives should be avoided (as inefficient), and discretion in granting incentives should be minimized as this can encourage corruption. 4.26 More specifically, tax holidays, which are widespread in developing countries, have well-known shortcomings. They basically consist o f applying a rate o f corporate income tax lower than the regular rate - often zero - to companies in a certain sector or region for a pre-specified period. The main objective o f tax holidays i s to increase investment. They raise several issues: 0 By exempting profits irrespective o f their amount, tax holidays benefit an investor who expects a highrate o freturn and probably would have investedeven without the incentive. 0 Tax holidays provide strong incentives for tax avoidance, as taxed enterprises could enter into economic relationships with exempt ones to shift their profits to the latter through transfer pricing. The duration o ftax holidays also i s prone to abuse. 37 0 Time-bound tax holidays, if they have any impact at all, tend to attract short-run projects, which typically are less beneficial for the economy. In Afghanistan, over 40% o f tax holidays granted before mid-2004 were reportedly awarded to construction companies. 0 The revenue cost i s seldom transparent, unless enterprises are required to file proper tax returns and the aggregate fiscal "loss" i s calculated. Conservative estimates by M o F put the revenue loss due to exemptions and concessions grantedprior to July 2004 at $30 million. 0 Tax incentives could be o f questionable value to a foreign investor because the benefits may be offset whenprofits are repatriated through an increasedtax charge inthe home country. 4.27 In 2004, tax holidays and exemptions were replaced by two tax incentives: accelerated depreciation and unlimited loss carry-forward. These have the least o f the shortcomings associated with tax incentives. Since merely accelerating depreciation o f an asset does not increase the total allowable depreciation beyond its original cost, little distortion in favor o f short-lived assets i s generated. Neither i s there much incentive for an enterprise to engage intax abuse. Accelerated depreciation also does not have the other negative elements associated with tax credits. C.Medium-termRevenueProjections 4.28 Underlying the fiscal scenarios outlined in Chapter 3 are illustrative revenue projections driven by: (1) economic growth and inflation: (2) customs and tax policy reforms; (3) customs administration reforms; and (4) tax administration reforms. The macroeconomic assumptions are similar for each scenario, with real GDP growing at around 10% over the first four years and falling to around 7.5% thereafter. Inflation also falls from an initial rate o f 10% per annum to 5%, while the nominal exchange rate remains stable. Several alternative scenarios illustratethe implications o f different reformpaths over a ten-year period: 0 Scenario I (baseline case) assumes gradual implementation of theJive-year customs and tax administration reforms. These include the wage withholding tax (raising a total o f $120 million inthe first three years); selected service taxes including roll-out to other urban centers ($36 million in the first five years); and the airport tax ($2.7 million over three years). Improvements in compliance are 10-15% per annum during the first six years before falling to 5% per annum subsequently. 0 Scenario 2 (high-case) assumes accelerated`implementationof theJive-year customs and tax administration reforms combined with additional revenue measures. Improvements in compliance rates are higher than in scenario 1, 20% p.a. inthe first four years, then gradually falling to 5% annually, assuming stronger performance from Customs administration and the LTO. There is also a gradual move toward a consumption tax which broadens the tax base, and eventually royalties from the mineral sector gradually come on stream. 0 Scenario 3 (low case) assumes no further tax policy reforms. Based on reforms so far, compliance rates improve by around 10% p.a. during the first six years before falling to 5% p.a. thereafter. N o additional major policy reforms are implemented. 4.29 The revenue ratios for each scenario are shown as a percentage o f GDP in Figure 4.2 (see also Statistical Appendix Table A.6). These ratios may provide useful benchmarks for assessing Afghanistan's revenue path. As discussed in Chapter 3, in the low case scenario Afghanistan would fail to meet a number o f its policy objectives. While the high-case scenario would require considerable efforts and thorough reform implementation, experience in other post-conflict countries suggests that it could be achieved. Although the revenue to GDP ratio rises by in excess o f one percentage point per year in the first four years, some other post-conflict countries have achieved this level o f revenue performance. For example, the creation o f the revenue authority in Uganda helped to double the tax revenue/GDP ratio from 6% to around 12% in the 1990s (Figure 4.3b). InRwanda the tax revenue/GDP ratio increased sharply from a low o f 4% in 1994to around 12%inlate 2002 (Figure 4.3a). 38 Figure4.2: Medium-termrevenueforecast (YOof GDP) 12 SAF revenueprojections 10 8 6 Low-case revenue 4 2 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012113 Source: SecuringAfghanistan's Future (SAF) and IMF estimates. Figure 4.3: Comparative GrowthinRevenueCollection(YOof GDP) a. Rwanda b. Uganda I t 12 O N , , I , I I , , , , , I I ' ' ' ' ' ' ' ' ' 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2oM) 2001 2002 ' ' ' ' ' ' 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Zoo0 2001 2002 2003 39 CHAPTER5. PRIORITIZINGEXPENDITURES 5.1 A second key desired outcome o f the PFM system is effective allocation o f public sector resources to different sectors, programs, and activities. This should be within an affordable budget envelope (as discussed in Chapter 3); in accordance with Afghanistan's development strategy, sector strategies, and national policies; and in a manner that achieves good development outcomes. With Afghanistan facing immense needs inmost sectors, the issue clearly boils down to prioritization: What is the right balance across and within sectors? What i s the right sequencing between competing priorities? What are the most effective interventions? This chapter first reviews the broad pattern o f expenditures across sectors and recent trends, and discusses why it i s so important to prioritize. The second section looks at prioritization o f expenditures across sectors, the third section at prioritization within sectors. A. SectoralExpenditurePatterns:Why It is Important to Prioritize THEALLOCATIONOFSPENDINGACROSS SECTORSAND RECENTTRENDS 5.2 Afghanistan's budgetary expenditures can be classified along programmatic lines, and divided into four broad pillars established by the Government in its strategy documents (Figure 5.1): human capital development (25% o f total spending in 2004/05), physical infrastructure (27%), general administration (9%, including trade and investment and public administration and economic management), and security (39%). Aside from the fact that security i s the dominant sector, with a rough balance between the human capital and infrastructure pillars and much less going into the fourth pillar, it i s not possible to say much at this very aggregated level. However, the pattern o f expenditures differs significantly as between the Core Budget and the External Budget. Human capital development i s the largest pillar in the Core Budget, reflecting substantial spending in the recurrent budget on teachers' salaries and in the development budget on the National Solidarity Program (NSP) and on basic health services (contracted out to non-government providers - see Chapter 8). On the other hand, security i s considerably smaller than human development in the Core Budget but twice as large in the External Budget (reflecting in particular massive spending on the Afghan National Army). Infrastructure comprises a much larger share o f the External Budget than o f the Core Budget. Figure 5.1: 2004/05 ExpenditureComposition (YOof total of each category) a. Total Expenditures HumanCapital Development Physical Inhstructure General Admmistmtion 33 % 41 17% 9% Source: MoF (preliminary annual statementsfor Core Budget andpreliminary annual reportfor External Budget). 40 5.3 Within the broad pillars, spending can be divided among 16 national programs, o f which the major ones include Education; Health; Livelihoods and Social Protection; Transport; Energy, Mining, and Telecommunications; Public Administration and Economic Management; National Police; and Afghan National Army (Table 5.1 and Table A.7 inthe Statistical Annex). The ANA was the largest single sector interms ofexpenditures in2004/05; the transport sector, mainly consisting ofroads, was second. Table 5.1: 2004/05 NationalBudget Expendituresby Program(!$million) Core Budget External Budget Total Budget berating Developme )evelopme %of ofwhich Exp. nt Exp. Total ntExp. total developt. HumanCapital Development 192 157 349 494 843 25 651 1.1 Refugeesand IDP return 2 2 73 75 2 73 1.2 Education 115 7 122 102 224 7 108 1.3 Health 25 39 63 91 154 5 130 1.4 Livelihood and Social Protection 37 111 148 192 340 10 303 1.5 Culture, Media, Sport 13 0 13 36 50 I 36 PhysicalInfrastructure 33 119 152 749 901 27 869 2.1 Transport 6 78 83 497 581 1 7 575 2.2 Energy, Mines, and Telecom 12 33 45 177 222 7 210 2.3 Natural ResourcesManagement 14 2 15 30 45 1 31 2.4 Urban Management 1 7 8 45 53 2 52 General Administration 58 23 81 217 298 9 240 3.1 Trade and Investment 3 6 9 31 40 1 37 3.2 Public Administration and Economic Management 55 17 72 186 258 8 203 Security 275 17 292 1,036 1,328 39 1,053 3.3 Justice 21 0 21 24 45 1 24 3.4 National Police and Law Enforcement 146 17 16? 177 340 I O 194 3.5 Afghan National Army 108 108 681 788 23 681 3.6 Mine Action 96 96 3 96 3.7 DDR 59 59 2 59 Unallocated 8 8 0 8 Total 558 317 874 2.503 3.378 100 2.820 Source: See Figure 5.1. WHYPRIORITIZE? KEYTRADE-OFFS 5.4 First, like any country Afghanistan faces major resource constraints over the medium term (Chapter 3), and prioritization i s essential to ensure that the available resources are well used. Substantial external assistance i s available now but inevitably will decline over time, so malung good use o f these extra resources duringthe current window o f opportunity i s critical for success. 5.5 Second, with the bulk o f public spending externally financed and most o f it donor-executed, ifthe Government does notprioritize, the allocation of expenditures will be determined in an ad hoc manner by thefragmentedpriorities of others. The External Budget has a very large influence indetermining the ex post prioritization o f expenditures. In2004105, for example, the relative ordering o f total expenditure by pillars (security, infrastructure, human development, general administration) was determined by the External Budget, whereas expenditure was allocated quite differently in the Core Budget (Figure 5.1). Moreover, the two top national programs in terms o f funding (ANA and Transport) were determined by the External Budget, as these were only fourth and fifthinterms o f spending inthe Core Budget. 5.6 Third, prioritization is a political process (Box 6.1), which needs to be informed by and to interact with the national development strategy at the broader level (and strategies for different pillars like security as well as sector strategies at the narrower level), and sound technical inputs and guidance are required for budget decision-making (by Cabinet, later by Parliament). While at one extreme there i s no technical "magic formula" and it i s unrealistic to expect a purely technical prioritization o f expenditures to be feasible (or even desirable), at the other extreme political decision-making in the absence o f any sound technical inputs can lead to disastrous outcomes. 41 5.7 Fourth, Afghanistan faces especially severe issues in terms of prioritizing expenditures over time, in particular related to the downstream O&M consequences o f public investment decisions made today (in both Core and External Budgets). The large amounts o f recurrent expenditure in the External (Development) Budget will generate enormous pressures later (Chapter 3, Section C). B. How to Prioritize: Allocating Expenditures Across Sectors 5.8 There are n o easy answers on how to prioritize in allocating public expenditures across broad pillars and sectors. Sophisticated technical approaches can require a great deal o f resources and even so are highly imperfect and can result in misleading conclusions. Simpler rules o f thumb may work better. Judgment i s called for. And inevitably political factors will be prominent. 5.9 A first critical point is that inter-sectoral allocation decisions need to be anchored in an encompassing national strategy for development. The national strategy document should provide general guidance and a framework for prioritization, as well as broad strategic priorities. The Government's National Development Framework (NDF - 2002) and subsequently Securing Afghanistan 's Future (SAF -2004), as well as documents prepared for Afghanistan Development Forum meetings, have provided a strategic vision and framework. In the SAF, a detailed medium-term investment program also was put forward. However, prioritization and sequencing were not concretely spelled out, and the discipline o f a resource constraint was not applied. The Government i s currently preparing its interim Afghanistan National Development Strategy (ANDs). It i s hoped that the ANDs, and the consultative process around it, will provide guidance on broad prioritization o f spending." Prioritization will also be facilitated by setting a medium-tern fiscal resource envelope (Chapter 3), and progressively strengthening its sector content. 5.10 Without overall strategic guidance, intersectoral allocation decisions can be arbitrary and counter- productive. A good example i s security where, most strilungly, thejustice sector has received only 3% o f total security funding over the past three years (Figure 5.2). Police reforms have progressed slowly compared with the rapid development o f the ANA. Based on the PFM Review's analysis (see Volume V), contributing factors to these anomalies include (i)the earlier lack o f an agreed national security strategy (not surprising in the initial period following the end o f major conflict in late 2001); (ii) the overwhelming dominance o f the External Budget in overall security spending; (iii) the above problems, and fragmentation of decisionmalung across donors, malung it more difficult for the National Security Council to fully carry out its mandated strategic, leadership, and coordination functions, although the situation has improved in 2004 and 22005; and (iv) fragmentation across sectors possibly exacerbated by the "lead donor" approach (whereby a lead donor was assigned for each sector withinthe Security Pillar). 5.11 More generally, identifiing and mitigating gross anomalies in inter-sectoral allocations of expenditures can be a powerful tool for prioritization, if informed by strategy and mediated by the resource envelope. Outliers can be identified interms o f their relatively low (or very high) shares of total expenditures, or through comparing expenditure growth rates for different sectors. Agriculture may be a good example o f a critically important sector (inthe natural resources management national program) that has been underfunded." Onthe other hand, the size and growth o f expenditures on roads, while reflecting massive rehabilitation needs and the highpriority assigned to this sector by the Government and partners, do seem to have far outrun investments inother major infrastructure sectors. 5.12 It is very important to ascertain the reasons behind gross anomalies in inter-sectoral expenditure patterns. Where major underspending ina sector i s related to limited capacity or poor performance of the loSuch prioritization at a strategic level is challengingbut possible, with success dependenton having a meaningful linkage to the budgetprocess (medium-termframework and annualbudgeting). For example, Armenia's draft Poverty Reduction Strategy Paper (PRSP) proposed interventions in over 100 areas, but intensive consultations supported by a software tool facilitated reachingconsensusto concentrateon 12priority areas linked closelyto the MTEF andwell-specifiedpolicy actions. In 2003104, less than $40 million was spent in the Development Budget on projects under the Ministry of Agriculture's oversight, ofwhich only $1millionwas inthe CoreBudget. 42 concerned ministry, then additional funding alone will not solve the problem and on the contrary could waste resources. Sometimes gross anomalies may be in large part attributable to donors' preferences, as they may gravitate toward visible, popular, or well-performing sectors and away from others. Thus rectifying underspending needs to be accompanied by measures to address the underlying problem that resulted inlow spending inthe first place. Figure 5.2: Composition of Security Expenditures, 2002/03-2004/05 Nahonal Police and Law Enforcement 28% ustice 356 Afghan National e6% Action A m y 60% DDR 3% Source: Volume V. 5.13 At the opposite extreme in terms o f technical sophistication i s use of cost-benefit analysis as a guide to inter-sectoralprioritization. While superficially attractive, this approach carries heavy technical and informational requirements, and even in the industrialized countries it i s not used systematically except in the evaluation o f new project proposals. Moreover, there appear to be biases across sectors. For example, power projects and roads tend to have high estimated economic rates o f return, whereas those for large irrigation projects tend to be more marginal. Since such differences are common across projects and country situations, they most probably reflect differences in the methods applied, data, etc. rather than mainly underlying differences in the development impact across sectors. Thus using cost- benefit analysis to prioritize across sectors can be misleading and counterproductive. 5.14 Information on international patterns of spending and outcomes across sectors may provide some guidance, although it must be appliedjudiciously in line with the country context. An international comparison highlights the very large share o f Afghanistan's public expenditures allocated to the security sector (Statistical Appendix Table A4). It also suggests that, ifthe External Budget i s taken into account, the share o f education and health intotal spending i s lower than inother countries (11%against 18%). 5.15 An approach based on identifying bottlenecksand directing expenditures to alleviate them also may be useful in the short run. One example i s building an electricity transmission line from northern Afghanistan to Kabul to deal with the power supply bottleneck facing the city. Strengthening the judiciary and creating a solid statistical system appear to be important enabling factors for many other sectors. Similarly, identifying successes can lead to scaling up or replicating them, for example in the case of the National Solidarity Program. 5.16 Another important lesson i s the needfor inter-sectoralprioritization to be comprehensive- no sector (e.g. defense) should be considered sacrosanct or immune fiom scrutiny. 5.17 Incertain countries, the Government may allocate domestic resources so as to counterbalance donors' allocations of donor-executedassistance, thereby ensuring that the overall pattern o f spending i s inline with national strategic priorities. InAfghanistan, however, the ExternalBudget is so large that it would be impossible to use the limited resources in the Core Budget to offset donors' priorities manifested in the External Budget. Hence there i s no alternative to close cooperation with donors to ensure that the overall allocation o f resources inthe External Budget corresponds with national priorities. But this is difficult, and the problems faced by the Government inmanaging the External Budget provide strong grounds for more external assistance to go through Core Budget channels (Chapter 6). 43 5.18 From an incentive standpoint, it may make sense to have some linkage between sectoral allocations and performance. For example, ministries/sectors that do better in terms o f spending the money they are allocated would get continuing or higher allocations in the future, whereas those that are not able to spend their allocations would get reduced allocations. However, a strong linkage to performance can result over time inunderperforming sectors systematically getting less resources than the importance o f the sectors concerned in the national strategy would call for. Since donors in their own allocation decisions respond to perceived good performance, they can exacerbate such imbalances. Thus while some short-term performance linkage may be desirable, from a medium-term perspective it i s essential to correct the reasons for underperformance and ensure that well-performing sectors do not become grossly overfunded inrelation to their strategic importance, within overall resource constraints. C.How To Prioritize: AllocatingExpendituresWithin Sectors 5.19 As in the case o f inter-sectoral allocation, prioritization o f spending within sectors (e.g. health, power, highways, irrigation) needs to be anchored in sound sector strategies, which should provide general guidance on priorities. Technically, prioritization within sectors i s more manageable than inter- sectoral prioritization because the alternatives tend to be more comparable. Cost-benefit analysis, which cannot be used uncritically for broad intersectoral allocation decisions, i s much more useful for prioritizing projects or programs/activities within a sector. Moreover, insectors where benefits cannot be easily expressed in value terms or doing so i s controversial (e.g. health), cost-effectiveness analysis can be applied, whereas this could not be usedfor the most part across sectors. 5.20 A very important considerationis the appropriate role of Governmentfor the activity. The two key reasons for public interventions are because o f public goods or externalities (i.e. goods and services that would not be adequately provided by the private sector alone) and for redistribution. Even when there i s a clear case for Government intervention, the type o f intervention can vary: regulation, financing, or direct provision. Public expenditures should therefore be reviewed against these criteria (Box 5.1). 5.2 1 Geographical and gender disparities comprise an important dimensionfor prioritization within sectors. For instance, in education disparities in enrollments between boys and girls, rural and urban areas, and provinces suggest that adjustments in expenditure allocations are necessary (Figure 5.3). A review o f the education sector suggests that these issues can be addressed by an appropriate strategy, including recruitment o f teachers (notably female teachers), targeted programsto increase enrollment, and use o f local participatory mechanisms such as School Management Committees (Volume IV, Chapter 2). Figure 5.3: Net Enrollment Ratios, Grades 1-6 (2003) 100 90 80 70 60 50 40 30 20 10 Source: VolumeIV, Chapter 2. 44 5.22 Technical analysis can be used to evaluate individual projectproposals and as general guidance for prioritization o f expenditures within a sector. Health i s a good example o f the latter approach, where the cost per life saved (cost per DALY12saved) o f different types o f interventions can be compared. In general, basic health interventions are much more cost-effective than more expensive curative care including hospitals (Table 5.2). A general rule o f thumb i s that when the cost per DALY saved o f a health intervention exceeds average GDP per capita (around $250 inAfghanistan today), allocating public resources to that activity constitutes a relatively inefficient use o f resources (particularly if activities with lower cost per DALY saved are underfunded). Based on these considerations, the Government has prioritized provision o f an agreed Basic Package o f Health Services, and has tried to resist pressures to make major new investments inhospitals, particularly inKabul. Table 5.2: Cost-Effectiveness of DifferentHealth Interventions Note: DALY is a disability adjusted life year. Source: VolumeIV, Chapter 1. 5.23 Inmajor infrastructure sectors where large projectsare involved, decisions on individual projects form a primary element of intra-sectoral resource allocation, and cost-benefit analysis can be very useful. The highway sector i s a good example. Afghanistan already has a large highway rehabilitation program underway, with a total cost exceeding $1 billion. Particularly in view o f the future O&M requirements, any major new highway project proposals must therefore be treated with great caution and carefully prioritized. Benefits depend greatly on traffic forecasts (both freight and passengers), and on available alternative routes. Costs per kilometer may be fairly standard (depending on the type o f road construction involved and the condition o f the existing road), although there has been a great deal o f cost variation in Afghanistan - depending in part on mountains, the number o f bridges, tunnels, etc. (Box 7.3). High altitude and steep grades also affect the costs o f users and reduce the net benefits to them. Another factor i s extra security costs, estimated at 3-15% for highway rehabilitationprojects. 5.24 Some other considerations are very important for the intra-sectoral allocation of spending. One i s complementarities between different projects. A good example is the power sector, where gross imbalances between investments in generation, transmission, and distribution would sharply reduce returns. This highlightsthe need for a robust sector strategy, in this case an Energy Master Plan, which provides the rationale for prioritization, based on sequencing investments, taking into account lags in buildinglarge-scale infrastructure equipments, and adequately financing O&M. 5.25 Prioritization o f public investments within sectors also must include the time dimension. Medium-term targets for service delivery (whether for physical infrastructure or social services) need to inform investment decisions and O&M implications - not investments driving O&M. In addition there are critical sequencing issues in terms o f what comes first, what comes later. As public investment accumulates over several years the O&M requirements become truly daunting, leading to questions about what size o f investment program i s affordable over the longer run, and to the critical need and potential for cost recovery to partially or fully fund O&M requirements (Chapter 8). 5.26 The hydrocarbon and mining sectors provide good examples o f how important, and intricate, sequencing issues can be. Managing these underground resources is important for the Government, because o f their potential impact on both domestic revenues and private sector development (including in DALY is a disability adjusted life year, which provides a common measure for comparingthe burden of disease taking into accountbothdeath anddisability arisingfrom illness. 45 downstream sectors). Capacity development i s an urgent priority, as lack o f capacity to shift to an effective regulatory approach, limited political leadership to shift toward private development o f undergroundresources, lack o f policy consistency, and limited external assistance can combine to prevent progress, whereas a breakthrough i s required. 5.27 Finally, technical criteria including cost-benefit analysis can be used to clean out excessive "wish lists" of proposedprojects inthe portfolios o f different ministries. A damaging practice that i s sometimes found inother countries i s to maintain much too large a portfolio o f approvedongoing projects, for which the likely funding available will be far from sufficient to complete all o f the projects. This must be avoided through ruthless culling o fnew project proposals as well as dropping approved projects and even those under implementation that have little hope o f being completed. InAfghanistan some progress has been made in this regard: the 2004105 development budget included 540 projects out o f the 1,110 proposed by line ministries, o f which only 210 were in the Core Budget. But this most likely i s still too many, and the slow pace o f project implementation suggests that better project screening and budget realism i s required (Figure 3.1). The preparation o f the ANDs creates an opportunity to strengthen this process and create a lean, high-quality portfolio o fpublic investment projects (Box 5.1). Box 5. 1: Prioritizing When Everything is Critical Prioritization is technically and politically a daunting task in the Afghan context, with limited data and analytical capacity and enormous needs. However, prioritizing is important technically (not all projects will have the same economic and social impact and some projects require other projects to be completed first), fiscally (unrealistic budgets either produce large fiscal imbalances, or, not being implemented, make the budget an irrelevant tool), and politically (poorly prioritizedbudgets raise expectations, leading to disenchantment). The first filter for prioritization is that each project should contribute to the ANDs's objectives and be part o f the sector's master plan. The contribution to objectives o f the ANDs should be clearly specified. Beyond this filter, projects could be ranked based on four simple criteria: Rationale (Weight 15%): Should this be done by the public sector or can the private sector adequately undertake the activity? Does the project target the poor? Cost-effectiveness (Weight 15%): For the output o f the project, has the least-cost alternative been identified? Are multi-year implications laid out, and is the project likely to be self-financing after completion? Benefit-cost (including identification o f beneficiaries) (Weight 50%): Have benefits (e.g. social, financial) been quantified? Do benefits exceed costs? Risk and mitigation (Weight 20%): I s the project likely to be completed on time? Have allowances been made to address potential physical and financial contingencies? Are there institutionallmanagerialltechnical constraints in carrying out the project? Are there any environmental risks? Based on this ranking, the top-priority projects can be included inthe budget up to the sector resource envelope. It should be noted that a framework along the lines put forward above would need to be flexible in accordance with sector-specific circumstances. For example, some sectors like security are not expected to be self-financing, so the point inthe second criterion about a project beingself-financing after completion would not apply. Based on Swaroop (2004). 46 CHAPTER 6. MAKINGTHE NATIONALBUDGET THE CENTRAL INSTRUMENTOFPOLICYAND REFORM 6.1 Previous chapters have looked at important outcomes o f the PFM system; the next two focus on the P F M system itself, starting with the central process o f national budget formulation. The first section o f this chapter reviews the legal and institutional framework o f the budget. The second analyzes progress and challenges in making the budget an effective policy tool. The third outlines issues o f ownership and political buy-infor the budget. The last section looks at the implications o f the large amounts o f external aid for Afghanistan and discusses how donors can support the Government's efforts to make the budget the central instrument o fpolicy andreform. 6.2 The issue of Afghanistan's weak institutional capacity (Chapter 9) has to be taken into account. It would be futile to reform budgeting without improving the Government's overall management framework. Ultimately, performance i s a management issue, and there would be no demand for a performance-oriented budget without performance-driven management. Moreover, effective budgeting i s demanding interms of information and analysis. Thus appropriate sequencing i s essential to bringabout improvements that achieve their objectives. A. Budget PreparationProcess 6.3 The 2004 Constitution provides the legal authority for preparing, approving, and executing the national budget. The Public Finance and Expenditure Management (PFEM) Law, approvedinJune 2005, further specifies the processes and responsibilities. The Government i s responsible for preparing the budget (and executing it, see Chapter 7), with M o F leading the process. The budget i s then to be presented for approval to the bicameral legislature no later than 45 days before the start o f the fiscal year, and approval i s required no later than a month after the start o f the fiscal year. Inprevious years, as the Parliament had not yet been established, the Cabinet approved the national budget through a Presidential Decree (see also Box 6.1 on the political economy o f the budget). 6.4 As reflected in PFM indicator 11, progress has been made toward an orderly annual budget process. Inparticular, it has moved from a pure bargaining process to a more policy-oriented discussion. Modernization o f Soviet-inherited practices has started. The mid-year review, established since 2004, has significantly improved the quality of budget management by allowing transparent adjustments during the year and creating a feedback loop from one year's budget implementation to the next year's budget preparation. Finally, significant information i s publicly available on the Government's website, notably thebudget guidelines,budget decrees, and fiscal reports (PFMindicator 10and Box 7.2) 6.5 But much more remains to be done. A review of the seven steps of the budget formulation process (Table 6.1) reveals two key issues (in addition, some features o f the process weaken the policy content o f the budget, see next section). First, as in many countries the schedule i s very compressed, starting after the mid-year review (late December 2004 for the 2005/06 budget). Either the process ends on time (for the operating budget) but with limited opportunity for substantive analysis and discussions, or it ends after the start o f the fiscal year (for the development budget), constraining budget implementation. Timing i s now even more critical since the Parliament will need sufficient time to review the budget. Second, the ,early stages o f the process are weak, with no pre-determined budget envelope or sector ceilings. In many countries, the Cabinet reviews and agrees on the fiscal envelope (updating the MTFF, see Chapter 3) and on budget ceilings by sector, which sets a clear direction for aggregate fiscal discipline and broad allocations across sectors, reinforcing the budget's strategic content. The absence o f such ceilings weakens the budget's policy content (Section B) and increases political bargaining (Section C). Section D looks at the role o f external assistance inthe process. 47 Steps Description Issues 1.Issue Budget Preparedby Budget Department and issuedto Schedulegives Guidelines line ministries -- Development little time for analysis and operatingbudgetsremain 2. Decide Budget Envelope 3. Prepare Expenditure Proposals Source: "Budget Formulation in Post ConjlictAfghanistan (2005/06)", backgroundpaperfor the PFM review. B. PolicyOrientation of the Budget COMPREHENSIVENESSAND CLASSIFICATION 6.6 There are two technical requirements for the budgetto be a useful policy tool (see PFMindicators 5 to 9). First, the budget needs to give an overall picture o f public finances; if parts (e.g. humanitarianor defense spending) are excluded, the Government cannot assess the impact o f the budget on the macro- economy or make meaningful sector allocations. Second, the budget must follow a classification system that allows meaningful analysis. 6.7 Progress has been made toward a comprehensive budget. Thanks to strong aid coordination efforts, the broader budget concept captures most external assistance (Figure 6.1). However, for the External Budget the Government relies on voluntary reporting by donors. The quality of data i s poor (data are incomplete, based on somewhat different definitions, classifications, and reporting periods) and actuals are often not available (see inFigure 6.1 the discrepancy between the Government's 2002/03 data and OECD data). Another issue i s that the fiscal risks from public entities other than the Central Government are not monitored: there i s very limitedinformation on SOEs and on municipalities. Figure 6.1: Comprehensivenessof Budget and ExternalAssistance ( X GDP) I 50 40 30 20 10 Source: Government data based on MoFfiscal reports andfinancing indicated in 2005/06 budget (security expenditures in the External Budget are separated out); OECD data based on reportsfrom donors to OECD /DAG (in principle excludes military support) -data available up to 2004. 48 6.8 Progress toward a meaningful classification system also has been commendable. Since March 2005 a chart o f accounts, covering all Government transactions, has been in place and should provide information in line with internationally accepted standards. 6.9 The main classification problem i s the lack o f integration between operating and development budgets. The dual budget structure has no policy basis, as the operating budget includes 7% o f capital expenditures and the development budget includes almost 40% o f recurrent expenditures (based on simple estimates, Figure 6.2). Lack o f integration i s very detrimental to the effectiveness o f public spending, as investments are made without planning (and financing) for maintainingthem. There is also a bias toward investments, crowding out operation and maintenance o f completed investments to actually deliver services, which can also lead to political difficulties (investments raise hopes that are disappointed afterward). Dual budgets also potentially open up the opportunity for double-allocations for the same purpose (e.g. maintenance costs ina ministry's operating budget and ina donor-financed project). Figure6.2: OperatingandDevelopmentBudgets($ million) 1,500 1,250 1,000 750 500 250 Operatmg Core ExternalBudget Secmty m Development w/o secmty ExternalBudget W Wages, Goods and S e m e s , Transfers Cap&l Investments I Source Based on MoF reportsfor 2004/05, staff estimatesf or development budget (no oflcial classification IS available). 6.10 Achieving progress toward integrated budgeting requires three reforms. The first i s to structure MoF's Budget Department along sectoral lines, with no organizational separation between the operating budget and the development budget, and build up its capacity, complemented by similar reforms in line ministries. This is already beinginitiated. The second i s alignment o f classifications and the presentation inbudget documentation, accounting, andreporting systems, which has already started with the Chart of Accounts implemented in March 2005. Finally, donor practices need to support budget integration (see Section D). 6.1 1 A number o f other accounting issues also need to be addressed inorder to improve the quality o f budget presentation, notably full adoption o f a cash-based presentation instead o f mixing cash flows with commitments. The current presentation o f the financing framework i s weak and does not show properly the amount o f cash that i s carried over fi-om one year to the next on the Government's accounts. These shortcomings contribute to weakening the realism o f the budget and its usefulness as a policy instrument. FROM POLICY TO BUDGET 6.12 Building on these two important technical aspects, the budget can actually become a policy instrument if the country's development strategy i s translated into annual budgets. First, the budget should reflect the Government's multi-year strategy. At the aggregate level, annual budgets represent the year-wise articulation o f the path toward fiscal sustainability (see Chapter 3). To facilitate this linkage, the MTFF should be updated every year with Cabinet approval. Within sectors, budgets should anticipate the flow o f resources necessary to operate and maintain new investments, thus requiring integration o f operating and development budgets as discussed above. 6.13 Second, across and within sectors, the budget should operationalize the Afghanistan National Development Strategy (ANDs), whose interim version i s currently at an advanced stage o f preparation. 49 This means aligning with the main strategic trade-offs across sectors (see Chapter 5); hence it is advisable for the Cabinet to agree on broad sector envelopes early in the annual budget formulation process. For example, while improvements inthe justice sector are high on the development agenda, spending has not been in line with this priority (see Figure 5.2). The strategic linkage also implies careful prioritization o f projects within each sector, as discussed in Chapter 5, and adequate input from a monitoring and evaluation system, as discussed inChapter 8. 6.14 Finally, the budget should reflect cross-cutting issues such as gender, environment, counter- narcotics, and provincial allocations. The first entry points are the sector strategies themselves (e.g. recruiting female teachers to reduce disparities in school enrollment between boys and girls). This strategy o f "mainstreaming" i s consistent with the view that these issues cut across many sectors. In particular, the issue o f gender needs to be reflected in budgeting, not only in sectors like Health and Education but also in Agnculture, Public Administration and Economic Management, etc. where women's contribution to the economy i s significant and growing (see World Bank, 2005b). This approach can be complemented by expenditure analysis. For instance, some countries have created virtual "poverty funds": expenditures that seek to reduce poverty (e.g. in health, education) are tagged in the Government's accounting system to allow analysis. Such an approach may also be useful in the counter-narcotics strategy. In some cases, targeted expenditure programs can be created for these cross- cutting issues, but their size should remain under control to avoid wasteful overlaps with sector programs. C.Enhancing Ownership and PoliticalBuy-In 6.15 Political buy-in i s o f utmost importance for making the budget a credible policy tool - which reflects the Government's strategy and i s effectively implemented- and an instrument o f nation building. This requires the involvement o f many stakeholders at various stages o f the strategic and budget cycle, which i s extremely challenging in Afghanistan, with a political transition, a fragmented Cabinet, the absence hitherto o f a Parliament, a fragmented civil society, and extensive foreign involvement. As outlined in Box 6.1, this means that it i s difficult to assess who decides on the budget and therefore the extent o f buy-infor the budget. 6.16 This section outlines two key directions to enhance political buy-in: participation and transparency inbudget preparation. Implementation o f the PFEM Law i s a priority inthis regard. Box 6.1: W h o Decides on the Budget? The budget is one o f the most important political tools, the one that provides financial means to implement policies. In Afghanistan the rules are straightforward; the Cabinet collectively prepares the budget, the Parliament approves it, the Government implements it, and the Parliament has final oversight with respect to the audited financial statements. The management o f this politicaltool has been complex, however. First, in the absence o f an elected Parliament, the budget has been approved by the Cabinet with the President signing Budget Decrees. There was limited external demand to release financial statements and audit findings. Second, during the transitional administration, budgets were based on imperfect information and thus reflected the outcome o f a game among key political actors rather than a consistent strategic framework. Much progress has been made on this front, but political constraints have remained significant. Third, the President, both by his authority to sign decrees with financial implications (including decrees to increase establishment lists) and by his authority over large contingencies, retained significant power to reallocate budgets during the year. While contingencies have been significantly reduced in 2005, a more cohesive approach to budget management will require strengthening o f the Cabinet Secretariat. Fourth, MoF i s the counterpart for a number o f donors, including the ARTF, in terms o f resource allocation. Inthe absence o f a strong Cabinet Secretariat and a comprehensive budget, projects have been proposed for financing without going through the due process o fbudget formulation. Fifth, both line ministries (by allotting funds to provincial departments) and MoF (by allocating cash to provinces) can influence the distribution o f spending between Kabul and the provinces, with little external oversight. Finally, the international community retains considerable influence over the budget, mainly by directly financing ring-fenced projects but also through involvement inpolicy dialogue. 50 6.17 Participation starts within the Government. At the technical level, ownership - as well as the quality o f the budget, as argued in Chapter 8 - will be greatly enhanced when provincial departments have a substantive input into the national budget. This should inparticular correct the gross imbalances in resource allocation between Kabul and the provinces in sectors like education. At the political level, buy- in-as well as strategic content, see the previous section-will be stronger ifthe Cabinet is engaged early inthe budget formulation process throughthe establishment ofaclear fiscal strategy and sector envelopes (preparation and approval o f the ANDs also contributes to this ~bjective).'~Inmany countries the budget process starts with a "strategy period", which gives the opportunity for collective political buy-inon high- level decisions such as the multi-year fiscal framework and sector ceilings for the budget. Progress in these areas will also require modernization o f the processes and rules o f the Cabinet (including strengtheningo fthe Cabinet Secretariat). 6.18 Participation o f the Afghan people also should be increased (directly and indirectly through media or civil society organizations). This i s mainly relevant at the policy-malung stage (preparation o f the ANDs), while budget preparation itself can be viewed as a more technical exercise requiring decisions under a tight deadline. Even without formal public participation, greater transparency in the Government's budget preparation process (e.g. through improvements in publicly available budget documentation, with translation into D a n and Pashtu) can help strengthen ownership. Chapter 8 explores various mechanisms to give citizens a voice inservice delivery. 6.19 The election o f Parliament in September 2005 also provides a basis for more participation and transparency in decision-malung. Although there are risks o f the Parliament becoming fragmented and paralyzed, the Parliament can also emerge as a major element o f peace-building, notably through its role in representation and reconciliation (WBI 2005). From the P F M perspective, the two major roles o f Parliament are to provide democratic participation in the budget process (through its budget approval function) and oversight o f budget execution (by reviewing the Government's annual financial statements and external audit reports). In both cases, Parliament can generate demand for the Government's accountability - provided that its internal rules allow for constructive politics and its capacity (notably that o f the dedicated Committees for Fiscal Affairs and Public Accounts) i s built up. The role o f donors, including their role in ownership and political buy-in,i s the subject o f the next section. D.Role ofDonorsinImprovingthe BudgetProcess 6.20 As emphasized in several chapters, the international community plays a significant role inPFM. This is a major opportunity for Afghanistan, as donors bring significant financial resources as well as technical advice based on worldwide experience. But history suggests that this opportunity could be short-lived, as many post-conflict countries experience at best only a few years before levels o f assistance go down to more normal levels (World Bank, 2003a). 6.21 With respect to PFM performance, the involvement o f donors does give rise to challenges (Box 1.3). First, at the policy level, donors influence the agenda and the directions in a way that, when uncoordinated, reduces Government ownership and policy consistency (e.g. in education, see B o x 7.1). Even high-level prioritization across broad pillars can be altered by donor funding (Figure 5.1), and the same i s true o f prioritization across sectors and programs. Second, when the strategy i s translated into budgets, donors can undermine budget choices by funding projects that are not high priorities for the Government or that the Government cannot afford to operate and maintain. Third, the sheer magnitude o f external assistance creates fiscal risks, since aid flows tend to be unpredictable over the medium term. As discussed in Chapter 3, this i s especially problematic in the security sector. By negotiating projects directly with line ministries, donors may also undermine the hard budget constraints that M o F seeks to enforce (including with respect to decisions on affordability o f O&M). Finally, reflecting their own l3The disconnectbetweendevelopment programs andministerial portfolios - partially addressedby the December2004 Cabinet reshuffling- also contributesto fragmentation, lowersprograms' coherence, andlessens political buy-in. 51 fiduciary interests, donors often promote fragmented, ring-fenced projects that bypass Government systems and use donors' own fiduciary processes, with lack o f harmonization across donors (see PFM indicator D2). All o f these issues arise not by design but because donors' presence itself alters the incentives o f various actors -to fkagrnent projects, attract funding that bypasses fiscal discipline, etc. 6.22 Should the Government go so far as refusing aid? In extreme cases, this might well be appropriate. For instance, donor-driven investments in public hospitals are sometimes referred to as "Trojan horses" because o f their large operating costs which crowd resources out o f priority areas such as the basic package of health services. Inmost cases, however, the recommendation i s not to refuse aid but to progressively integrate donor assistance inthe budget. 6.23 As suggested by Table 6.2, there are various degrees of integration with national systems. First, in 2002 the Government established the concept of an "External Budget": even if resources are not flowing through the Government's accounts, donors should coordinate their interventions through the Consultative Group process and report financial data to the Government. Second, the introduction of the Core Budget concept in 2004 has formalized a Consolidated Fund under Government control. In most cases, the Government can decide on the allocation o f resources in the Core Budget (e.g. for IDA allocations, the Government develops the work program inconsultation with the World Bank), funds flow through the Government's accounts, and accounting and reporting follow national procedures. Often external audit i s done independently, but this could be reformed since Afghanistan's Auditor General already audits the whole o f Government accounts, which include these funds. Similarly, donors often require the use o f their own procurement guidelines, but this could change once the new, modem Procurement Law i s implemented (Chapter 7). budget Externalbudget 1,629 hl ii Off-budget unknown 52 6.25 These various positive factors, together with donors' commitment to move toward aid harmonization as exemplified in the Paris agenda (see www.aidharmonization.org), suggest a way forward. It i s hoped that the ANDs will open up the possibility for multi-donor support, as inmany other countries that have prepared Poverty Reduction Strategy Papers (PRSPs). The annual budget will be the vehicle to implement the ANDs and could be complemented by an annual plan o f policy actions. Such a package o f prioritized short-run actions could thus comprise a "platform" from which the next set o f actions would take off, with monitoring and feedback to guide the process at each stage.14 Moreover the package, rather than any single measure or mechanical target, would be what the Government takes responsibility for and could form the basis for dialogue and agreements with the international community. 6.26 Donors could also commit to: (i) increase the predictability o f their financing (Le. provide clear indications of their commitments early in the budget process and, to the extent possible, multi-year commitments); (ii) condition their project assistance on the inclusion o f the projects concerned in the national budget; and (iii)provide timely and harmonized reports to the Government (including an agreement on simple, harmonized definitions for concepts such as expenditures, disbursements, and commitments). The Government for its part cannot decree donor harmonization, but progress in PFM performance will certainly contribute to substantial progress on this aspect. In particular, the Government's continuing commitment to maintaining adequate fiduciary standards and taking actions to further improve fiduciary performance, and demonstrated performance in this regard, will be critical, as will be strong performance indomestic revenue mobilization (Chapter 4). Progress inbudgeting, notably the recent adoption of an MTFF and strengthening o f the MTFF over time, will also help deepen the partnership with the international community. I Box 6. 2: The CriticalRoleof the AfghanistanReconstructionTrust Fund The Afghanistan ReconstructionTrust Fund (ARTF) was set up in the spring of 2002 to provide coordinated external financial support to the Governmentbudget. Up to the end of 2004105, it has receivedcontributionsworth S847 million from 24 donors, disbursed$610 million, and committedan additional $98 million. The World Bank is the administratorof the ARTF, while the World Bank and three other multilateral institutions (ADB, IsDB, and UNDP) are members of the ARTF Management Committee. The main outcome achieved is that the ARTF has financed Afghanistan's civilian recurrent budget ($508 million, mainly focused on education and a few other civilian sectors, as police and military expenditures are not eligible), while a dozen investment projects are under implementation. In 2005, an external evaluation stressed the importanceof the ARTF, in particular its recurrent cost window, in strengthening the Government's financial management systems. ARTF provides stability and predictability to Afghanistan's public finances and strengthensMoF andthe budgetprocess. Throughthe use of a Monitoring Agent - an external firm which reviews recurrent expenditures prior to reimbursement by the ARTF - it is also strengtheningfiduciary standards andprovideswhat is in effect an internalaudit functionto MoF. The evaluation also made recommendations to further strengthen the ARTF, notably the need to place ARTF support in a medium-termframework and to address the issue of capacity in public finance management. DFID's recent three-year commitmentis a major step inthe direction of multi-yearpredictability. Based on external evaluation ofARTF, March 2005. l4See DFID (2005) for a discussion of the background, rationale, and modalitiesof how such a "platform" approachcould work in the context of country situations involving externalbudget support. This type of approachis currently being usedincountries like Cambodia, EastTimor, andRussia. 53 CHAPTER7. ENHANCINGTHE EFFECTIVENESSOFBUDGET EXECUTION 7.1 If institutions and processes lead to a well-conceived budget (Chapter 6), the quality of expenditures depends on four conditions. Are service delivery units receiving budgeted funds on time? A r e there adequate processes and controls with respect to the use o f these resources? Are procurement methods generating value for money? And i s the Government accountable for its management o f public finance? This chapter discusses these four conditions (PFM indicators 13-28), and also looks at the very important issue o f corruption inthis context. A. Getting Fundsto Service Delivery Units 7.2 To deliver services, units must receive the necessary funding in a predictable and timely way. Despite muchprogress, there i s evidence that this i s not yet the case. Substantial achievements have been made inreforming the payments system. However, further improvements are needed at two main levels, M o F and line ministries. Indeed, timely flow o f funds to ministries i s mainly MoF's responsibility in terms o f managing cash assets and reviewing budget implementation. But line ministries need to allocate these funds to their departments, including inthe provinces and districts. 7.3 The development o f mid-year reviews and basic fiscal reporting (to line ministries and the Cabinet) has been an early success (inpart due to investments in a computerized system, see Box 7.2). In addition, some cash flow projections have been developed, and cash management has been greatly facilitated by the ARTF (Box 6.2). Finally, despite remaining challenges the Central Bank, which acts as the Government's banker, has made great progress since 2002 in ensuring timely payments throughout the country (see B o x 8.3 and, for the example o f teachers' salaries, Box 7.1). Box 7.1: Tracking Public ExpendituresinEducation A small Public Expenditure Tracking Survey (PETS) was conducted in the education sector as part o f the PFM Review. The study involved interviews with 217 teachers and 109 head teachers across 109 schools randomly chosen from 36 districts in nine provinces. The findings were as follows. First, schools as well as district and provincial education departments have little record o f expenditures, which made analysis difficult. There are inconsistencies o f records on financial allotments, notably for non-salary expenditures, because o f miscommunication between Kabul and provinces and between the Ministries o f Finance and Education. When available, records on expenditures were found to be largely consistent, however. Second, staffing controls through financial allotments were found to be effective. Three quarters o f the teachers receive their salary every month, even though a third o f them complain about delays. One third o fthem indicated that they had to pay a modest commission, usually to a bonded trustee, for the receipt o f their salary every month (on average equivalent to about tsl), a practice that was considered "normal" as a fee for travel expenditures. Third, it was confirmed that schools spent hardly anything on non-salary expenditures (books, maintenance, etc.). Kabul allots only 15% to non-salary expenditures; only 65% of this goes to provinces; provinces allot little if anything to lower levels; and Mustoufiats regularly reject requests from schools because o f cash constraints. In fact, hardlyhalfo f schools made a request for non-salary allotments inthe last quarter of 2004105, and half of the requests were rejected. In sum, schools incur almost zero non-salary expenditures, and they mainly get books, desks, chairs, etc. in kind, usually from non-governmental organizations (although sometimes from the Ministry o f Education itself). Given the uneven presence o f these organizations around the country, significant disparities were found across schools. Fourth, schools and even district and provincial departments do not provide much input into budget formulation. inparticular for the non-salary budget. However, they are often visited by the district or provincial department, sometimes even by the Ministry, to review performance or financial records. Finally, most schools report having a Parent-Teacher Association and a School Management Committee, usually to organize activities, sometimes to supervise teachers. Almost no school or teacher reported any payment from parents. Source: World Bank Education Expenditure Tracking Survey. 54 7.4 Much work remains to be done by MoF to ensure that service delivery unitsreceive the cash and information necessary to perform their operations. A first issue i s that o f quarterly allotments. Because their management i s unclear and not tied to cash availability, service delivery units sometimes receive conflicting information, and M o F may turn down their requests because o f misunderstanding or lack o f cash. A second issue i s cash management, which needs to be further strengthened to optimize the use o f resources and avoid unnecessarily constraining expenditures by service delivery units. Inthe medium term, these two instruments (allotments and cash planning) will have to be complemented by recording and controlling expenditures at the commitment stage, before it i s too late interms of incurring a liability. 7.5 An effective mechanismto make progress on this front would be to developfinancial reporting. Such information is useful for all managers inthe Government. The Cabinet and the President should get information to hold ministers accountable. The Afghan public, and soon the Parliament, should also get such data. Routinely, basic data on spending should be prepared from the Government accounting system. In addition, the use of public expenditure tracking surveys should be explored to focus on specific issues. The small-scale traclung survey on education conducted as part o f the PFM Review highlights the wealth o f analysis available from this instrument (Box 7.1). Another need i s for gender disaggregated data to facilitate proper monitoring of social inclusiveness. 7.6 The role of the central line ministries also i s problematic. Less than 20% o f the non-salary allocation i s given to provinces, whereas more than 50% o f the salary allocation (and staff) are in provinces (Figure 7.1; see also Box 8.2). This situation i s all the more problematic inthat provinces tend to have better capacity to absorb these resources, as their share o f estimated actual non-salary expenditures i s higher than their share o f allotments. In addition, local units have little influence over allotments, receive conflicting information on them, and are subject to unexpected changes in allotments. Local departments, since they have more local information and are themselves accountable to deliver services, should be given a role in the budget formulation process (Chapter 6) and receive clearer information on budget allocations. In turn, central ministries should be held accountable for the constraints they impose on service delivery units, for instance through frequent monitoring by M o F and the Cabinet on allocations made to provinces and the frequency of changes inthese allocations. Figure 7.1: Share of Budget Allocated to Provinces (% of Total Expenditur !S) 55% P ProvincialShare of Salaries ProLincialShare of Non-salary Exenditures Source MoF (AFMIS)and staffestimates. B. Controllingthe Use of Public Funds 7.7 A solid control framework should ensure that Government organizations are worlung within their legal and policy responsibilities and are achieving the results set for them. This implies that basic fiduciary standards must be observed, i.e. funds are spent only when there i s an approved budget, a secure process i s followed to authorize expenditures, and the process is documented. At a more advanced level, thisprocess will be simplified to shiftits focus towardperformance, while the systems themselves willbe 55 controlled continuously through audit. Treasury functions range from the commitment o f expenditures to recording, payment, and accounting. Internal audit provides a feedback loop that can foster improvements in expenditure management. The overall system needs to be complemented by effective enforcement mechanisms (Section D). TREASURY FUNCTIONS 7.8 Basic fiduciary standards for civilian recurrent expenditures have reached decent levels (Figure 7.2). Simple budget controls, accounting, recording, and reportinghave muchimproved, notably with the implementation o f a computerized system in M o F (Box 7.2). Progress has also been made on payroll control; a review by ARTF's Monitoring Agent suggests that there i s almost n o problem o f "ghost workers" in several central ministries. Another significant improvement i s the introduction o f a Treasury Single Account system (Chapter 4). Although still to be completed, this has already made cash management easier, bank account reconciliation simpler, and control o frevenue collection more robust. Figure 7.2: Share of ExpendituresReimbursedby ARTF 2a. Payroll 2b. Non-Salary I 2004105- 2004105- 2004105 2004105- 2005106- 2W5106 2003104 2004105- 2004105- 2004105- 2004105- 2005106- 2005106 I 2 3 4 1 2 1 2 3 4 1 2 I I 7.11 M o F i s implementing a "Verified Payroll Plan" to ensure timely and accurate salary payments to Government employees. More than 24,000 staff in 17 agencies inKabul are paid directly (as opposed to the bonded trustee system) and 400 o f them receive their salaries directly in their bank account; 6,000- plus employees in Kabul have received an individual identification card; and 10 central ministries are using a payroll database to prepare the payroll. The roll-out o f these three components - automated payroll, individual identification system, and transfer to bank accounts - will move payroll management a long way toward an effective, secure process. Box 7.2: CanE-GovernmentPlay a RoleinAfghanistan? In 2001, Afghanistan had very limited telephone land lines and computer skills. However, the Government had a great need to communicate effectively with district and provincial offices and to promptly deliver services. Although e-Government seemed promising - to deliver services quickly, reach remote districts, and transparently publishinformation- it seemed impossible due to lack of skills and infrastructure. Four examples demonstrate that this pessimism was unwarranted. The success o f cell phones demonstrates the ability o f modem technology to be quickly implemented and that demand for it is large; the mobile footprint now covers 50-60% o f Afghanistan's population, and mobile prices have declined by 70% during the last year and a half. In addition, a Government Communication Network has been set upinnineprovinces and is expectedto be expanded. A simple computerized general ledger was installed in record time in 2002. Initially set up to print checks, the system now records all expenditures and revenuesas well as budgets. It has allowed breakthroughs inreporting and transparency. Basic financial controls (e.g. checking the availability o f a budget) are now much easier. Annual financial statements were produced in the second year o f the transitional administration whereas none had been produced inprevious decades. Requests for reimbursements to ARTF were streamlined with this system, facilitating cash management. The system i s now operational inthe central MoF, two central ministries, and a pilot Mustoufiat. The Government has created a website (www.af, where a number o f important documents are posted. Inparticular, all budget decrees and fiscal reports are available publicly. This website can also be used for communication campaigns, notably a tax informationpage (httD:/iwww.mof. eov.af/taxiindex.htm). The ongoing implementation o f the Automated System for Customs Data (ASYCUDA) will increase transparency (easy reporting, international classification, performance monitoring) and control (including through a reduction in interface between customs officers and traders), while facilitating trade (more standardized, predictable, transparent procedures). Transit i s considered a priority: all goods entering into Afghanistan will be reported to the system and automatically discharged when they reach their destination. The first application is expected to be operational shortly; it will cover the transit corridor between the Torkham border point with Pakistan and Kabul and will be extended to the border with Uzbekistan. Some lessons emerge from experience in Afghanistan, as well as that o f other countries. First, e-Government i s no1 a panacea and needs to be well conceived. The "Systems Study" that has been initiated by MoF i s welcome since it will define a master plan for financial management systems. Second, e-Government works when it i s combined with improvements in business processes that seek to put the client (citizens, businesses, other parts o f the Government) at the center. Third, e-Government requires strong leadership at the higher levels, as implementation i s a long-term effort which disrupts habits, and widespread training at lower levels is needed. Fourth, some implementation issues - such as translation of English-based software or insufficient bandwidth for communication -have already emerged from early experience inAfghanistan and should be factored into any future proposal. 7.12 For non-salary expenditures, the most critical improvements are related to procurement (see next section). Other controls are well established inthe central M o F with the computerizedsystem but needto be rolled-out to provinces and line ministries, complemented by systematic commitment controls in the medium-term. 7.13 A final control issue is related to asset management, including the assets generated through donor-funded projects. The development o f a simple asset registry i s a priority for the Government if it wants to retain these assets and be able to account for them. Other asset management issues, related to SOEs and underground resources, are discussed inChapter 5 (see also Volume 111, Chapter 4 and Volume IV, Chapter 4). INTERNAL AUDIT FUNCTIONS 7.14 Reports from central ministries and from provinces suggest significant audit activities, by internal audit departments, the Control and Audit Office, the Attorney General's Office, and sometimes the 57 Governor or the National Security Office. Unfortunately, all o f these activities review transactions only to assess compliance with rules and detect errors, irrespective o f their cause or impact on performance. The only effective feedback loop has been an extraordinary arrangement whereby the ARTF Monitoring Agent, after reviewing expenditures to assess eligibility for reimbursement, transmits performance reports to MoF, which has tabled them at Cabinet meetings and has included a provision inthe Budget Decrees to reduce budgets o f agencies not performing against these eligibility criteria (Box 6.2). Improvements shown inFigure 7.2 are a testimony to the impact o fthis innovative practice. 7.15 Sustainable arrangements for internal audit needto be developed, however. The PFEML a w calls for a centralized system, in which MoF's internal audit department has authority to audit all Government operations. M o F has developed a plan to implement this rule and expects to focus on compliance audit. This seems appropriate in the short run given lack o f capacity. MoF's internal audit department will, however, need to strike a balance by promoting improvements in financial management practices without paralyzing the system through excessive intrusion in line ministries' management. Over the longer term, line ministries themselves would be expected to play a more active role in internal audit, with M o F shifting to systems andperformancereviews. C.Procuring goods and services 7.16 In2004/05, the Government spent more than $500 million on goods and services as part of the Core Budget (of which $100 million was for capital investments). Giventhe size o f these expenditures, a 10% increase in efficiency and value for money generated by procurement reforms would be equivalent to a 20% increase in domestic revenues. In addition to promoting efficiency gains, a good public procurement system should also adhere to key principles o f non-discrimination, equal treatment, and transparency. Although the rules in Afghanistan until recently fell short o f these objectives, actual performance has been more encouraging thanks to temporary external support and ongoing legal reforms. Nevertheless, there are numerous issues and difficulties, as illustrated by the experience with highways (Box 7.3). Box 7.3: Procurement for Highways A review of major road projects implemented inAfghanistan over the last three years reveals some key issues and constraints to obtaining value for money with adequate accountability and transparency. Major variances o f unit costs, from $123,000 to $589,000 per kmo f road, were found. First, all projects incurred higher than expected security costs (3-15% of total project costs). While there were also other exogenous factors such as weather conditions, security remains a major constraint on project implementation and value for money (with further implications for delays, safety of personnel, etc.). Second, the price, availability, and quality o f supply, often imported, raised unit costs. For example, unit costs for asphalt pavements were estimated to be 30-45% higher than inneighboring countries. Third, lack o f participation by the Afghan private sector in bidding processes has somewhat reduced competition and increased costs. Constraints include lack o f experience in managing large-scale projects and in following donors' financial management and procurement processes; lack o f up-front capital (or financing) to mobilize equipment; and difficulties inmobilizing construction labor. Fourth, many o f the projects were contracted with limited competition. Other factors also contribute to higher unit costs. For example, it was noted that implementation delays occurred because o f red-tape on the Government's side, for instance to clear the importation of equipment. Source: VolumeIV, Chapter 5. 7.17 The existing regulatory framework has some essential positive features, for instance basic financial controls and an emphasis on competitive bidding, but it i s outdated. There has been excessive focus on prices - at the expense o f quality - and an absence o f complaints and review mechanisms. Similarly, the practice of public procurement has positive features - such as some familiarity with modem basic processes, but also significant shortcomings - such as absence o f implementation regulations and monitoring mechanisms. The line ministries lack capacity to define and communicate effectively their 58 Constraint PossibleWay Forward Mistrust inthe way the public Simplify and disseminate rules (awareness campaign), increase sector operates transparency Cost o f doing business with Simplify procurement rules and develop standard contracts; promote e- Government (e.g. procurement procurement; develop direct payments to vendor (ongoing in the and payment delays) Treasury); focus on simplification o f procedures for small enterprises No quality standards Develop standards Poor capacity to prepare bids Prepare model contracts and detailed handbooks; train private sector Lack o f financial capacity Creating leasing businesses (e.g. through reforming tax rules); develop financial sector (require legal reforms) Lack of skilled labor Develop crash courses in selected topics; build-up capacity o f tertiary / vocational education system Preference given to SOEs IIDiscontinue these preferences 7.19 Turning to external support, an international firm, the Procurement Agent reporting to the Ministry o f Economy, has processed more than 200 contracts with a total value o f almost $400 million, mainly for projects financed by ARTF and IDA, but increasingly for projects financed by the Government itself. Performance under this arrangement has been reasonably good, using competitive methods, with adequate timing (to prepare and evaluate bids) and transparency (advertisement o f bids and publication o f awards). However, the overall performance o f this mechanism i s constrained by the limited capacity o f line ministries to prepare all the necessary documents (see above), as well as sometimes delays in obtaining requiredapprovals. Private sector participation has been limited - with typically less than five bidsper offer. Similar extraordinary external support is also available ina number o f line ministriesthat have set up Program Management Units (PMUs). The short-term benefits and potential longer-term shortcomings o f these practices are discussed inChapter 9. 7.20 In addition to building capacity, further progress requires improvements in the regulatory framework. The new Procurement Law that was recently approved by the Cabinet will establish a solid legal framework. The accompanying regulations, including handbooks, operational procedures, standard contracts, etc., will be critical for operationalizing this framework. Progress i s contingent on expanding Government ownership. The lack o f a clear champion in the Government, probably reflecting low capacity and misunderstanding o f the scope o f the procurement function (e.g. a belief that only major works or purchases are "procurement"), has been detrimental to reforms. Creation o f a Procurement Policy Unit in M o F would represent an important step forward. This unit will be able to take the lead in preparingdetailed regulations, overseeing training, and monitoringprocurement performance. D.HoldingGovernmentAccountable for the Use ofPublicFunds EXTERNAL SCRUTINY OF THE BUDGET 7.21 Development of the external audit function has been a key component o f the Government's efforts to enhance transparency. The PFEM Law mandates audit o f annual financial statements by an independent auditor. The Control and Audit Office (CAO), which reports directly to the President, i s performing this function. With the support o f an international firm, it has made progress in preparing audits, first for ARTF and then for the Government's accounts. It has also audited a number o f departments and has reportedits findings to the Cabinet twice a year. 59 7.22 Much remains to be done, however, to create demand for control and accountability. The CAO's capacity i s weak and its focus i s limited to basic financial audits o f annual accounts. Moreover, while M o F has been preparing follow-up matrixes, there i s little evidence of actual follow-up on audit findings, and reviews o f departments are never followed up. Part o f the explanation probably i s that audit reports are not publicly disseminated. This also relates to the absence, until recently, o f a Parliament to hold the Executive accountable for addressing audit findings. While the PFEM Law (and Constitution) mandate approval o f the Budget and review o f audited annual financial statements by Parliament, this has not been possible so far without a Parliament. The President and Cabinet have had legislative powers, but with obvious limitations in terms o f separation o f duties. Civil society - weak media, challenged NGOs, frequently unorganized communities - has been unable to perform its role either, in part due to lack o f capacity to analyze PFM issues. Donors have been active in promoting external scrutiny, but their unwillingness to use Government systems for most aid has undermined the Government's efforts to improve P F M performance. On the other hand, the ARTF has played an important role in enhancing external scrutiny o f budgetary spending financed by ARTF (Box 6.2). FIGHTING AGAINSTCORRUPTION 7.23 There are widespread allegations o f corruption, sometimes referring to unintended misuse o f funds (for example by error or lack o f capacity to follow the rules) or to traditional practices (for instance payment o f a commission to a bonded trustee who arranges salary payments, Box 7.1), sometimes referring to outright corruption. Box 7.4 reviews vulnerabilities to corruption. This issue i s part of the broader challenge of governance. As highlightedby an international comparison largely based on polls and surveys (and hence subject to significant margins o f uncertainty), Afghanistan ranks among the countries for which governance problems are the most severe (Figure 7.3). While significant progress has been made in some areas (voice, accountability, and Government effectiveness), Afghanistan remains among the bottom one-eighth o f all countries interms o f all six dimensions o f governance measured, and i s among the worst-off interms of rule of law, regulatory quality, and control o f corruption. Figure 7.3: Six Dimensions of Governance Voice and Rccountability Political Stability E Governnent Effectiveness E+ 1 b No data L Regulatory Quality No data Rule of Lau Control o f Corruption Kl e 25 58 75 1 6 Conparison betueen 2664, 2662, 1996 (top-botton ord -) Country's Percentile Rank (6-166) Note: The three bars represent,for each dimension of governance, Afghanistan's percentile rankfor 2004, 2002, and 1996 (top- bottom order). The line represents the standard deviation of this measure (in other words the margin of uncertain@). Source: Kaufmann, Daniel, Kraay, and Mastruzzi (2005). 60 Box 7.4: Vulnerabilities to Corruption While specific evidence o f corruption is difficult to find, concerns in a number o f areas point to important vulnerabilities to corruption in Afghanistan. For example, significant "unofficial payments" reportedly are paid by Afghan firms (World Bank, 2005e). Corruption is also one of the main concerns o f many Afghans (HRRAC, 2004). First, there are widespread allegations o f abuse of power, bribing, and other forms of corruption in thejustice and police system (HRRAC, 2004). Land management is an important vulnerability for corruption (Volume 111, Chapter 5). The recent Parliamentary elections also have generated rumors o f corruption. Second, by all indications there is massive corruption associated with the drug industry, with many Government officials directly or indirectly benefiting from narcotics revenues. Other criminal activities also are likely to be associated with (lower levels of) corruption. Third, the State has traditionally been built on a patronage system. While recent civil service reforms are tryng to reduce this factor (Chapter 9), the vulnerability to this form o f abuse o f power (power o fhiring) remains. Fourth, vulnerability in the area o f revenue collection has been reduced by a number o f administrative reforms (Chapter 4). Nevertheless, the existence o f numerous small "nuisance" taxes with obscure rules and the continuing collection o f revenues by actors other than the MoF create a climate o f uncertainty that i s conducive to corruption. The lack o f a clear privatization process could generate additional vulnerability to corruption if privatization occurs without this lacunabeing corrected (Chapter 3). Fifth, concerns with regard to the misuse of public funds have been much less widespread. However, anecdotal evidence - such as striking differences inprocurement delays in the power sector (unusually speedy process for one contract; unnecessarily protracted process for another, Volume IV) - suggest that vulnerabilities are real. Vulnerabilities seem more pronounced at the local level (notably in municipalities - see Volume I11- but also in provinces and districts). In addition, even with strong PFM processes, PFM institutions could still be captured b j private interests. Finally, these vulnerabilities are not yet well offset by opportunities for civil society to hold public officials accountable. While inthe PFM area the Government has beentransparent inreporting on budgets and expenditures the capacity of the media and the people to holdGovernment accountableremains to bedeveloped. 7.24 Corruption i s widely considered to be a symptom o f poor governance. This suggests that treating the symptom (corruption) directly most likely will be less effective than addressing the underlying problems that create or exacerbate vulnerabilities to corruption, including problems in the PFM system. International experience bears this out, although experience also demonstrates the importance o f setting a strong example at the top in terms o f talung action against high-level corruption on the part o f government leaders. 7.25 Tackling these issues requires a holistic approach, based on further analytical work and possibly leading to adoption of an anti-corruption strategy by the Government. International experience offers a number o f lessons for the design o f such a strategy. First, anti-corruption strategies work better through prevention than through investigation and prosecution. 7.26 Second, sound anti-corruption strategies need to cover five pillars:l5(i) institutional restraints on power (independent and effective judicial system and prosecution and enforcement, legislative oversight); (ii)political accountability (political competition, credible political parties, transparency in party financing, asset declaration, conflict o f interest rules); (iii)civil society participation (freedom o f information, public hearings on draft laws, active role o f media and NGOs); (iv) competitive private sector (economic policy reform, competitive restructuring o f monopolies, regulatory simplification for entry, transparency in corporate governance, collective business associations); and (v) public sector management (meritocratic civil service with monetized, adequate pay; budget management - coverage, treasury, procurement, audit; tax and customs; sectoral service delivery - e.g. health, education, energy; decentralization with accountability). While this report focuses on the last pillar, the other four pillars include a number o f areas where progress i s needed. Examples related to PFM include a strengthened judicial system, rigorous processes o f legislative oversight, transparency to enhance civil society participation, simplification o f regulations, and a merit-based civil service. See the World Bank's website on anti-corruption strategy, httu://www1.worldbank.ore/Rublicsector/anticorrupt/strategies.htm. 61 7.27 Third, improvements in PFM practices will help very much in reducing vulnerabilities to corruption, by limiting transactions that provide opportunities for graft, reducing likely benefits from corruption, and enhancing information, transparency, and oversight. This report includes recommendations to simplify taxation rules and strengthen revenue administration (Chapter 4); improve the coverage o f the budget (Chapter 6); develop internal controls, procurement, reporting, audit, and external scrutiny (this chapter); explore innovative accountability mechanisms (Chapter 8); and strengthen the management of SOEs and municipalities (Chapters 3 and 8, see also Volume 111). 7.28 Finally, while these recommendations should constitute the core o f the anti-corruption strategy, an additional question i s whether an Anti-Corruption Commission would play a useful role, such as the Commission created in Afghanistan by the 2004 Anti-Corruption Law, which reports to the President. While there are some examples in other countries o f such approaches having a positive impact, many experiences have been less positive, with sometimes the commission losing all credibility in its fight against corruption. A first issue inAfghanistan i s that the 2004 Law does not distinguishbetween corrupt practices and mistakes or negligence. A second issue i s the lack of capacity to prosecute and try corruption cases. Additional issues include risks o f politicization, predation, diversion o f attention and resources from other necessary areas, and bureaucratic duplication. 7.29 I n summary, enhancing the effectiveness of budget execution should be at the center of the governance and anti-corruption strategy. Among the key priorities - both to increase the effectiveness o f public spending and to reduce vulnerabilities to corruption - are a further increase in transparency (transparent allotments to service delivery units; release o f annual financial statements and external audit opinions; publication o f bid requests and contract awards), strengthening o f the audit function (both internal and external), and drafting of procurement regulations. The recent approval o f the PFEM Law and the Procurement Law provides the legal foundation for pursuing these priorities. On this basis, capacity o f line ministries (and o f the private sector to participate in public procurement) will need to be enhanced (Chapter 9) and PFM responsibilities gradually devolved to line ministries; the role o f the Parliament as a promoter o f accountability can be developed; and, over time, more ambitious PFM reforms can be introduced. 62 CHAPTER8. DELIVERINGSERVICESTO THEAFGHAN PEOPLE 8.1 The ultimate objective o f the PFM system i s to deliver services to the Afghan people. Service delivery can take various forms, ranging from direct delivery by Government-paid civil servants, to delivery by non-governmental organizations financed from private or foreign Contributions, to private provision funded by user fees. While the main message o f this chapter i s unequivocal-people shouldbe put at the center o f the service delivery framework -the institutional implications are more diverse. The chapter reviews the institutional framework for public service delivery, buildingon examples o f particular sectors and drawing lessons from experience. It also looks at PFM issues related to subnational administration from a service delivery perspective. A. Six Models of Public Service Delivery 8.2 A simple conceptual framework illustrates different models of service delivery and their implications (see World Bank, 2003b). There are three main actors: the people (as citizens and consumerdclients o f services), the state (as political body and policymaker), and the service providers. The interactions by which these three actors influence and are accountable to each other form an "accountability triangle" (Figure 8.1a), and there are correspondingly financial flows among them (Figure 8.lb). The people can affect service delivery by influencing policy makers (e.g. by voting or through the advocacy role o f civil society), who in turn exert influence on the service provider (this i s the so-called "long route" o f accountability). The people can also directly influence service providers (the "short route" o f accountability) by selecting the provider (when there i s competition), using their voice (e.g. complaints), or malung financial contributions (when there i s cost recovery). Another important set o f actors i s the donor community, which exerts influence through its dialogue with policymakers, its financing, and its direct contracts with service providers. Figure 8.1: Service Delivery Framework a. Accountability b. Financing Choose; Assess service Source: Adaptedfrom World Bank (2003b) CENTRALIZEDGOVERNMENT SERVICE DELIVERY 8.3 The "traditional" service delivery model in Afghanistan i s that the State itself provides services through its centralized bureaucracy (Figure 8.2). Ninety-sevenpercent o f Afghan students are enrolled in public schools operated by civil servants under the management of the Ministry o f Education (MoE). This model directly links the Government with the population and thereby may enhance the former's perceived legitimacy; it facilitates propagation o f nationwide service content (e.g. curriculum); and it may be able to exploit economies o f scale. Certainly Afghanistan has seen an unprecedented expansion o f elementary education since 2001, based on the application o f this centralized model (Figure 8.3 - also see Volume IV, Chapter 2). Both girls' and boys' enrollments have reached levels far higher than at any time inAfghanistan's history, includingprior to the conflict. 63 Figure 8.2: Service Delivery by Central Government (Education) Payroll, taxation \\ Management 8.4 The centralized service delivery model has major weaknesses, however. Management o f the primary education system, for example, i s highly centralized. All important decisions, and even relatively less important decisions like appointments o f teachers, are made by M o E inKabul, often a very long distance and at least three management layers away from the schools and the students they serve. Schools have very little access to non-salary budget allocations, and many inputs are provided in-kindby donors and NGOs (Box 7.1) - leading to disadvantages associated with the "outside government" mode o f service delivery discussed later. The distribution o f expenditures, especially non-salary expenditures, between Kabul and the provinces i s skewed against the providers, and there are major gaps in service delivery between urban and rural areas (which when combined with gender and regional disparities become enormous).'6 The needs o f each school are not effectively responded to by distant decision- makers. The efficiency of the centralized system is reducedby management difficulties, inparticular how to hold teachers and principals accountable for their work from a distance. Another issue i s the role o f donors, who may provide financing to the State for education (through ARTF or project financing), but also may provide salary top-ups, direct payments, and in-kind contributions directly to schools. The result o f all o f these shortcomings i s poor quality o f education and serious geographical as well as gender disparities (see Figure 5.3), reducingthe returns to the large investments being made ineducation. Figure 8.3: Enrollment Growth (Grades 1-12) 5,000,000 4,500,000 4,000,000 3,500,000 3,000,000 2,500.000 2,000.000 1,500,000 1,000,000 500.000 0 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1993 1999 2002 2003 2004 Source: MoE, UNESCO,and UNICEFfor variousyears 8.5 This centralized model, or variants o f it, also exists in other sectors, a legacy o f Afghanistan's history. Whereas in education (and in the pre-war health system), as well as inthe military and national police, the service provider i s an arm o f the central ministry, inother cases the provider may be a State- l6In the health sector, similarly large disparities can be seen in Government-provided services like doctors and hospitals. For example, there i s one doctor per thousand population in Kabul, one per 100,000 in BamiyanProvince. Operating expenditures of the Ministryof Public Health (MoPH) per capita of population are more than four times as high inKabul as the national average, and there i s great variation acrossprovinces as well (see Volume 11, Chapter 1, Figure 1.11). 64 Owned Enterprise (SOE) closely supervised by a parent ministry (e.g. power, water supply, major imgation schemes). In the case of electricity, the national power utility, Da Afghanistan Breshna Moasessa (DABM), i s supervised by the Ministry o f Energy and Water (MoEW) and has neither an appropriate governance structure nor financial resources to improve the country's electricity services (Volume IV, Chapter 3). The result i s very limited access to power (only an estimated 10% o f the population has access to gnd power supply) and poor quality o f supply (unreliable service, frequent interruptions, and hightechnical losses). 8.6 There are several ways to improve service delivery while staying within this basic model. First and foremost, mechanisms need to be created whereby the people can use their voice directly with the service providers and monitor them. Good examples in the education sector include Parent-Teacher Associations / School Management Committees. Community monitoring, for example through the elected Community Development Councils (CDCs), i s a promising approach for certain sectors. Second, within the State, options include de-concentration and delegation to improve budget allocation decisions and bringmanagement closer to the provider. Inthis context, ensuring that non-salary budget allocations reach and can be utilized by the service providers (e.g. schools) i s a priority. Greater management autonomy for service facilities can yield good dividends based on international experience, including through flexibility in managing and utilizing the budgetary resources made available to them. Third, information sharing and greater transparency (both from the state and from the service providers) can enhance the voice of the Afghan people, and a free and capable press and other watchdogs can enhance oversight and accountability. Donors providing financing need to respect the Government's systems and work through them rather than providing resources directly to providers outside the Government's ambit (which undermines the accountability relationship between service provider and State). 8.7 Short o f privatization (the fourth model, see below), better financial management o f SOEs, along with higher cost recovery as appropriate, can improve service delivery. In the power sector, required reforms include corporatization o f DABM; computerization o f billing and accounts; enhanced cost recovery; capacity building in MoEW; and prioritization of investment projects. Private sector participation inpower generation also can be pursued on a limitedbasis. OUTSIDEGOVERNMENT 8.8 This second model is common inconflict and post-conflict countries, with a failed State. Donors directly finance and contract or manage non-government providers (NGOs, private providers, sometimes international agencies e.g. UN agencies) to deliver services, often o f a humanitarian nature, with little or no direct involvement on the part o f the Government (Figure 8.4). This model i s still being used for a number o f activities in Afghanistan, but much less than inthe 1990s when it was the primary mode of international assistance. A variant consists o f setting up a narrow Project Implementation Unit (PIST), with weak or no linkages to the Government, specifically to oversee a donor-financedproject. And even in sectors where the basic model is different (e.g. centralized government service delivery ineducation, see above), there may be elements o f direct donor involvement through direct provision o f various inputs, financial top-ups, etc. The appeal o f this model i s its simplicity and that it i s straightforward and visible for donors as well as service recipients. Moreover, by bypassing Government procedures (and often normal international competitive procurement procedures), this model enables a rapid response to humanitarian crises and other urgentproblems. 8.9 However, this approach does not support longer-term national capacity building and does not exploit the opportunity for aid to help build up the State's legitimacy but on the contrary undermines it. From a policy point o f view, it also defeats the purpose o f policy coordination discussed in Chapter 6. Finally, it can lead to higher unit costs (for example in the case o f building schools) and weaker cost controls. Thus this model i s not desirable beyond short-term emergency and humanitarian interventions. 8.10 The overallthrust o f Afghanistan's development strategy duringthe past several years has beento move away fi-om humanitarian interventions isolated from the Government and toward longer-term development activities under Government leadership. The way forward need not be through direct 65 Government provision, however, but rather through the Government taking policy leadership and control of the contracting and funding process vis-&vis non-government service providers, thereby making the latter accountable to the State and supportive o f the Government's efforts to build up its credibility and legitimacy. This i s the thirdmodel of service delivery, discussed below. There i s inany case a great need for strengthening monitoring o f service providers by the people and their communities. For PIUS,the best approach - strongly advocated by the Government - i s that they be fully integrated in the concerned line Ministry and provide services (e.g. procurement, financial management) to its entire development programrather than only to a specific donor-funded project or program (Chapter 9). Figure 8.4: Service Delivery Outside Government (Humanitarian) ,I' Direct paymentsand lmrnagement Somm local influmnsm GOVERNMENT OUTSOURCINGTO NGOS/PFUVATE SECTOR 8.11 Inthe thirdmodel, the Government is incharge of (and accountable for) policy, financing, and monitoring, but non-government service providers (NGOs, non-profit entities, or private firms) are put in charge o f delivery on a competitive contractual basis by the Government (Figure 8.5). This model i s being applied in the health sector, for delivery o f the Basic Package o f Health Services (BPHS) through Performance-based Partnership Agreements (PPAs) ineight provinces. Figure 8. 5: Service Delivery by Outsourcing to NGOs (part of Health services) i Finance, taxation Contract, Manage L- a . , '.,.: Some local influence 8.12 There can be significant efficiency gains from this model as compared with centralized Government provision of services (the first model). A review o f global experience found that, in selected examples, unit costs for contractual provision o f health services were 11-50% lower than for government provision (Loevinsohn and Harding, 2004). InAfghanistan, the per-capita cost o f the BPHS covered by PPAs averages around $3.80, which i s fully in line with patterns o f health expenditure in other low- income countries. A competitive process overseen by the Government thus contains unit costs within reasonable limits. By contrast, the same BPHS package provided under donor-executed projects in the External Budget have more widely varying and often substantially higher unit costs (Table 8.1). Thus while it is too early to come to very strong conclusions, it appears that open competition for contracts to provide basic health services holds down unit costs and reduces the variance incosts. 8.13 Moreover, as compared with the second model, this approach provides for Government leadership on policy, contracting, and financing, and thereby potentially also can enhance the legitimacy o f the Government. In addition, this model can exploit existing capacity in the non-government sector and harnesses it to meetingnational development priorities, as has occurred inthe health sector. 66 DonorProgram Unweighted Average Lowest Highest Ministryof PublicHealthStrengthening 4.1 d a d a Mechanism(World Bank-supported) Performance-basedPartnershipAgreements 3.8 3.3 5.2 (WorldBank-supported) Rural ExpansionofAfghanistan's 6.1 0.3 46.5 Community-basedHealthcare(USAID) Support to HealthService Delivery in 4.2 2.2 5.3 Afghanistan (EuropeanCommission) 'Asian DevelopmentBank 4.8 d a d a It should be notedthat these figuresrepresent contracted levels; actualunit costs may differ and inthe case of PPAs appear to be lower than contractual levels. Also, cost figures are not strictly comparableacross donors/programs; for examplethe costs for the USAID-REACH programdo not include the cost of drugs and most training (which are included in the PPAs), whereas the EC programincludessome training and constructioncosts. Overhead costs of running different programs vary widely, ranging from $1.5 million duringthree years for the first two programs(runby MOPH) to $23 million for the USAID-REACHprogram, althoughthe latterincludes extraTA to MOPH (Stronget al, 2005). This includes the possibility of trylng out various models for increasing gender targeting and inclusion, both in terms of staffing and beneficiaries, drawing on experiences from the NGO sector. This will lead to greater potential for inclusive strategies(vulnerablegroups, minorities etc.) 67 appropriate regulatory frameworks can be a technical and political challenge. Effective competition, as has been achieved in the case o f telecommunications, i s critically important. Where competition i s impossible (in some network industries), the pressure on the regulatory framework to ensure efficiency and reasonable pricing i s all the greater.lg Regular monitoring and evaluation by the Government, includingreliance on feedback from customers, can play an important role inthis regard. Figure 8.6: Service Delivery by Private Sector (Telecommunications) Choose and Pay Box 8.1:Forms of Private Sector Participation inInfrastructureKJtilitySectors There are numerous forms o f public-private partnerships. The simplified figure below shows the main categories, where the extent of participation o f the private sector grows from left to right: Public Private Partnership 7 Works Build & Operate Services Transfer Low b High Extent of Private Sector Partnership Works and services contract: A works or services contract i s an arrangement in which the public utility contracts out specific works (e.g. rehabilitating and expanding a distribution system, construction o f a new power plant, etc.) or services (e.g. technical assessmentsor project management for large works contracts) to the private sector. Operation and maintenance contract: An operation and maintenance contract i s an arrangement by which a private company is entrusted with various types o f tasks usually performed by the public authority, such as day-to-day operation and maintenance o f existing electricity operations (e.g. a power plant). Concessions; A concession i s an arrangement under which a public entity, owner o f the asset concerned (e.g. power generation plant, transmission line, etc.), delegates to a private entity (concessionaire) the responsibility for providing and maintaining a specified level o f service to users in exchange for the right to collect revenue from the users. Unlikethe previous forms o f public-private partnerships, a concession shifts some o f the financial risk to the private sector. Concessions may take various forms: Under a Build Operate Transfer (BOT) concession, the responsibility o f the concessionaire is not limited tc operation and maintenance o f the infrastructure but also includes initial construction, upgrading, or majoI rehabilitation component. Large investment and consequent mobilization o f private funding sources is therefore required from this company and i s to be repaid from the revenue collected from users. BOT stresses thc responsibility o f the private entity during construction and operation o f the asset and the handing over (transfer: o f the assets to the public entity at the end o f the concession period. The high initial investment required from the private sector and the consequent long concessionperiod make the distribution o frisk between the parties i key element o f success in such schemes. Many variations o f this type o f contract have been implemented: . 9 . BOO: Build, Own, and Operate type project financing BOOT: Build, Own, Operate, and Transfer type project financing BOT: Build, Own, and Transfer type project financing Source: VolumeIV,Chapter 3. It may be possible to unpackage services so that competition can be introduced at least for some components (e.g. power generation as opposed to transmission and local distribution). Moreover, in some cases introducing contestability (periodically re-opening service contracts or other arrangements for bidding on a competitive basis, so that a monopoly is not permanent) can reap benefits associatedwith competition. This approach does, however, raise important technical design issues. 68 8.18 Forpublic utilities, there i s actually a range o f options betweenservice delivery by a Government enterprise or department and purely private service delivery. Inparticular, various types o f contractual arrangements involving different degrees o f private sector participationare possible (Box 8.1). COMMUNITY-BASED 8.19 The fifth model involves community decision-making and leadership (Figure 8.7). In the National Solidarity Program (NSP), the most notable example in Afghanistan, communities elect by secret ballot Community Development Councils (CDCs) which prioritize community development needs with support from a facilitating partner. The communities are then provided block grants from the Government which are used for the identifiedpriority activities. Implementation i s overseen by the C D C itself. 8.20 This approach has turned out to be a powerful instrument for building community social and governance capital, empowering rural people, and enhancing the legitimacy and credibility o f the State. Moreover, there are built-in incentives for the communities to contain costs. In addition to small-scale rural roads at the village level, NSP has hnded other small-scale infrastructure (for example irrigation, drinhng water, sanitation, rural energy). This approach also holds the biggest promise for involving women on a widespread basis across the country in local level decision-making and as active participants ineconomic activities (Boesen, 2004). Figure8.7: Service Delivery by Communities (National Solidarity Program) / Finance Some 1 \\ Finance, taxation Elect, Oversee 8.21 Although the NSP has achieved good initial results, there are significant issues: 0 Continuity - What will be the roles and responsibilities o f CDCs beyond their work with respect to block grants? H o w will the community assets created under NSP be maintained? 0 Capacity - H o w much and how quickly can the CDCs, a very new and possibly fragile institution, take on additional activities? 0 Coordination - H o w can it be ensured that community development activities mesh with development activities o f other local actors, for example with respect to local roadnetworks? 0 Scale - Are there possibilities for communities/CDCs to be involved in larger activities that encompass more than one community? H o w can this be managed? 0 Linkages with local administration - What are and what should be the linkages between CDCs and district administration? There are bothbenefits and risks from such linkages.20 8.22 The way forward with respect to NSP would therefore involve, first, a review of the implementation o f NSP so far, its accomplishments and shortcomings, including an honest assessment o f the capacity and potential o f CDCs. An enhanced, structured role for CDCs inmonitoring local service delivery by other (State and non-government) organizations inthe village would appear to be a promising option. On the other hand, expanding the role o f CDCs very widely into, for example, credit provision or revenue collection would appear to be excessively burdensome andrisky. 2o Potential benefits include infusing a more participatory, accountable approach into local administration; potential risks include "bureaucratization" o f CDCs if they come to be seen as defacto the lowest level o f government administration, and increased possibility of "capture" by local powerbrokers. 69 MUNICIPAL SERVICES 8.23 The sixth model (Figure 8.8) pertains to municipal authorities, which are the only level of government in Afghanistan with a measure o f de jure autonomy (Chapter 3). They have their own revenues to fund municipal government costs and delivery o f certain municipal services (e.g. trash collection, recreation, park services). However, their tax and fee rates are set by MoF, and they are under the supervision o f the Ministry of Interior which approves their budgets (except for Kabul Municipality which has the status o f a ministry itself). This model has some promising aspects, since the management o f the service provider, and inprinciple the political accountability, are located close to where the service is delivered, potentially strengthening both the "long" and "short" routes o f accountability. Moreover, municipalities' autonomous budgets and their fees and levies could provide some linkage between fundingand service delivery. Figure 8.8: Service Delivery by Municipalities Some Finance, taxation Vote and Pay Taxms 8.24 However, the municipalities o f Afghanistan, exemplified by Kabul, face a host of problems and constraints, including most notably the following: 0 Confused roles vis-&vis central ministries, especially the Ministry o f Urban Development, that are involved in planning and investments and sometimes supervising service delivery (e.g. water supply). Some of this confusion stems from lack of clarity inthe legal Eramework. 0 Dj.sfunctiona1 management arrangements with municipal service providers. For example, in Kabul Municipality the current institutional arrangements and existing departments do not adequately address the full range of service delivery challenges. Several key functions are not being performed, such as environmental sanitation and property management. Although solid waste collection, maintenance o f urban roads, cleaning o f ditches and drains, and apparently some responsibilities with respect to water supply and sewerage are handled by a couple o f departments in the Municipality, these departments are hamstrung by poor capacity, lack o f funding for capital investment projects, excessive bureaucracy (i.e. being subject to the full gamut of administrative procedures), and lack of incentives to effectively provide services.2' Service providers seem to have very little in the way o f management autonomy, incentives, or accountability for service delivery, either to the Municipality or to beneficiaries. e Lack of political accountability to the urban population. Although elections are the ultimate form o f political accountability and elected Mayors and city councils are called for in Afghanistan's Constitution, Mayors of Municipalities currently are appointed by the administration instead o f being elected. 0 Weak governance, budgeting, and financial management. In Kabul, budgets are prepared mechanically and never revised within the financial year; expenditures are constrained by cash availability; and the Mayor can approve expenditures outside the budget. Capacity i s sorely lacking in the vast bulk of municipalities. A s a result, municipalities are vulnerable to corruption (Box 7.4). "Infact, oneofthemunicipaldepartmentsresponsibleforservicedeliveryinKabulwaspreviouslyanSOEbutwastransformed into a department of the municipal government in 2004. 70 8.25 The reform agenda for the municipalities i s formidable, but an emphasis on service delivery will help provide focus. Key priorities include (i) clarifying responsibilities (in particular between larger municipalities like Kabul and the Ministryo f Urban Development); (ii) improving the budget process and execution controls; (iii)building human and organizational capacity in core municipal functions (including P F M functions); (iv) reforming the relationship between municipal governments and municipal service providers, including consideration o f contracting for private or community provision where appropriate; and (v) ensuring effective outreach and response to community needs, especially those o f vulnerable and marginal groups and women. B. Lessons andImplications 8.26 As demonstrated in the previous section, there are a variety o f different modes o f service delivery, and rich experience both in Afghanistan and internationally. The first lesson i s precisely that "one size does notfit all" - different approaches can and should be used depending on the characteristics o f the service concerned and the specific context. However, it i s essential that the underlying accountability relationships support and encourage adequate service provision. 8.27 A second lesson i s that, ingeneral, serviceproviders need to be made more accountable directly to consumersheneflciaries, through enhanced channels for voice, choice o f providers, or in some cases bypaying for the service and thereby having a greater incentiveto be concerned about quality. The need to strengthen the "short route" o f accountability applies to most o f the different models outlined above, and to virtually all types o f services - a strihng example inthe security sector i s the police. 8.28 Third, different models of service delivery have different implications for state-building. Several dimensions seem to be important inthis regard: (i) actual role playedby the Government (e.g. the policy, financing, management, provision); (ii)the visibility o f the Government in relation to consumersheneficiaries; and (iii) the effectiveness o f the services (access, quality, inclusiveness with respect to women and other vulnerable groups, etc.). Except for the second model (delivery completely outside government), the Government does play an important role o f one lund or another in all of the models, but in some o f them (e.g. outsourcing to NGOs or private sector provision) there may be a need for measures to enhance the Government's visibility. 8.29 Fourth, where possible competition (or at least contestability) among providers can play a useful role. Mobile telecommunications i s a notable example o f competition resulting in major improvements in access, quality, and efficiency. In health, competitive contracting (even though there ends up being only a single service provider ineach province) can help control unit costs. 8.30 Fifth, although it seriously constrains service delivery in the short run, the lack of capacity in traditional centralized service delivery systems provides an opportunity to rethink, modifi, or move awayfrom this model. Inmost sectors (whether social services or public infrastructure), Afghanistan has much more o f flexibility in choosing options than other countries with well-established service delivery institutions and networks. Depending on the sector, this may involve de-concentration within the Government hierarchy, moving toward more commercialized modes o f provision by a utility, private sector participation, etc. 8.31 Thereare important lessonsfor donors as well, including the need to work through Government budget channels and not build up parallel structures, avoid undermining service providers' relationship with the State by providing resources (including salary top-ups) directly to providers, resist "flagging" o f projects/activities that they finance to the detriment o f the Government's visibility, and encourage rather than undermine cost containment, for example by supporting competition in their own and Government procurement procedures. While the need for quick impacts and concerns about weak Government capacity are valid, the various models outlined above highlight the importance o f placing donor interventions under Government leadership and, when possible, using Government institutions (Chapter 6 discusses the use o f the Government's budget andbudgetary systems). 71 FINANCINGSERVICEDELIVERY THEROLEOFCOSTRECOVERY OF AND 8.32 Financing and cost recovery deserve attention from both fiscal and service delivery perspectives. The fiscal impact of downstream O&M requirements associated with public investments and (re)building o f social service networks can be truly daunting (Chapter 3), and cost recovery where appropriate can be an instrument for managing these downstream costs and reducing associated fiscal risks. Cost recovery can also be an instrument for demand containment (for example adequate power and water tariffs, and nominal charges for health facility use or higher education to discourage unnecessary or excessive use). 8.33 Whether and to what extent it i s appropriate to charge for any given service depends on the characteristics o f the service (whether people can be excluded from benefiting from the service or not - e.g. defense versus domestic pipedwater supply, whether charges can vary with usage), demand (whether not charging will result in overuse, e.g. in the case o f electricity), and consumers (whether poor and non- poor consumers can be distinguished?whether the poor are disproportionate - or at least equal -users e.g. primary education?etc.).** Inmany cases significant cost recovery may be neither feasible nor desirable - Government-provided primary education i s one o f the best examples. Nevertheless, there are good technical grounds for additionalcost recovery for many public services. 8.34 Inaddition to its fiscal implications, cost recovery can be a tool o f accountability for improving service delivery. People who pay for a service are more likely to be concerned about whether and at what quality level they are benefiting from the service, and to be more vigilant and selective, positively influencing the behavior o f providers. These beneficial effects depend on whether there i s competition among providers (and therefore the possibility o f exercise o f choice by the consumer), on other factors that enable payment for services and exercise o f voice to reinforce each other, and on receipts transparently accruing to the service provider. 8.35 Turning to some examples, although willingness to pay for power is demonstrated throughout Afghanistan (widespread use o f expensive small generators with full cost recovery), in Kabul (and other largely hydel-based cities like Kandahar) the effective gnd tariff i s only $O.O2l/kwh, far below the actual cost o f supply. Nationally, the effective tariff i s roughly estimated at $0.05 l h h , compared to a cost o f supply o f $0.123/kwh7 implyingan average cost recovery rate o f 40% (Table 8.2). Although donors are funding various power investments as well as recurrent costs such as fuel (diesel fuel providedbyUSAID for the Kabul Northwest Power Plant cost about $40 million in 2004/05), this pattern o f financing i s fragmented and unsustainable. Region Effective Share of PowerBilledas Tariff costs YOof System Total (USVkwh) Kabul* 2.1 15.4 49% IKunduz I 4.6 I 3.2 I 4% I Balkh 7.4 6.5 17% Herat** 2.5 5.2 7% Nangarhar 7.3 1.3 6% Note: Data apply to DABM. Source: Volume I E Chapter 3. 8.36 Thus there is an urgent need to raise power tariffs where they are very low. Moreover, it i s important not to exacerbate the problem o f low cost recovery. For example, M o E W recently purchased several diesel generating sets for use inKabul. Without higheruser charges, it will be impossible to cover 22 See World Bank (2003b, Box 4.4, p. 71) for a schematic summary of these considerations (reproduced and discussed in the contextofAfghanistaninWorld Bank, 2005a, p. 85). 72 the estimated $10 million annual cost o f diesel fuel for these generators. In June 2005, MoEW sharply reduced power tariffs (by on the order o f40%) inHerat, reflecting the lower cost o f imported power from Iran. This will damage the ability o f the power system in Herat to finance expansion o f its distribution network to improve coverage. These examples highlight the importance o f fully incorporating financing and cost recovery issues indecisions. 8.37 The natural gas sector also faces very serious financing and cost recovery issues. As noted in Box 8.2, private investment based on commercial pricing will be essential if this sector i s to be revived and developed. Currently natural gas tariffs are far too low to fund rehabilitation, O&M, and necessary expansion o f gas production and infrastructure. Moreover, non-payment by end-users is common even at present low prices, raising serious doubts about whether they will be able to pay full cost-recovery prices. It seems clear that development o f this sector will require a set o f complementary investments including indownstream end-user industries,basedon effective Government regulationandeconomic pricing. Box 8.2: Natural Gas Pricing After years o f neglect and underinvestment, many steps are needed to revive and develop Afghanistan's natural gas sector. Given the characteristics and investment requirements o f the sector, large amounts o f private investment will be essential. Two critical first steps for attracting private investment in natural gas are (i) pricing natural gas on commercial terms, and (ii)establishing a track record o f full and timely payment for gas by end-users. Gas tariffs today are too low to cover costs o f rehabilitation, O&M, and expansion o f gas production and infrastructure. For incremental gas production, tariffs may be less than half o f what they ought to be. Moreover, the Government collects no taxes or royalties from the natural gas sector, missing out on a potentially substantial source o f revenue. Gas transmission and distribution tariffs tend to be regulated by the Government in all countries. Gas prices at the wellhead are generally determined by market forces, although some Governments such as Pakistan have established pricing formulas to attract investors. Determination o f economic natural gas prices requires, among other technical work, separation of reasonable and prudent costs from unnecessary expenditures for cost recovery purposes. For the benefits o f regulation to exceed its costs, regulation should be simple, workable, and cost-effective. The Government needs to strike a balance between abuse o f market power (in the absence o f competition and effective regulation) and over-regulation, in light o f the available resources and its own capacity to regulate. A closely related question is the ability and willingness o f consumers to pay prices that allow cost recovery. International experience demonstrates that creditworthy large-volume consumers form an important market base for the natural gas industry. In Afghanistan, non-payment by consumers has been a serious problem despite the current very low gas price. It is not clear whether current consumers, inparticular the Kud Bergh fertilizer/power plant (the only large-volume consumer at present), can pay for gas at economic pricing levels. There is a need to identify potential creditworthy large-volume gas consumers to help develop the gas sector. A new power plant based on natural gas may be such a candidate. At the same time, what to do about current gas users who may not be in a positionto pay for gas at cost-recovery levels needs to be carefully considered. Finally, the discussion above strongly suggests that development o f the natural gas sector in Afghanistan will have to involve a set o f complementary private investments in gas production, infrastructure, and major gas end-use activities, with effective Government policies and regulation. 8.38 Higher education provides a good illustration from the social sectors of the need for cost recovery and sustainable financing arrangements. Given its enormous needs and the high socio-economic returns to investments in elementary education, this i s being prioritized in terms o f budgetary allocations (see Chapter 5). But delivering quality higher education services leading to significant numbers o f well- qualified graduates will be essential for state-building and capacity development in both public and private sectors. While charging fees for higher education i s not popular inany country, the private returns to higher education are substantial enough that partial cost-recovery i s well-justified, with scholarship or loan programs for those unable to afford fees. Moreover, an astonishing 40% o f the recurrent budget o f the Ministry o f Higher Education i s spent on running costs and food at university dormitories. Since another 50% o f the operating budget o f the Ministry goes for salaries, very little i s left for indispensable pedagogical inputs such as internet access, textbooks, journals, lab materials, etc. Even if charging fees for higher education i s not possible inthe short run, expecting students to cover part o f living costs - at least food - would be reasonable and would fkee budgetary resources to improve the quality o f higher education. 73 MONITORINGANDEVALUATION 8.39 Information flows and monitoring and evaluation (M&E) are critically important for effective service delivery, as shown in the various models and examples discussed. Some progress has been made inimproving statistics and information flows, butthere is a long way to go infurther improving these and exploiting them for effective M&E leading to better service delivery. In particular, there have been substantial improvements in flows o f financial information on expenditures (Chapter 7). Good examples o f progress in developing timely information flows include the Health Management Information System and Education Management Information System. But little information i s available on outputs and outcomes. Moreover, gender-disaggregated data is largely lacking. These deficiencies will need to be corrected over time. 8.40 Transparency and dissemination o f information are extremely important as well. In addition to, and often more important than, vertical flows o f information to and within the administrative hierarchy, information needs to flow to the public and civil society in a horizontal manner. 8.41 Better information flows are only part o f the story, however. Appropriate performance benchmarks need to be set and regularly monitored. For example, performance benchmarks will help in identifying where bottlenecks occur, e.g. in terms o f girls' retention rates in schools, women's access to maternal health services, causes o f high maternal mortality, etc. Setting the right benchmarks requires a clear strategy and objectives. Indeed, setting benchmarks and monitoring performance can be a good device to clarify objectives and make them truly operational. 8.42 Finally, institutions need to take on the role o f monitoring and evaluation and treat it as a serious part o f their work. This is clearly one o f the important roles o f Parliament at the national level. The elected Provincial Councils can and should play a role inmonitoring at the provincial level, and similarly the elected District Councils (mandated under the Constitutionbutto be formed later) at the district level. At the community level, as discussed above the CDCs could play a very important monitoring role vis-&- vis local delivery o f services. C. Sub-National Strategy 8.43 Afghanistan i s dejure a highly centralized state, with lower levels o f administration serving as arms o f central ministries. Provincial Governors are supposed to play a coordinating role and are responsible for security, but they report to the Ministryo f Interior, and they have no supervisory authority over line departments in provinces. Only municipalities have some degree o f fiscal autonomy (Section A.6). The defacto situation, however, i s that central control has been weak and non-state power bases at regional and local levels strong. The structures o f a unitary state, provided for in Afghanistan's Constitution, are inplace inrudimentary form, but there are glaring capacity weaknesses. With formation o f the elected Parliament and Provincial Councils (and later District Councils), Afghanistan's administrative and political structure will be as shown inFigure8.9. 8.44 The need to improve service delivery and the challenge o f buildingan effective, accountable state are closely linked; both are at the core o f Afghanistan's reconstruction agenda. An effective state delivers services to the people, and by ensuring that services are delivered with widespread access by citizens (including women and other vulnerable or marginal groups), the state can build its legitimacy. Subnational levels o f administration (provinces and districts) are very important from both perspectives. A service delivery perspective provides some discipline inconsidering subnational strategy and focuses it on relevant outcomes. For example, there have been improvements in financial management and other aspects o f subnational administration (Box 8.3), but these may be only beginning to translate into improved service delivery. 74 District Municipalities Box 8.3: Improvementsand Challengesin SubnationalAdministration A recent update report on subnational administration, based on field visits to six provinces and interviews in Kabul, found that financial management at the provincial level has improved significantly over the past two years or so: 1 Inmost cases civil servants at subnational levelreceivepaymonthly, and on time. .. 1 Access to cash balances for non-salary expenditures is much improved, particularly invery poor provinces. 1 Financial reports are being submitted on time for the most part. The single expenditure and revenue accounts are working well. Audits are beingconducted, although their impacts arenot yet clear. However, there are still significant issues to be addressed: 1 Low pay continues to be a major complaint. 1 T h e takhsis (budget authorization) and tashkeel (administrative authorization for staff positions) arrive late in many cases, which can resultin delays in pay and other expenditures; there can also be disconnects between the line ministries' takhsis and that reaching the Mustoufiats (provincial finance departments). 1 The system o f transferring positive balances into the provinces' single expenditure account, while effective . ingetting cashout to provinces, is still placingunnecessary constraints or delays on spending. Non-salary allotments are still low, and there continue to be inequities between provinces. There i s still a lack o f written manuals, and more training i s needed. 1 Management o f the education budget i s still a major issue. Moreover, little has changed in terms o f recruitment and staff appointments. Although senior appointments are being channeled through the Independent Administrative Reform and Civil Service Commission (IARCSC), political and ethnic considerations appear to feature prominently inrecruitment. Many (though not all) o f these issues can be addressed relatively easily. However, there i s confusion among a number o f parallel and separate provincial coordination mechanisms supported by different external agencies (including for counter-narcotics), which detracts from the functionality and effectiveness o f subnational administration (Lister, 2005). Source: Adapted from Evans and Osmani (200.5, pp. 3-4); also see VolumeIII,Chapter 3. 8.45 Decentralization i s an important political choice made by a country, which may have a variety o f motivations. Although better service delivery i s often one o f the avowed objectives o f decentralization, international experience has been mixed. Moreover, there are examples where rushed and not carefully thought through decentralization initiatives turned out to be counterproductive. Among the risks o f decentralization which may be particularly salient inAfghanistan are the following: Possible "capture " of the decentralizationprocess and structures by local elites -this risk is particularly great in Afghanistan given the importance o f non-state forces (often retaining some armed capabilities despite the DDR process). Capture o f national revenues by Governors and other local power-holders has sometimes occurred and i s problematic for fiscal stability. 75 Associated risk o f inadvertently undermining the effort to build up a credible unitary state. Weakening of financial controls and fiscal discipline. Afghanistan's existing mechanisms and continuation o f the Government's strong "no-overdraft'' policy will likely enable strong fiscal discipline to be maintained. However, decentralization mechanisms such as block grants to lower levels o f Government administration, without proper financial controls and monitoring, could increase fiduciary risks. The experience in 2002/03 (until the creation o f the Treasury Single Account) shows the risk that revenues will be spent by provinces and poorly accounted for and that, even inresource-richprovinces, liabilities will be created. Issues related to ensuring reasonable equity in funding (and ultimately service delivery) across the country. 0 Design risks that could result in counterproductive outcomes, for example inadequate accountability at lower levels o f Government administration. 8.46 Nevertheless, Afghanistan's geographical characteristics and regional and ethnic diversity would suggest that over the longer run, decentralized modes o f governance may well be appropriate and could be considered. From a service delivery perspective, as seen earlier, there i s very often a case (depending to some extent on sector-specific circumstances) for management and oversight o f service delivery to be close to the providers and consumers. However, the current situation and the legacy from conflict provide ample grounds for exercising caution in this difficult area in the short run, and in particular not underminingthe state-building agenda. 8.47 Keeping within the present unitary state structure and takmg a practical, results-oriented service delivery focus, there are some promising options which include: Enhancing the functioning of the present subnational administration, including through measures to address the problems noted in Box 8.3. One concrete example would be increasing the share o f non-wage recurrent budgets going to lower levels o f administration. Making available information on budgetary spending byprovinces (and to the extent possible districts). Providing greater management flexibility to provincial and district line departments in oversight o f service delivery facilities (e.g. for provincial and district education departments in overseeing schools). This can include an extended role for local user groups (e.g. the Parent-Teacher Association for primary schools). Giving provincial line departments (and through them districts) an explicit role in the national budgetformulation process, which could include proposals, consultations, etc. 8.48 Finally, "top-down'' measures to strengthen subnational administration and enhance its capacity to oversee service delivery can complement "bottom-up" community-based efforts such as the NSP. There i s a risk o f the two approaches clashing, particularly ina highly politicized atmosphere, but this can be mitigated by maintaining a service delivery focus. As demonstrated by the initial successes achieved by NSP, community-based organizations can play important developmental and governance-improving roles, building on the traditional role o f communities in decision-malung and on participatory mechanisms (including elections), and promote social inclusion (securing representation not only o f local minorities but most remarkably also o f women). ! 76 CHAPTER 9. INSTITUTIONALREFORMSAND CAPACITY DEVELOPMENT 9.1 While the previous chapter reviewed options for organizing effective delivery o f services, this chapter focuses on buildingindividual and organizational capacity to improve PFM performance. At the outset, the "sequencing" or "time horizon" dilemma must be emphasized - there i s a dichotomy between the need to deliver services quickly to the people (Chapter 8) and the need to develop effective, sustainable national public institutions. The trade-off is complex, as both are necessary to build a legitimate, well-functioning State. H o w to quickly deliver services while progressively developing capacity over time? The chapter opens with a discussion o f the broader long-term agenda o f capacity development. It then reviews the reform process o f MoF, the core o f the PFM system. The third section analyzes the challenges o f building PFM capacity beyond M o F - in line ministries, provinces, and municipalities. A. Developingthe Government's Capacity 9.2 A quarter-century of wars and civil strife left Afghanistanwith a collapsed administration. While there was some institutional memory o f rules and processes (notably with regard to financial management, see Evans et al., 2004b, and Box 8.3), service delivery by the Government had largely ceased. Skills have been depleted, due to both the weak education system and a major "brain drain" (Box 9.1). The Government i s committed to building an effective administration, with the Government not delivering all services but often acting as a financier andfor regulator. T h i s requires significant enhancement o f Government capacity. Box 9.1: Harnessing Afghan Expertise Severe capacity constraints in Afghanistan's civil service result from a number o f factors. Millions o f Afghans, including many with skills, were killed or fled the country during the conflict. Universities and schools were closed or operating at low capacity. This has left the country deprived o f expertise, knowledge, and experience in managing the development process. Women in particular suffered from lack o f education opportunities. Many Afghans who left gained a better education and entered into skilled and professional employment in the region (especially in Pakistan and Iran) and also farther away (in countries such as the USA, UK,France, and Germany). In the early reconstruction phase, the Afghan civil service suffered further losses from among the (very few) skilled staff who had remained or returned immediately after the end o f conflict, as NGOs and donor agencies offered much better salaries and working conditions than were available in the Government. It would be misleading, however, to conclude that Afghanistan has no usable expertise; while external assistance i s very much needed, tapping existing Afghan expertise is crucial. The first option has been to make use o f the expertise and experience o f Afghans working inNGOs and also some o f the international agencies (these groups are generally referred to as the "second civil service"). Given the importance (and, in some cases, size) o f NGOs delivering services in the 1990s (Chapter 8), they have accumulated a wealth o f management and technical expertise. In the early days o f the transitional Government, several ministries tapped these human resources to quickly build very able management teams. The second option i s to tap human resources from the Afghan diaspora. A number of talented Afghans returned from abroad, sometimes working as volunteers, and brought important skills and experience to the Government. In order to better tap both of these reservoirs of skills, the Government has created the Lateral Entry Program (LEP) and the Afghan Expatriate Program (AEP), both o f which are being managed by the IARCSC. Under the LEP, the Government will inject up to 1,500 professional and skilled Afghans into senior and middle-level positions o f the civil service, with enhanced pay that goes beyond PRR scales. The primary target o f this program i s the "second civil service" and Afghans inregional labor markets. Under the AEP, the Government will appoint up to 100 highly skilled Afghan expatriates to work in advisoq positions inpriority areas, o f which around 50 are already inplace. This program's success i s based on a number o f factors: (i)a dedicated Executive Committee; (ii)clear guidelines on recruitment and pay levels; and (iii) a thorough review o f both the capacity o f the line ministry to use an expatriate Afghan effectively and o f the expatriate to bring the required skills. 77 THREESTRATEGIC COMPONENTS 9.3 The Government has outlined its strategy as part o f the Public Administration Reform and Economic Management (PAREM) program, one o f the 12 NDFprograms. This strategy, currently under revision for the ANDs, focuses on three key aspects. 9.4 First, individuals (civil servants) must be well qualified and motivated. Recruitment based on merit has been found to be the most effective way to raise skills and motivation, even though it can be politically sensitive ina fragmented society with a history o fpatronage. A diverse workforce -notably in terms o f gender and ethnic group - i s necessary to builda legitimate State. Women are especially needed inthe civil service for delivery and management o fpublic services for the female part of the population. However, the workforce composition needs first to be adequately understood and monitored before plans can be designed to enhance diversity. A talented workforce also requires an adequate pay and grading structure, providing attractive wages and potential for promotion. Wages and promotion should also be linked to performance. A strong ethos - a sense o f service and accountability - i s very important. While partly linked to issues o f pay levels, this also requires leadership by management teams that model these values. Inaddition, capacity can be created from inside by training, not only formal training but also on- the-job training or coaching from senior staff or external advisors (Box 9.2). Finally, tapping slulled Afghan expatriates may be a way to quickly bringtalented Afghans into the civil service (Box 9.1). 9.5 Second, organizations (ministries and agencies) need to function well. In addition to recruiting the right people, this requires institutional reforms to (i) specify each organization's mission (in line with the ANDs); (ii)improve (simplify and standardize, with clear reporting lines) the organizational structure, and (iii)review and improve business processes. The solution o f creating new organizations to solve each problem should be resisted (for instance, Chapter 7 highlights issues and problems associated with anti-corruption agencies). Adequate management processes (including sharing information, making decisions, implementing decisions, and following up) need to be established, which often requires some disconnect between, and clear accountability of, the political level (worlung toward and making high- level policy decisions) and the administrative level (providing information and analysis and implementing policy decisions). Some core functions need to be created or modernized, such as HR, finance (see below), and communications units. Implementing these sweeping changes requires preparing a vision for the agency, translated into a medium-term strategy and annual business plans with performance indicators, and implementedwith regular monitoring. It also requires physical investments (buildings, IT, equipment). 9.6 Third, the institutional environment needs to be supportive of effective individuals and organizations. The PFMReview has highlighted a number o f policies which are important in this regard. Policies should be matched to capacity, because a commitment to unenforceable policies i s economically costly (risk o f red-tape and corruption - Chapter 7, poor service delivery - Chapter 8) and politically harmful (failure to keep promises). In addition, decision-malung should be supported by adequate coordination mechanisms throughout the Government. As indicated in Chapter 6, the Cabinet Secretariat will need to play an important role in coordinating policies and malung sure that Cabinet decisions are informed and implemented. The capacity o f the justice system to enforce laws i s also critical for the Government's capacity to deliver services. Finally, the role and capacity o f the Parliament will be necessary for the Government to operate under adequate oversight (Chapters 6 and 7). 9.7 Progress on these three dimensions has been uneven. At the institutional level, a number o f policies have been adopted related to PFM that seek to simplify the regulatory framework, and M o F envisages further simplifyingthe tax code. Projects are underway to reform the Cabinet Secretariat. At the organizational level, the Priority Reform and Restructuring (PRR) scheme (which allows departments to place staff on an elevated pay scale in exchange for organizational restructuring) i s being implemented for 15,000 positions. Finally, at the individual level, more than 100 senior appointments are n o w processedmonthly by the Independent Appointment Board, following merit-based procedures. More than 100 appeals on recruitment, transfer, and termination of employment have been evaluated. A short-term 78 capacity building framework (including programs for Afghan expatriates and lateral entrants from neighboring countries, Box 9.1, and funding for short-term technical assistance and feasibility studies) i s under implementation. However, there remain many issues, at the institutional (weak enforcement capacity in the justice system, for instance), organizational (lack o f in-depthorganizational reviews and reforms), and individual (uncoordinated training programs and slow recruitment) levels. Two KEYCHALLENGES 9.8 This uneven progress should be viewed inthe context o f two important challenges that capacity development faces. First, international experience has amply demonstrated that capacity development - recruiting and training people as well as reforming and strengthening institutions - i s a long-term endeavor. It involves technical assessments and preparation o f detailed strategies, politically sensitive changes in practices, and learning by doing. Faced with the "time horizon dilemma", an asymmetric approach - inwhich reforms are implementedin selected agencies first and gradually rolled-out may be - the only solution. But this i s a risky strategy, which can be threatened by frustration at lack o f more widespread results; social discontent arising from disparities among pay levels; slow progress because unreformed departments weaken the performance o f reformed ones; and ineffectiveness due to uncertainties surrounding institutional reforms. These risks can only be mitigated by high-level political leadership, a large and effective injection o f technical assistance (Box 9.2), an adequate monitoring mechanism, and communication efforts. Box 9.2: Using Technical Assistance Effectively According to OECD/DAC data, more than $400 million o f external assistance to Afghanistan was spent on Technical Assistance (TA) in 2003. Such a high level o f TA may well be necessary given the low capacity in the Government and the need for a short-term injection o f capacity (including for highly specialized skills). Nevertheless, various problems suggest that T A could be better managed. The following lessons can be learned from experience: .TA should be based on explicit demand and leadership from a clearly identified counterpart, and should report to that counterpart (with quality control as needed from the donor); this should increase follow-up on TA recommendations. TA proposals should include clear and monitorable indicators and should be monitored, with outcomes evaluated; this would enable assessment o f the quality o f T A and better management o f it. More specifically, there needs to be differentiation between (i)operating TA (directly carrying out certain Government functions - this should be on a strictly temporary basis untilnational capacity i s builtup); (ii) advisory TA (a small number o f advisors on a long-term basis, with provision for short-term advisory support on specific issues and for .identified analytical work); and (iii)capacity-building TA (intended to directly support sustainable capacity building). Ministries could also create dedicated processes to ensure adequate TA oversight and coordination; while a clear ministry strategic framework with specific milestones should help, more structural coordination may be required. 1To the extent possible, T A provision should be aligned with organizational structures to avoid unnecessary fragmentation, which would avoid in particular conflicting advice (some departments have up to six or more TA providers). This may, in some cases, lead to assigning one TA provider per department or agency, or pooling donor resources together (as is done, for instance, inTanzania and Cambodia for PFM reforms). 1TA contracts should be subject to international procurement standards to seek maximum value for money. 9.9 The experience with the PRR highlights these difficulties. Technically, implementing this scheme requires significant assistance, as organizing ministries and gradingjobs require specialized skills. Implementation i s also contingent on having experienced management capacity. Politically as well, the asymmetric approach requires solid high-level leadership and a good communication campaign to maintain support for - or at least acceptance o f - this strategy, especially when its impact i s difficult to quantify and takes time to materialize. As indicated above, an asymmetric solution may well be the only option, but the risks have to be mitigated. 9.10 Second, donors, while bringing financial and technical assistance, can distort incentives and the labor market (impacts on the budget are discussed in Chapter 6). Given the shortage o f skills (notably 79 management, language, and computer shlls), competition between local offices o f the international community and the rest o f the market has led to higher salaries, especially inKabul, reaching levels that the Government, in need o f similar skills, cannot afford. In some cases, donors even directly pay civil servants' salaries or "top-ups", which i s not fiscally sustainable for the Government and can disrupt accountability and loyalty relationships. Such relationships are also harmed when separate units dedicated only to supporting donor projects are created (Section C). In some cases donors approve salary/wage levels that are well above prevailing rates for whole categories o f employees - such as for the ANA, National Security Directorate, and Counter Narcotics Police - but which are never costed or assessedfrom a fiscal standpoint. Finally, external technical assistance, while very much needed to "buy capacity" inthe short term, needs to be well managed to be effective (Box 9.2): otherwise it can result ina waste o f resources that generates resentment among less well-paid staff without having much positive impact. B. Reformingthe Ministryof Finance 9.11 As emphasized throughout this report, MoF has a crucial stewardship role in public finance management. Its capacity to fulfill its role in P F M i s therefore critical for overall PFM performance as captured by the PFM indicators (Chapter 2). Progress in mobilizing revenues also requires the development o f a professional cadre inthe tax and customs administration(Chapter 4). Progresshas been substantial - under the leadership o f a strong management team and with external assistance supporting, sometimes operating, core systems such as AFMIS. But the challenges remain enormous. 9.12 Inthe first halfof2005, MoFtherefore initiateda strategic planningprocess. The first step was to put forward a vision for the Ministry, based on five goals: (i) mobilizing revenue and managing Government finances; (ii)supporting economic management and promoting economic growth; (iii) managing public wealth; (iv) promoting good governance; and (v) being a best-practice leader within Afghanistan's public sector. 9.13 A review o f the Ministry's capacity, from an individual and organizational point o f view (policy issues are discussed in Chapter 6 and 7), suggests the following issues: Organizational adjustments: The newly-created Fiscal Policy Unit in MoF should help strengthen analytical capacity; the Treasury Department and the Accounting Department are being merged; a Procurement Policy Unit should be created as a clear champion o f procurement inthe Government (see Chapter 7); the creation o f the Large Taxpayer Office i s intended to help mobilize more revenues; and the planned creation o f Revenue Offices in the provinces, with the Mustoufiats becoming sub-Treasuries, would clarify accountabilities at the provincial level. Departmental reforms: With these organizational adjustments, the Ministry could submit a Ministry-wide PRR (as a further step to existing PRR programs in several departments); this should cover the strengthening o f the Budget Department (along sectoral lines, with no separation o f operating and development budgets), the Internal Audit Department (in line withthe PFEMLaw), andthe Finance and Administration Department. Institutional strengthening: A number o f processes should be created or strengthened to improve the Ministry's effectiveness, including the budget process for MoF's own budget; human resource strategy; internal and external communications strategy; information sharing and decision-making processes; and TA management (Box 9.2). 9.14 M o F may want to pursue a rigorous institutional reform process inorder to develop its capacity as steward and custodian o f Afghanistan's public finances. The next step would consist o f translating MoF's vision into detailed and quantified performance targets to measure progress; specific and realistic annual business plans; and consistent strategic plans, performance targets, and annual business plans at the department level. Implementation will require sustained efforts and regular monitoring. A possible 80 vehicle for this purpose would be an institutional reform working group / monitoring cell within MoF, with support from MoF's top leadership. C.BuildingPFMCapacityThroughout Government 9.15 Beyond MoF, strengthening overall PFM performance calls for financial management reforms and capacity development inline ministries (including those responsible for the security sector as well as other sectors). Most o f them have very weak capacity and operate based on outdated practices. Most ministries continue to have two departments, sometimes reporting to two different deputy ministers, separately working on their operating and development budgets. Many processes - including the expenditure process - involve dozens o f signatures. This overabundance o f ex-ante controls not only adds nothing to the effectiveness o f internal controls but also leads to slow processing and creates opportunities for corruption. 9.16 Financial management unitsor budget units inline ministries should therefore be reformed. M o F intends to pilot a "model office" in one ministry, before rolling out reforms more widely in line ministries. This would need to be coordinated with the procurement reform, according to which procurement authority i s to be decentralized in line ministries. Over time, such reforms could allow a number o f responsibilities to be devolved to line ministries, notably access to AFMIS for reporting and, later, posting o f expenditures (Chapter 7). This process would need to be subject to well-defined criteria, building on the experience o f Thailand, where under the so-called "hurdle approach", PFM functions were devolved to line ministriesafter they met performance standards. 9.17 The creation o f Program Management Units or Program Implementation Units (PMUs/PIUs) can support better P F M performance but also gives rise to risks. Donors often favor having such units to manage their funds. These units have proven useful, even indispensable, in cases where the civil service and the existing institutions have no capacity. However, autonomous PIUs can become "super- ministries", which require a long time and high spending to establish, absorb civil servants from ministries and generate resentment from others, expand their purview, and become over-stretched and entrenched. As a result, PMUsPIUs easily can become an obstacle to national institution building and sustainable capacity development. It i s therefore extremely important to ensure that enclave/isolated PMUs/PIUs do not become the norm, and that those already established are reformed. Inparticular, these units should not cover only donor programs but rather should be integrated in line ministries (OED, 1998). The Grant and Contract Management Unit inthe Ministry o fPublic Healthhighlights these issues: well integrated in the Ministry,its strong capacity not only facilitated implementation o f donor-financed projects but also contributed to the overall achievements o f the Ministry. Nevertheless, tensions and resentments (including among donors) were not easy to manage. 9.18 At the subnational level, the PFM system works through the Mustoufiat network. Analysis has shown that administrative and financial regulations are adhered to in provinces (Box 8.3). However, as the Government seeks to implement the budget more effectively outside Kabul, Mustoufiat reform i s essential. At the individual level, this requires the implementation o f pay reform and training. At the organizational level, it i s recommended to separate Treasury functions from revenue functions, with a view both to specialize these units (with different slulls and different clients) and ensure adequate separation o f duties. At the institutional level, strengthening Mustoufiats requires a commitment from the central ministries (MoF in particular) to deliver funding in full and on time; to take into account local realities and demands in the formulation o f the budget; and to provide clear manuals, training, and other forms o f support (e.g. implementation o f AFMIS) to implement modernized PFM rules throughout the country. Finally, devolving responsibilities to line ministries (see above) would require developing PFM capacity not only inthe central ministries but also intheir provincial line departments. 9.19 While municipalities have an institutional set-up that has potential to effectively deliver urban services (Chapter 8), their P F M capacity i s currently very weak. A study o f Kabul Municipality highlights the constraints: unclear institutional responsibilities between the municipality and line 81 ministries (notably the Ministry o f Urban Development and Housing); outdated functions including some (e.g. price control) inherited from the era o f the Soviet occupation; an organizational structure that has been adjusted but i s still inneed o f further improvement (see Volume 111, Chapter 5); poor budgeting (e.g. separation o f operating and development budget functions), financial management, procurement, and record-keeping processes (for instance, if there i s no line item in the budget for a certain type o f expenditure, the expense i s simply not entered inthe ledger); and opaque revenue collection mechanisms applied to numerous low-yield local revenue sources. Therefore, P F M processes and capacity need strengthening as a prerequisite for realizing municipalities' potential to deliver services inurban areas. 9.20 Insummary, the capacity development agenda is far-reaching, as it covers the three dimensions outlined in Section A and cuts across the central Government, its provincial and district departments, and municipalities (to which SOEs and the private sector - for participation inpublic procurement- shouldbe added, see Chapters 3 and 8). Further developing MoF's capacity -both inKabul and inthe Mustoufiats -is the short-term priority as MoF is the core ofthe PFMsystemandthe custodianofpublic funds. Buildingon this, organizational reforms (budget, financial management, procurement departments) and training can be rolled-out to line ministries and municipalities while, ina sequenced way, authorities (e.g. for procurement) are de-concentrated to line ministries based on performance. 82 CHAPTER 10. KEYCHALLENGESAND A ROADMAPFOR PFMREFORM 10.1 The previous chapters o fthis report have highlightedboth the progress achieved and the problems faced by Afghanistan's PFM system. Building on the analysis presented earlier and looking to the future, this concluding chapter briefly summarizes the major challenges that need to be addressed and puts forward a roadmap for strengthening and reform o fthe PFMsystem. A. EightMain Challenges 10.2 Good public finance management i s a key enabling factor for implementation o f Afghanistan's development strategy and for achieving central national objectives o f state-building, sustained rapid economic growth, and poverty reduction. The following broad challenges for Afghanistan's PFM system have been identified and discussed inthis report. 10.3 First, and from a macroeconomic and growth perspective o f critical importance, there i s the challenge of fiscal sustainability. Afghanistan needs to make continuous and sustained progress toward a fiscal position where domestic revenues cover all operating expenditures. This will require, inter alia, (i) growthofdomesticrevenues(seebelow); (ii) rapid containment o f overall expenditures in line with medium-term resource constraints; (iii) ensuring that downstream expenditure liabilities created by public investments and other spending decisions are affordable; and (iv) incorporating recurrent expenditures funded directly by donors through the ExternalBudget inthe fiscal sustainability equation(notably inthe security sector), and bringingthese expenditures progressively into national budget channels. 10.4 Second, there i s the challenge of sustained rapidgrowth of domestic revenues, which will be the lynchpin o f progress toward fiscal sustainability). Afghanistan's revenue-to-GDP ratio at around 5% i s well below half the level even in other very poor countries. Revenues have to grow much more rapidly than expenditures on a sustained basis. At the same time, a sound revenue system should be conducive to development o f the private sector. Key challenges inraising revenues include: (i) lack o f capacity inthe tax administration system as well as among taxpayers to calculate and pay taxes; (ii) existence o f numerous, low-yielding "nuisance taxes", illicit revenue collection by many local authorities, instances o f double-taxation (mainly between municipal and national taxation), and ambiguities in the tax laws; (iii) the dominance o f agriculture and the informal sector in the economy, which are difficult to tax in any country; and (iv) widely perceived corruption inthe tax administration. 10.5 Third, there is the challenge of apolicy-based, well-prioritized budget with strongpolitical buy- in. This will require a close linkage with national and sector strategies that themselves will need to be oriented toward private sector-led growth and have widespread national ownership. A Medium-Term Fiscal Framework (MTFF), integrally linked to annual budgeting, also will be critical. Effective strategic prioritization o fbudget allocations will be difficult (including politically) but essential. Key cross-cutting issues such as gender will need to be mainstreamed inbudgetary decision-makmg. 10.6 Achieving this gives rise to a fourth challenge: the challenge of making the national budget the central instrument for policy and reform. Key constraints include: (i) capacity; (ii) dominant limited the share o f expenditures directly by donors outside Government systems; (iii) serious timingdilemmas and a compressed schedule which weakens the quality o f the budget; and (iv) lack o f a medium-term strategic and fiscal perspective to guide annual budget formulation. Meaningful processes o f political engagement and approval (including by Parliament) and public communications will also be needed. Finally, there i s an urgent need for greater inputs and participation by lower levels o f the government administration (provinces and districts) inthe budget formulation process. 10.7 Fifth,there is the challenge of effective budget execution. While considerable progress has been made in some areas, a great deal remains to be done. Relying on sound institutions and adequate 83 capacity, strengthening financial controls, procurement, recordingheporting, and audit will be key elements o f better budget execution. Meaningful implementation o f the Public Finance and Expenditure Management (PFEM) Law and the Procurement Law will support effective budget execution. Strengthening accountability and reducing vulnerability to corruption will be critical. International experience demonstrates that the latter i s best achieved by addressing the underlying governance problems (in the PFM system and elsewhere), rather than through investigation and prosecution o f corruption cases. Thus improvements in the P F M system, particularly in the area o f budget execution, will needto be a key element of a holistic anti-corruption strategy. 10.8 Sixth, there i s the challenge of service delivery, the ultimate outcome o f a PFM system. With a few exceptions, public service delivery in most sectors i s poor, as demonstrated by available data on outcome indicators, giving rise to frustrations and weakening the credibility o f the Government. Accountability relationships and incentives for service providers need to be gotten right, although institutional arrangements can and will vary by sector. Sustainable financing and, where appropriate, adequate levels o f cost recovery, will be essential. Donors will need to ensure that their interventions strengthen rather than undermine national service delivery mechanisms. 10.9 Seventh, closely related, there i s the challenge of sustainable national capacity development. Unless this happens, Afghanistan will remain excessively dependent on foreign expertise indefinitely, and highly vulnerable to reductions in aid flows and external TA. Capacity development will require public administration reforms (recruitment, pay, HR), organizational restructuring and improvements, and training. Better management o f the massive amounts o f external TA currently being injected into Afghanistan will be neededto ensure that it supports longer-termnational capacity development. 10.10 Finally, there i s the challenge of coordination and communication within an overall national strategic and budgetary framework. Coordination within the Government needs to be improved (this can be done around a widely-owned national development strategy and a sound budget process), and public communications and transparency will be extremely important in fostering Government accountability vis-a-vis citizens and their elected representatives. Coordination i s a major challenge for donors as well, which also needs to occur around the national development strategy and budget process -the l o w share of external assistance going through national budget and treasury channels i s a serious constraint in this regard. 10.11 While the challenges outlined above are numerous and daunting, Afghanistan can build on earlier progress, and prospects for success are better than it might superficially appear at first sight. In fact, the major progress already achieved - for example in Treasury functions, in currency reform and monetary policy, in expansion o f education and delivery o f basic health services - gives ground for hope. The Ministryof Finance has developed a vision to guide it inidentifying and implementing a wide range o f actions required for strengthening its current functions and capacities and moving forward from the emergency needs of the last two years toward a modern and effective public financial management ~rganization.~~Other supportive factors include positive developments in the financial reporting system, solid institutional reforms in a few line ministries, the establishment o f basic processes to manage the budget, and the positive collaborationwith the international community, notably through the ARTF. 10.12 However, there are also many downside risks and threats, ranging from insecurity to difficulties in mobilizing domestic revenue, lack of core capacity in Government, and fragmentation o f much international assistance in the External Budget. Much will depend also on establishing the functionality and processes o f Parliament as an essential element promoting accountability. Finally, it must be recognized that improvements will not be easy and will take time, not least because a stable political culture will take time to develop. Given the "time horizon dilemma" that was mentionedearlier related to achieving quick results versus longer-term national capacity building, interim and asymmetric mechanisms have been used and will continue to be required, but they need to be applied in a way that does not detract from sustainable progress over the mediumterm. 23"AfghanistanMinistry of Finance-The Way Forward", backgroundpaper for the World BankPFMReview. 84 B.A Five-Point ReformAgenda 10.13 The Government has fully recognized these challenges and the centrality of the budget process and PFM performance for Afghanistan's future. It is preparing the interim Afghanistan National Development Strategy (ANDs), in which the PFM agenda needs to be anchored. This report concludes with a suggested five-point agenda to concretely move forward. 10.14 First, the path towardjiscal sustainability needs to be mapped out and steady progress in this direction achieved. Based on a sound MTFF (which can be rudimentary at the outset but improved over time), this will require particular attention to increasing domestic revenue mobilization (through administrative improvements leading to better compliance as well as sound tax policies); maintaining control over aggregate expenditures within the overall resource envelope; setting and implementing clear policies with respect to key expenditure components notably the Government wage bill; and ensuring that strategic and public expenditure decisions are affordable over the medium-term, including downstream O&M implications ofpublic investments and service delivery networks. 10.15 Second, a number of policy decisions and actions are needed to improve service delivery. This will require clarifying the roles o f the State (operational, financing, policy, and regulation) in service delivery in different sectors; strengthening accountability; ensuring sustainable financing arrangements including cost recovery where appropriate; and determining the responsibilities o f various levels o f Government administration (central ministries, provinces, districts). As discussed inChapter 8, a variety o f different models of service delivery are possible within Afghanistan's Constitutionally-mandated unitary state structure, and the choice among them should depend on sector-specific conditions. Although "one size does not fit all," the effectiveness o f service delivery i s largely dependent on the degree o f ownership behind the approaches and the coherence o f the models in terms o f accountability, incentives, and financing. (including multi-year fiscal implications). 10.16 Third, the capacity of the key PFM institutions should be developed. Without stronger capacity and institutional reforms, improvements in PFM performance as well as progress toward key national development objectives will be difficult to achieve and impossible to sustain. Priority areas include training, recruitment practices, pay structures, organizational reforms, detailed business processes, specific IT systems, etc. While these issues cut across all levels o f Government, MoF's capacity i s o f particular importance. Since it i s chief custodian o f the P F M system, its capacity to develop and implement policies i s at the core of PFMperformance. Progress toward an M o F Strategic Plan - initiated in May 2005 - will be very important in developing MoF's capacity and performance. PFM capacity development inline ministries also i s very important but will take more time. 10.17 Fourth, and related to the previous recommendations, the Government should adopt a clear action plan to improve PFM performance, with time-bound milestones, quantified objectives, and clearly identijied responsible agencies. Table 10.1 seeks to provide a comprehensive, high-level roadmap o f policy reforms required. It i s comprehensive in the sense that all areas o f PFM are covered. It is high-level in the sense that some concrete next steps are identified for the next 6-12 months, but many actions are less specific and will need to be converted into detailed action plans later by the responsible Government departments. Also, some key accountabilities, left out o f the PFEM Law, will need to be fully specified, especially with regard to the respective roles o f M o F and the line ministries (the Procurement Law clearly gives authority to line ministries). Progress along this roadmap could be measured by key fiscal outcomes and implementation and progressive improvement o f the MTFF, and by the PFM performance indicators presented in this report. Monitoring o f progress, along with a meaningful feedback loop into subsequent decision-making, will be as important as the action plan itself. 10.18 Finally, continuing and further deepening collaboration between the Government and the international community will be critical for the success of the reform program. The Government's development strategy - which may include plans for progressing toward fiscal sustainability, more effective service delivery, capacity development, and strengthened PFM performance - should serve as the basis for a framework of mutual accountability. Donors would align their support to national strategic 85 objectives and the nationalbudget and increasingly use the Government's systems, while the Government would implement the decisions it has announced and further improve the performance o f the PFMsystem. In addition, the content, mode, and capacity-building contribution of technical assistance will be very important and needs to be well-managed and well-coordinated in line with Government leadership and national budgetary and other capacity-building processes. 10.19 Wayforward. Not surprisingly inview o f the major challenges that Afghanistan faces, this i s a multi-faceted, demanding, and ambitious medium-term agenda. Each o f the five elements includes difficult sub-agendas and important actions required for success. This points to the need to prioritize and focus on a set o f realistic yet meaningful short-term measures that will achieve significant improvements and step-changes, thereby setting the stage for further progress. A package o f prioritized short-run actions could thus comprise a "platform" from which the next set o f actions would take off, with monitoring and feedback to guide the process at each stage.24 Moreover the package, rather than any single measure or mechanical target, would be what the Government takes responsibility for and could form the basis for dialogue and agreements withthe international community. 10.20 Inthis process, the Government first needs to develop and reach agreement on its strategy. The Afghanistan National Development Strategy (ANDs) i s expected to be completed on an interim basis shortly. It needs to include or to be complemented by the MTFF, as well as a P F M action plan that would be adopted by the Government and an M o F reform, strategy. Clear sector strategies also are needed to guide prioritization and sector expenditure programs. The ANDs and sector strategies needto specify the roles o f the State and put forward sound approaches to ensure effective service delivery, differentiated by sector as appropriate (see Chapter 8). The sector strategies will take varying amounts o f time to develop and, like the interim ANDS, MTFF, and PFM action plan, will need to be improved over time and adjusted in the light o f experience. It i s critically important that these strategies have wide national ownership, both within and outside Government. 10.21 The strategy as it applies to public finance management would then be translated each year into annual plans that would consist o f three building blocks: (i) an updated medium-term view o f the fiscal path (MTFF); (ii)an annual budget that i s consistent with the MTFF and with the ANDs; and (iii) an "annual action plan" (a list of a small number - certainly less than 20 - key actions to be accomplished during the year). Successful implementation of this package (not necessarily all o f the measures but at least a critical mass of most o f them) would build the "platform" on which the following year's action plan could be based, leading to further progress subsequently. This type o f process would help in coordinating the expectations of all partners as to what the Government realistically will be able to achieve every year and, with adequate monitoring, it would help the Government to establish a strong track record o f implementation. Adequate monitoring and evaluation should also feed back into the next year's annual plan and possible revisions to the overall strategy - this requires significant development o f statistical and analytical capacity. 10.22 In conclusion, the five-point reform agenda outlined above i s ambitious, requiring sustained efforts over a considerable period of time based on meaningful steps each year. But as seen earlier, the challenges in building an effective, accountable, financially self-sufficient state that facilitates sustained economic growth, ensures adequate delivery o f services to the Afghan people, and reduces poverty are enormous and pressing-requiringa commensurate response. The Government, with strong support from the international community, has demonstrated its willingness and commitment to embrace an agenda along these lines and move forward with implementation ina determined yet realistic manner. ~~ 24See Brooke (2003) for a discussion of the background, rationale, and modalities ofhow such a "platform" approach could work inthe context ofcountry situations involving external budget support. 86 s 3-10 - I i 2-2 m-f . . -4 + 1 5 N I W I - 0 N m h n N iA N W 3- N N m 4 0 T- I B .r .r .- cn E .r F Esi .r .r 2E r c r STATISTICALAPPENDIX 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Estimates Projections A. Output GDP (US$million, excluding opiumproduction) 2,463 4,084 4,585 5,975 7,130 8,562 9,951 GDP (millions ofnew Afghani) 133,987 182,862 225,108 284,504 355,338 426,708 495,946 Real GDP growthrate ... 28.6 15.7 8.0 14 11 11 GDP per capita (in US dollars) 123 182 199 253 293 342 386 Ooiumoroduction (metric tons) 185 3.400 3.600 4.200 4.100 ... ... . L Opiumvalue (millions of US dollars) 56 1,300 2,300 2,200 ... ... ... B. Prices CPI (% year-on-year change) (43.4) 52.1 10.2 16.3 10 8 5 C. Exchange rates AfghaniiUS dollars (averageduring solar year) 54.4 44.8 49.0 47.7 ... ... ... AfghaniilTS dollars (end-solar year) 31.0 52.6 50.3 48.7 ... ... ... D. Monetary indicators Domestic currency in circulation (percentage change) ..... . 20.1 40.9 37.5 28 23 ... Gross foreign exchange reserves (US$million) 426 816 1,261 1,691 2,036 ... Children 1-4 years that are disabled 2 3 3 Children7-17 years that are disabled 3 4 3 Childrenwith bothparentsdead 6 6 6 Preventive Health (YO) Children 12-23 monthsthat have received DPT 3 immunization 23 48 30 Children 12-23 monthsthat havereceived Polio 3+ immunization 43 71 51 Children9-59 months that havereceivedMeaslesimmunization 73 84 76 Childrenunder5 that havereceivedBCG immunization 55 73 60 Access to Sanitation and Water (YO) Householdswith drinking water from pump/protectedspring 31 61 40 Householdshavingaflush or pit toilet 59 87 67 Householdswith water source and latrine within 15 meters 77 60 67 Education Illiteracyrate: Male (%ofpop. aged 15 and over) 64 40 57 Illiteracyrate: Female (% of pop.aged 15 and over) 92 72 86 Primaryschoolnet enrolmentrate (%of 7 to 13 year olds) Male 61 81 67 Primaryschoolnet enrolmentrate (%of 7 to 13 year olds) Female 30 63 40 93 TABLE 3: FISCALSUMMARY (AFSMILLION) A. 2002/03 2002103 2003/04 2003104 2004/05 2004/05 2005106 Budget Est. Budget Est. Budget Est. Budget (In millions ofAfghanis) DomesticRevenue 2,822 5,864 9,000 10,168 15,380 12,812 16,150 Tax Revenues 6,262 9,546 Taxeson income,profits and capitalgains 363 995 Taxeson internationaltrade andtransactions 5,369 7,247 Othertaxes 531 1,304 NonTax Revenues 3,906 3,266 Donor AssistanceGrants (to operating budget) 12,828 9,430 15,750 10,074 14,952 14,984 16,733 ARTF 11,250 8,182 12,575 12,319 13,580 LOTFA andArmy TrustFund 4,500 1,892 2,377 2,583 3,153 Other grants 82 Donor AssistanceGrants (to core developmentbudget) 4,569 36,882 8,250 34,648 Total Expenditures(Core Budget) 15,650 15,514 24,750 30,054 81,562 41,710 91,513 Operating Expenditure 15,650 15,514 24,750 21,972 30,332 26,605 32,883 Wages andSalaries 12,954 13,818 18,027 17,842 22,311 Purchaseofgoods andservices 6,493 4,243 4,850 5,247 4,037 Debtservice andInterestpayments 25 278 97 404 Other recurrent 1,002 881 (12,438) 1,440 (345) Capitalexpenditure 1,683 3,005 1,752 1,979 2,301 ReservesandContingencies 2,618 17,863 4,175 Core budget developmentspending 8,082 51,230 15,105 58,630 By Program 24,750 30,054 81,562 41,710 91,513 HumanCapitalDevelopment 9,181 6,388 22,458 16,657 27,881 1.1 RefugeesandIDPreturn 132 81 127 105 187 1.2 Education 6,002 3,938 7,513 5,824 9,179 1.3 Health 1,256 1,233 2,327 3,007 2,850 1.4 Livelihoodand SocialProtection 1,192 627 11,870 7,079 14,537 1.5 Culture,Media, Sport 600 509 621 642 1,128 PhysicalInfrastructure 1,159 6,941 18,160 7,253 28,312 2.1 Transport 188 1,381 9,100 3,975 12,537 2.2 Energy,Mines, andTelecom 363 3,224 5,840 2,163 9,577 2.3 NaturalResourcesManagement 575 2,017 2,091 725 3,594 2.4 UrbanManagement 33 319 1,129 390 2,604 General Administration 2,358 5,203 6,155 3,876 12,521 3.1 TradeandInvestment 99 23 1 1,298 421 1,489 3.2 PublicAdministrationandEconomicManagement 2,259 4,972 4,857 3,455 11,032 Security 9,434 11,523 16,925 13,924 18,623 3.3 Justice 424 599 921 1,002 1,122 3.4 NationalPoliceandLaw Enforcement 3,250 5,515 10,689 7,792 11,288 3.5 AfghanNationalArmy 5,760 5,409 5,315 5,130 6,100 Reservesand Contingencies 2,618 17,863 4,175 Operatingbudgetbalance(excluding grants) (12,828) (9,650) (15,750) (11,804) (14,952) (13,794) (16,733) Operatingbudget balance(including grants) (220) (1,730) 0 1,190 0 Core budgetbalance (including grants) (220) (0) (5,243) (14,348) (5,665) (23,982) Float and Adjustment (656) 0 140 14,348 2,507 97 Financing 876 5,103 3,158 23,885 ExternalLoans(net) 4,886 14,754 23,853 Domestic(net) 876 217 - (11,596) 32 Memorandumitem: ExternalBudget 22,777 95,733 106,879 164,794 119,398 154,235 Source: MoF,IMF. 94 i o - ? 0 0 ' a : - TABLEA. 6: FISCALSCENARIOS 6a. Base Case 2002103 1 2003104 1 2004105 -1 2005106 2006107 1 2007108 1 2008109 I 2009110 1 20lOlll 1 2011112 1 2012113 is millioi Domestic Revenues 5,864 10,168 12,812 18,392 25,789 32,821 41,363 77,254 CoreExpenditures 15,514 30,054 41,710 59,483 102,602 114,482 118,044 125,978 of which OperatingExpenditures 15,514 21,972 26,605 32,883 43,284 53,049 60,664 76,882 ofwhich wages d a 13,818 17,842 22,3 1I 27,280 31,464 36,313 45,986 ofwhich non wage da 8,154 8,764 10,572 16,004 21,585 24,351 30,896 of which DevelopmentExpenditures 8,082 15,105 26,600 59,3 18 61,433 57,380 49,096 FiscalDeficit Before Grants 9,650 19,886 28,899 41,091 76,813 81,661 76,682 48,724 of which operating 9,650 11,804 13,794 14,491 17,494 20,228 19,301 (372) ExternalDvpt Expenditures 22,777 106,879 119,398 79,021 130,050 124,896 110,515 44,620 Total Expenditures ----- 38,291 136,933 161,108 138,504 232,652 239,379 228,560 170,598 As percentage GDP DomesticRevenues I 3.2 1 4.5 1 4.5I 5.2 6.0 1 6.6 1 7.2 1 7.8 1 8.1 1 8.2 1 8.3 Core Expenditures 8.5 13.4 14.7 16.7 24.0 23.0 20.6 19.6 15.8 1 of which OperatingExpenditures 8.5 9.8 9.4 9.2 10.1 10.7 10.6 10.7 10.2 ofwhich wages d a 6.1 6.3 6.3 6.4 1 ::: 1 6.3 6.3 6.3 6.0 of which DevelopmentExpenditures 3.6 5.3 7.5 13.9 12.4 10.0 8.9 5.6 Fiscal Deficit Before Grants 5.3 8.8 10.2 11.6 18.0 16.4 13.4 11.9 7.7 5.2 of which operating I 5.3 5.2 4.8I 4.1 3.4 3.0 2.1 y:; (0.0) ~ ~ ExternalD v p Expenditures 12.5 47.5 42.0 22.2 3::; 19.3 16.4 10.2 5.4 4.8 54.5 1 48.2 1 39.9 36.0 26.1 20.7 1 18.4 6b. HighCase I EWZI03 12003104 1 2004105 2005106 12006107 1 2007108 1 2008109 1 2009110 12010111 12011112 1 2012113 projection 's millions Domestic Revenues 5,864 10,168 12,812 63,766 Core Expenditures 15,514 30,054 41,710 128,579 of which OperatingExpenditures 15,514 21,972 26,605 71,212 o f which wages d a 13,818 17,842 42,753 o f which non wage d a 8,154 8,764 28,458 of which DevelopmentExpenditures 8,082 15,105 57,368 FiscalDeficit Before Grants 9,650 19,886 28,899 64,813 ofwhich operating 9,650 11,804 13,794 7,445 ExternalDvpt Expenditures 22,777 106,879 119,398 105,956 Total Expenditures 38,291 !36,933 234,535 Domestic Revenues 3.2 4.5 4.5 5.4 6.5 7.5 8.9 9.9 10.8 11.5 11.6 Core Expenditures 8.5 13.4 14.7 16.7 24.0 23.4 20.9 19.9 16.0 15.3 13.6 of which OperatingExpenditures 8.5 9.8 9.4 9.2 10.1 11.0 10.9 11.0 10.4 9.3 8.3 o f which wages d a 6.1 6.3 6.3 6.4 6.7 6.7 6.6 6.2 5.5 5.0 of which DevelopmentExpenditures 3.6 5.3 7.5 13.9 12.4 10.0 8.9 5.6 6.0 5.3 FiscalDeficit 5.3 8.8 10.2 11.4 17.6 15.9 12.1 10.0 5.2 3.7 1.9 of which operating 5.3 5.2 4.8 3.9 3.7 3.5 2.1 i 1.2 (0.3 , , (2.2 . , (3.3) . , ExternalDvpt Expenditures 12.5 47.5 42.0 22.2 30.5 25.1 19.3 16.4 10.2 5.4 4.8 Total Expenditures 20.9 60.8 56.6 97 6c. Low Case I 2002103 1 2003104 12CUMlO51 2005106 I 2006107 1 2007108 1 2008109I2WllO 1 2010l11 1 2 O l l i l l 1 2012113 Domestic Revenues 5,864 10,168 Core Expenditures 15,514 30,054 of which OperatingExpenditures 15,514 21,972 of which wages d a 13,818 o f which non wage d a 8,154 of which DevelopmentExpenditures 8,082 FiscalDeficit Before Grants 9,650 19,886 ofwhich operating 9,650 11,804 ExternalDvpt Expenditures 22,777 106,879 1 -- 136,933 II A S P entace of GDP Domestic Revenues I 3.2 4.5 4.5 Core Expenditures 8.5 13.4 14.7 of which OperatingExpenditures 8.5 9.8 9.4 o f which wages d a 6.1 6.3 6.2 1 of which DevelopmentExpenditures 3.6 5.3 FiscalDeficit 5.3 8.8 10.2 5.9 o f which operating 5.3 5.2 4.8 0.6 ExternalDvpt Expenditures 12.5 -- 47.5 42.0 Total Expenditures 1 20.9 -- 60.8 56.6 Source: staflprojections. TABLE 7: EXPENDITURESPROGRAMFROM2003/04 TO 2005/06 A. BY 2003/04 2004/05 2005/06 2003/04 2004/05 2005/06 Est. Est. Budget Est. Est. Budget US$million % of total budget HumanCapitalDevelopment 688 843 1,549 25 25 31 1.1 Refugeesand IDPreturn 33 75 66 1 2 1 1.2 Education 185 224 485 7 7 10 1.3 Health 168 154 259 6 5 5 1.4 Livelihoodand Social Protection 273 340 668 I O 10 13 1.5 Culture,Media, Sport 29 50 71 1 1 1 PhysicalInfrastructure 151 901 1,223 21 21 24 2.1 Transport 452 581 542 16 17 11 2.2 Energy, Mines, and Telecom 113 222 327 4 7 6 2.3 NaturalResourcesManagement 97 45 207 3 1 4 2.4 UrbanManagement 88 53 147 3 2 3 General Administration 251 298 594 9 9 12 3.1 Trade and Investment 8 40 78 0 1 2 3.2 Public AdministrationandEconomicManagement 243 258 517 9 8 I O Security 1,105 1,328 1,615 40 39 32 3.3 Justice 16 45 63 1 1 1 3.4 NationalPoliceandLawEnforcement 209 340 597 7 10 12 3.5 AfghanNational Army 797 788 830 28 23 16 3.6 Mine Action 63 96 77 2 3 2 3.7 DDR 21 59 48 1 2 1 R Rh n 2 Core Core Development -- Operating 165 317 1,209 6 9 24 614 814 1,881 22 26 37 External 2,182 2,503 3,180 78 74 63 Total Budget 2,196 3,318 5,067 100 100 100 Source: MoF (preliminary annual statementsfor operating budget and 200.5/06 Budget Decree estimatesfor development). 98 LIST OFADDITIONAL DOCUMENTS PREPAREDAS PART OFTHE PFMREVIEW Inadditionto the thematic chapters inVolume I1to V, a numbero fbackground paperswere prepared and shared with the Government duringthe process o f the PFMreview: Additional backgroundpaperson the budget and fiscal management: Budget Formulation inPost Conflict Afghanistan (2005/06), December 2004 Public Finance inAfghanistan: Trends and Prospects, February 2005 State-Owned Enterprises and the National Budget, February 2005 Operating Budget Allotment Control, February 2005 Overview of the 1384Budget, March2005 Reforming the Fiscal System and Achieving Fiscal Sustainability, March2005 Policy notes on PRR and pay reform, March - June 2005 Strengthening ARTF and Enabling Budget Support Through Selective Use o f Benchmarks, July 2005 Additionalbackgroundpapers on institutionalreforms: Reforming the Treasury inAfghanistan, December 2004 Afghanistan Ministry o f Finance -The Way Forward, December 2004 / February 2005 Institutions for Economic Policymalung: Key Considerations and Options, February 2005 Ministry o f Finance: Toward a Five-Year Strategic Plan, May/June 2005 Additionalbackgroundpapers on sectors: Public ExpenditureManagement inLivelihoods and Social Protection- summary o f findings, February 2005 Public ExpenditureReview of the Highway Systems in Afghanistan, February 2005 Public ExpenditureManagement inthe Security Sector: Initial Findingsand Conclusions, April 2005 Education Expenditure Traclung Survey -report on key findings, June 2005 99 REFERENCES Boesen, I.(2004). "From Subjects to Citizen: Local Participation in the National Solidarity Program (NSP)". Kabul, AREU. Byrd, William, and Chnstopher Ward (2004). "Drugs and Development inAfghanistan." World BankSocial Development Papers, No. 18 (December). Department for International Development (DFID, 2005). "A Platform Approach to Improving PublicFinancial Management", DFID Briefing. London, UK (July). Evans, Anne, N i c k Manning, Yasin Osmani, Anne Tully, and Andrew Wilder (2004a). A Guide to Government in Afghanistan. Washington, D C and Kabul Afghanistan: The World Bank and Afghanistan Research and Evaluation Unit. (2004b). Subnational Administration in Afghanistan: Assessment and Recommendations for Action. Washington, D C and Kabul, Afghanistan: The World Bank and Afghanistan Research and Evaluation Unit. Evans, Anne, and Yasin Osmani (2005). Assessing Progress: Update Report on Subnational Administration in Afghanistan. Afghanistan Research and Evaluation Unit Issues Paper Series (June). Government o f Afghanistan with international agencies (2004). Securing Afghanistan 's Future: Accomplishments and the Strategic Path Forward. Gupta, Sanjeev, Benedict Clements, and Gabriela Inchauste (2004). "Helping Countries Develop: TheRole of Fiscal Policy ",InternationalMonetary Fund,Washington, D.C. Heller, Peter (2005). "Understanding Fiscal Space", IMF Policy DiscussionPaper, 05/4. Human Rights Research and Advocacy Consortium (HRRAC, 2004). "Take the Gun Away: Afghan Voices on Security and Elections". Kabul (September). IMF(2005). Islamic State of Afghanistan: SelectedIssues and Statistical Appendix. Washington, DC: InternationalMonetary Fund(February). Kaufmann, Daniel, Art Kraay, and Massimo Mastruzzi (2005). "Governance Matters IV: Governance Indicators for 1996-2004", draft. Keen, Michael, and Alejandro Simone (2004). "Tax Policy in Developing Countries: Some Lessons from the 1990s and Some Challenges Ahead. Gupta, Sanjeev et al, "Helping Countries Develop: TheRole of Fiscal Policy ",Washington, DC: InternationalMonetary Fund. Lister, Sarah (2005). "Caught in Confusion: Local Governance Structures in Afghanistan." AREUBriefingPaper(March). Loevinsohn, Benjamin, and A. Harding (2004). "Contracting for the Delivery o f Community Health Services: A Review o f Global Experience." World Bank. Operations Evaluations Department (OED, 1998). World Bank's Experience with Post-Conflict Reconstruction,. Washington, DC: World Bank. PEFA Secretariat (2005). "Public Financial Management Performance Measurement Framework." Washington, DC: World Bank (June). Sarraf, Feridoun (2005). "Integration o f Recurrent and Capital `Development' Budgets: Issues, Problems, Country Experiences, and the Way Forward." World Bank Public Expenditure Working Group ImplementationNote. Schiavo-Campo, Salvatore, de Tommaso, Giulio and Mukherjee, Amitabha (1997). "Government Employment and Pay: A Global and Regional Perspective". World Bank Policy Research Working Paper No. 1771. 100 Strong, Lesley, Abdul Wali, and Egbert Sondorp (2005). "Health Policy in Afghanistan: Two Years of Rapid Change (A Review of the Process from 2001 to 2003)." London School of HygieneandTropical Medicine(LSHTM). Swaroop, Vinay (2004). Mimeo on Medium-TermExpenditureFramework. Ward, Christopher, and William Byrd (2004). "Afghanistan's Opium Drug Economy." World Bank SouthAsia RegionPREM WorkingPapers,No. 5 (December). World Bank (2003a). Breaking the Conflict Trap: Civil War and Development Policy. Washington, DC:World Bank. World Bank (2003b). WorldDevelopment Report 2004: Making Services Workfor Poor People. Washington, DC:World Bank. World Bank (2005a). Afghanistan-State Building, Sustaining Growth, and Reducing Poverty. Washington, DC:World Bank,2005. World Bank (2005b). Afghanistan National Reconstruction and Poverty Reduction - the Role of Women in Afghanistan 's Future. Washington, DC: World Bank,2005. World Bank (2006). Afghanistan: Investment Climate Assessment. Washington, DC: World Bank,forthcoming. 101