Report No. 25611-KH Cambodia Enhancing Service Delivery through Improved Resource Allocation and Institutional Reform Integrated Fiduciary Assessment and Public Expenditure Review September 8, 2003 Poverty Reduction and Economic Management Sector Unit Mekong Department East Asia and the Pacific region Cambodia Resident Mission A Joint Document of the World Bank and the Asian Development Bank CURRENCYAND EXCHANGE RATES Currency Unit -CambodianRiel ExchangeRate (as of August 2003): CR 3,835 FISCALYEAR (January 1 to December 31) Vice President Jemal-ud-din Kassum CountryDirector(World Bank) Ian Porter Country Director(ADB) Urooj Malik SectorDirector Homi Kharas SectorManager Barbara Nunberg Task TeamLeader (World Bank) RobertTaliercio TaskTeamLeader (ADB) Barend Frielink Abbreviations and Acronyms AC Asphaltic Concrete MPA Minimum Package of Activities ADB Asian Development Bank MPWT Ministry of Public Works andTransport ADD AcceleratedDistrictDisbursement MRD Ministry of Rural Development ADESS Agricultural DevelopmentSupportto Seila MSF Medicins Sans Frontiers APIP Agricultural Productivity ImprovementProject MTEF MediumTerm ExpenditureFramework AusAID Australian Agency for InternationalDevelopment MWVA Ministry of Women's andVeterans' Affairs BMC Budget Management Center NBC National Bank of Cambodia Bn Billion NCHADS NationalCenter for AIDS, DermatologyandSTD CAFIMEX CambodiaFishImport andExportCompany Control CAR Councilfor AdministrativeReform NGO Non GovemmentalOrganization CARD1 CambodiaAgricultural ResearchandDevelopment Institute NHA NationalHealthAccounts CDC Councilfor the Developmentof Cambodia NR NationalRoad CDRI CambodiaDevelopmentResourceInstitute NRRRP NationalRoadRehabilitationandReconstruction CENAT NationalCenter for Tuberculosis andLeprosy Program CIF Carriage,Insuranceand Freight NRSC NationalRoadSafety Council CMS CentralMedical Stores OD Operational District CNM Center for Entomology, ParasitologyandMalariaControl OPM Office of the PrimeMinister COM Council of Ministers PAP Priority Action Program CPA ComplementaryPackage of Activities PDAFF Provincial Departmentof Agriculture, Forestry and CR CambodianRiel Fisheries CSES CambodianSocio-EconomicSurvey PDB Partnership Database Danida DanishAgencyfor DevelopmentAssistance PDRD Provincial Departmentof RuralDevelopment DBST Double BituminousSurface Treatment PDWRAM Provincial Departmentof Water Resources and DFID Departmentof InternationalDevelopment Meteorology DHS DemographicandHealthSurvey PER PublicExpenditureReview DORD DistrictOffice of RuralDevelopment PHD Provincial Health Department DRR Departmentof Rural Roads PIP Public InvestmentProgram EFA Educationfor All PIU Project ImplementationUnit ESP EducationStrategic Plan PMU ProjectManagementUnit ESSP EducationSector Support Program PORR Provincial Oftice of RuralRoads EU EuropeanUnion PPWTD Provincial PublicWorks andTransportDepartment ExCom ExecutiveCommitteeof PRDC PRASAC Support Programfor the Agricultural Sector in FA0 FoodandAgricultural Organization Cambodia FBC FinanceandBankingCommittee PRDC ProvincialRuralDevelopmentCommittee FRMR Fundfor Repair andMaintenanceof Roads PRS Poverty ReductionStrategy GDP Gross Domestic Product PRSP Poverty ReductionStrategy Paper GMS Greater MekongSub-region PSP PrivateSector Participation GOC Govemmentof Cambodia FT Provincial Treasury ICB International CompetitiveBidding PTTC PrimaryTeacher Training College IFAPER IntegratedFiduciaryAssessment andPublicExpenditure RGC Royal Govemment of Cambodia Review RRG RuralRoadsGroup ILO InternationalLabor Organization RTTC RegionalTeacher Training College IMF International Monetary Fund SEDP Socioeconomic DevelopmentPlan JFPR JapanFundfor Poverty Reduction SIDA Sweden's InternationalDevelopmentAgency JICA JapaneseIntemationalCooperationAgency SWAP Sector Wide Approach LBAT Labor-BasedAppropriate Technology SWIM Sector Wide Management MAFF Ministry of Agriculture, Forestry andFisheries TA TechnicalAssistance MCE Military Corps of Engineers T O E Table of StateFinancialOperations MEF Ministry of EconomyandFinance TCAP TechnicalCooperationAssistance Program MIA Ministry of IntemalAffairs TSS Transport Sector Study MLMUPC Ministry of LandManagement,Urban Planningand W E T TechnicalandVocational EducationandTraining Construction UN UnitedNations MOC Ministry of Commerce UNDP UnitedNations DevelopmentProgram MOE Ministryof Environment UNICEF United Nations Children's Fund MEF Ministryof Economy and Finance USAJD UnitedStates Agency for InternationalDevelopment MOEYS Ministryof Education,Youth andSports VAT Value Added Tax MOH Ministryof Health vocs Vehicle OperatingCosts MOI Ministryof Interior WB World Bank MOP Ministry of Planning WFP World FoodProgram MOU Memorandumof Understanding WHO World HealthOrganization ACKNOWLEDGEMENTS The Integrated Fiduciary Assessment and Public Expenditure Review (FAPER)-a joint document of the World Bank and the ADB-has been prepared by a core team comprising several counterpart teams led by the Royal Government of Cambodia (RGC) and including the World Bank, Asian Development Bank, and InternationalMonetary Fund. The chief of the RGC Counterpart Team i s Dr. Hang Chuon Naron, Deputy Secretary General, Ministry of Economy and Finance, and Economic Assistant to the Prime Minister. Dr Naron oversaw the work from the Government side. Other RGC team members include: H.E. Chou KimLeng; Dr. Sok Saravuth; Dr. Chea Vuthna; Chhim Sareth; Ieng Sunly; Seang Kung; Hav Ratanak; Mr. Leang; and Phiyorin Tep. Counterpart team members from other Government ministries and agencies include: MOEYS-H.E. Pok Than, Secretary of State; Chhay Aun, Director General; Sar Nak, Deputy Director; Thong Borann, and Put Samith; MOH-Dr. Kuyseang Te, Director General, and Dr. Char Meng Chuor, Director; MAFF-Nep Chanthet, Director; Kith Seng; Che Savun; Sim Hang, Director; and Lauv Ny; MRD-Chan Sovandy; MOWVA-H.E. Mrs.You Ay, Secretary of State; MPWT-Chhin KongHean, Director; CAR- H.E. Sum Manit, Secretary of State; Ngo Hongly; Chhuon Chham, Under Secretary General; Svay Sovannrith; ChhimPhavin; Hok Peng Se; Hang Chham Chhan; Kong Sophy; Sar Chanthy; Ting Sokhay; Alain Benicy; and Michael Dalton; and officials from the SSCS and CDC. The World Bank and ADB would like to extend their most sincere thanks to the RGC Counterpart Team-including those officials not named here but who attended the IFAPER workshops-for its valuable guidance and significant contributions to the overall product. The overall World Bank-ADB counterpart team comprised the following members: 0 for the WorldBank -Geoff Dixon, Hout Chea, Amanda Green, and Adrian Fozzardand Rob Taliercio (Task Team Leaders); Sally Burningham, Imogene Jensen, Christina Malmberg- Calvo; Rosemary Bellew, LuisBenveniste, Chris Thomas, Peter Moock, Naoki Umemiya; Louise Scura, Marianne Grosclaude, Bill Magrath; Vincent Turbat; Wijaya Wickrema, Agnes Albert- Loth, Nancy Chen; Gillian Brown, Helen Brereton, Susan Razzaz, and Pamornrat Tansanguanwong; Magdi Amin; and consultants Paul Smoke, MalcolmGreen, Gwen Quere, Tim Ensor, and John Lee; 0 for the Asian Development Bank -Barend Frielink, Anthony Jude, Clay Wescott, Erik Bloom, Rottanak Keo, Chamroen Ouch, Gudrun Forsberg, Masahiro Otsuka, Indu Bhushan, Sandra Nicoll, Sukanda Lewis, Peter Broch, Ben Diokno, Catherine Dom, Mike Ratcliffe, and NihalFernandopulle. The IMF team consisted o f Robert Hagemann, Philippe Marciniak, Alejandro Lopez- Mejia, Matt Davies, JimBrumby, and Sophana Sa. Numerous others also contributed significantly to the review: Anders Frankenberg (SIDA); Hank Bekedam and Aye Aye Thwin (WHO); Omar Bargawi and Daniel Arghiros (DFID); Joanne Morrisonand Scott Leiper (Seila); Jean-Francois Frys (MEDICAM);Russell Peterson (NGO Forum); Dominique McAdams, Kaspar Bro Larsen, Ernest0 Bautista, and Richard Schiere (UNDP); Sr. Regina Pellicore (MaryknolYEDUCAM); Elaine McKay (PGE); Paula Quigley (GRZ); and David Salter (LO). Inaddition the team thanks the Australian Government for trust fundresources. The team has received overall guidance and direction from: Homi Kharas, Ian Porter, Urooj Malik, Barbara Nunberg, Kathy Krumm, Kazi Matin, Ronald Points, Bonaventure Mbida- Essama, Nisha Agrawal, Su Yong Song, Steve Schonberger, Tom Rumbaugh, I1Houng Lee, and BillDorotinsky. The team alsoreceived excellent support from Gloria Elmore. Peer reviewers were: David Shand, Jeni Klugman, Robert Ebel, Pierre Messali, and Vinaya Swaroop. The task team has benefited greatly from the reviewers' specific advice and direction. Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review" Table of Contents EXECUTIVE SUMMARY ........................................................................................................... vii Introduction:The Approach and its Findings............................................................................ vii ImprovingResourceMobilizationto EnsureAggregate FiscalSustainability......................... ... vi11 ReducingFiduciaryRiskto PublicFunds.................................................................................. ix RationalizingPublic ExpenditurePolicyandManagement....................................................... xi Pay andEmploymentReforminthe CambodianCivil Service ............................................... xiv Conclusion................................................................................................................................ xvi PriorityActionMatrix: Summary of Recommendations........................................................ xviii ... 1 THE MACROECONOMICFRAMEWORKAND MEDIUM-TERMFISCAL SUSTAINABILITY: OPPORTUNITIESAND CHALLENGESFORPOVERTYREDUCTION. Introduction................................................................................................................................. 1 ................................................................................................................................................. 1 Overviewof Performance............................................................................................................ 2 MediumTerm MacroeconomicFrameworkandFiscalSustainability........................................ 4 Conclusion................................................................................................................................... 8 2. Introduction............................................................................................................................... RESOURCEALLOCATION: TRENDS, OPPORTUNITIES, AND CHALLENGES.......11 11 Trends inthe FunctionalCompositionof Spending.................................................................. 11 Trends inthe Economic CompositionofExpenditure............................................................... 16 Externally-FinancedExpenditure.............................................................................................. 19 EDUCATION,AGRICULTUREAND ROAD TRANSPORT.................................................... 3. Conclusion................................................................................................................................. 22 POLICIES,EXPENDITURES,AND OUTCOMES INPRIORITY SECTORS:HEALTH, 23 Introduction............................................................................................................................... 23 Education................................................................................................................................... Health........................................................................................................................................ 27 38 RoadTransport .......................................................................................................................... 48 Agriculture................................................................................................................................. 55 Conclusion................................................................................................................................. 59 4 . REFORMINGINSTITUTIONSTO IMPROVESERVICEDELIVERY I: PUBLIC THE EXPENDITUREAND FINANCIAL MANAGEMENT SYSTEM............................................. 61 61 Aligning Resourceswith Policies and Strategies...................................................................... The ReformAgenda .................................................................................................................. 63 ImprovingBudget Execution.................................................................................................... -69 ReformInitiatives- PriorityActionPrograms.......................................................................... 74 BeyondPAP-Toward aUnifiedBudgetExecutionSystem.................................................... 77 Treasury Operations andCashManagement............................................................................. 79 84 Conclusion................................................................................................................................. StrengtheningTransparency andAccountability....................................................................... 89 CAMBODIANCIVIL SERVICE.................................................................................................. 91 5. REFORMINGINSTITUTIONSTO IMPROVESERVICEDELIVERY11:THE IntroductionandReformAgenda.............................................................................................. 91 CambodiainInternationalPerspective: The Cost and Size ofthe Civil Service....................... 93 ImprovingCivil ServicePay: Current PayLevelsandEvidenceofInadequacy..................... 96 * This combines the analysis of a public expenditure review (PER) and a country financial accountability assessment (CFAA). A country procurement assessment report (CPAR) has been carried out in parallel with this work. RationalizingCivil Service Employment: The Current Size and Distribution of the Civil 99 Pay Levels and The MediumTerm Wage Framework............................................................ Service ....................................................................................................................................... 103 Strengthening Civil Service Management: SafeguardingManagerial and Fiduciary Controls ................................................................................................................................................. 109 Conclusion............................................................................................................................... 112 6. REFORMINGINSTITUTIONS TO IMPROVE SERVICE DELIVERY 111: DECENTRALIZATION.............................................................................................................. 117 Reform Agenda and Institutional Framework......................................................................... 117 Decentralizationto the Communes and Sangkats ................................................................... 119 Deconcentration to Provinces and Municipalities ................................................................... 125 Implementation........................................................................................................................ Key Opportunities and Challenges inMovingForward : Strategy, Capacity Buildingand 131 7. CONCLUSION: TOWARD A STRATEGYOF REFORM.............................................. 133 TABLES Table 1: Key Issues inCambodian Public Expenditure: Policies and Institutions ....................... vii Table 2: Cambodia: Comparisonof Tax Revenue Structure with other Selected Countries .........ix Table 3: The Back-loading Problem: Percentageof Expenditures Posted inDecember. 2000 and 2001......................................................................................................................................... x Table 4: Linkages between Sector Policy. Expenditure. and Outputs/Outcomes .......................... xii Table 5: Share of Education Subsidy Received by Income Quintile ........................................... xiii ... Table 6: Median Monthly Remuneration...................................................................................... xv Table 7: Private Sector Monthly Wage Data. 1999 ...................................................................... xv Table 1.2: Cambodia: Comparisonof Tax Revenue Structure with Other Selected Countries .......36 Table 1.1: ExemptedImport Dutiesby ExemptionRegime............................................................ Table 1.3: MAFF RevenueCollections: Estimatedand Actual. 2001-2002.................................... 8 Table 2.1: Government and Externally Financed Expenditureby Summary Function. 1996to 12 Table 2.2: Financing Reallocation. 1998 to 2001......................................................................... 2001....................................................................................................................................... 13 Table 2.3: Treasury-Executed Spendingby Function, 1996-2003 ............................................... 13 Table 2.4: Treasury-Executed Spending by Function, 1996-2003 ............................................... 15 Table 2.5 : Treasury-Executed Civil Administration Expenditure by Economic Category, 1996- Table 2.6: Expenditureon Government Projects by Function, 1996-2001.................................... 2003....................................................................................................................................... 17 18 Table 2.7: Government Projects as Share of Total Sector Expenditure, 1996-2001 .................... 19 Table 2.8: Functional Structure of External Assistance, 1996 to 2000.......................................... 20 Table 2.9: Total External Assistance by Summary Function, 1996-2001 ..................................... 21 Table 2.11:Expenditureby Source of Financing and Administrative Level, 2001 ...................... Table 2.10: Expenditureby Source of Financing and Input, 2001 .............................................. 21 22 Table 3.1: Headcount Indices, Overall Poverty Line, 1997 and 1999........................................... 24 Table 3.2: The Health Sector Strategic Plan, 2003-2007 ............................................................... 28 30 Table 3.4: ReportedUser Charges inPublic Facilities, 1998-2001, CR millions.......................... Table 3.3: RGC Spending by Economic Classificationand Level of Spending............................ 32 Table 3.5: Trends in Selected Utilization, Access and Health Status Indicators ........................... 34 Table 3.6: Use of Health Facilities, Knowledge of Health and Health Status by Socio-Economic 35 Table 3.7: The Education Sector Strategy Paper and Strategic Plan, 2001-2005 .......................... Status .................................................................................................................................... 39 Table 3.8: EstimatedTotal Expenditure on Education, 1997-2002 ............................................. Table 3.9: Distribution of Government Recurrent Expenditure, 1997 - 2002 .............................. 40 41 11 Table 3.10: PrivateRecurrent Expenditure. 1997......................................................................... 42 Table 3.11: ESP Targets Compared with Actuals......................................................................... 45 Table 3.12: Share of Education Subsidy Received by Income Quintile ....................................... 45 Table 3.13: Per Student Recurrent Expenditure. Public and Private before and after PAP Subsidy ............................................................................................................................................... 47 Table 3.14: Road Transport Sector -NPRS Action Plan Matrix .................................................. 51 Table 3.16: EstimatedExpenditures on Roads Only. MPWT and MRD.2001............................. Table 3.15: Total Current and Capital Expenditures. MPWT and MRD. 1997-2001 ...................50 Table 3.17: Annual Routine Maintenance Cost (US$) on DifferentNetwork Size Options ........52 53 Table 3.19: Linkages between Sector Policy. Expenditure. and Outputs/Outcomes ..................... 60 Table 4.1: Variance inBudget Execution. 2000 and 2001 ............................................................ 70 71 Table 4.3: Chapter 13 and 11Budget Execution2000. 2001 and November 2002...................... Table 4.2: Back-loading: Share of Expenditure PostedinDecember. 2000-2001 ........................ 76 Table 4.4: Structure o f Government Financial Deposits............................................................... 81 Table 5.1: The National Program for Administrative Reform and Strategy to Rationalize the Civil 92 Table 5.1: Median Monthly Remuneration. 2002.......................................................................... Service. 2002-2006................................................................................................................ 97 Table 5.2: Private Sector Monthly Wage Data. 1999.................................................................... 97 100 Table 5-4: Civil Service Numbers. 2003...................................................................................... Table 5.3: Historical Size of the Civil Service by Source ........................................................... 101 Table 5.5: The Wage Bill. including Civil Administration and Defense/Security. as a Percentage of Recurrent Expenditures. Budgeted and Actual. 1998-2002............................................ 104 Table 5.6: Cambodia: MediumTerm Wage BillFramework. 2002-2007 .................................. 105 Table 5.7: Personnel Expendituresas a Percentage of Recurrent Spending. 1998-2005. in Education. Health. and Agriculture ..................................................................................... 106 Table 5.8: EstimatedExternalFinancingRequirementsunder Wage Bill Scenarios. 2003-2006 Table 5A.1: Civil Service Average Monthly Wage Projections. 2003-2007 ............................. ............................................................................................................................................. 108 113 Table 5A.2: Projected Civil Service Size (Year End) by Scenario. 2003-2006 .......................... 114 Table 5A.3: Projected Monthly Civil Service Wage (US$). 2003-2007 ..................................... Table 5A.4: Combined Wage Bill andRetrenchment Scenarios. Average Monthly Wage ........114 115 Table 6.1: Number of Districts/Sectors. Communes/Sangkats and Villages by Table 6.2: IntergovernmentalTransfers to Communes: Base Case Resource Envelope ............118 Province/Municipality inCambodia ................................................................................... 121 Table 6.3: Expenditureby Provincial Administrations as a Share of Total Expenditureby Function............................................................................................................................... 127 Table 6.4: Allocation of Health. Education and other Provincial Agency BudgetAllocations by Province. 2002 ................................................................................................................... 128 FIGURES Figure 1: Monthly Salary Payments inMOH. 2001 .......................................................... ix Figure 2: Health Benefits Incidence ConsumptionGroup .................................................... xiii Figure 1.1: Cambodia: Contributions to Real GDP Growth, 1998-2002........................................ 1 Figure 1.2: Cambodia: Total Tax Revenue(Centra1Government).................................................. 3 Figure 3.1: Health Sector Fundingby Source o f Finance.............................................................. 29 Figure3.2: Provincial ResourceAllocation, RGC Actual HealthCare Spending......................... 31 Figure3.3: Benefits Incidence Consumption Group..................................................................... 36 Figure3.4: Gender Benefits Incidence of Government Spending................................................. Figure3.5 Shareof Government Recurrent Expenditureby Level ofEducation, 1997-2002 .......36 42 iii Figure 3.6: Enrolment 1996/97-2001/02........................................................................................ Figure3.7: Grade 1Enrolment by Community Wealth Quintile, 1999-2001................................ 43 Figure3-8: Condition ofthe MPWTNetwork, 2001 ..................................................................... 44 Figure4.1: Monthly Salary PaymentsinMOH, 2001.................................................................. 54 Figure 4.2: Expenditureson Modified-Cash and CashBasis, by month 1999 to 2002 ................72 Figure4.3: DomesticRevenues andTreasury-Executed Expenditure(Cash Basis), 2000-2002 80 ..72 Figure 5.1: Total Central Government Wage Bill as aPercentageof Current Expenditures: Low and Middle IncomeCountries ............................................................................................... 94 Figure 5.2: Wage Bill as a Share of Total Revenues, Low and Lower Middle Income Countries 94 Figure 5.3: Civil Service Employment as a Share of Population: Low Income Countries...........95 Figure5.4: Age and Gender Distribution ofthe Civil Service in2000 ...................................... 102 Figure5.5: Cross-Country Comparisons of PersonnelExpendituresas aPercentage ofRecurrent Spending inEducationand Health...................................................................................... 106 BOXES Box 1: The Cambodia Integrated Fiduciary Assessment and Public ExpenditureReview (IFAPER) ............................................................................................................................. ... v111 Box 2: Corruption inCambodia: Evidence from Citizens, Firms,and Public Officials .................x Box 1.1: Corruption inCambodia: Evidence from Citizens, Firms, and Public Officials..............5 Box 1.2: RevampingFiscalBureaucracy: The Semi-Autonomous RevenueAuthority Model in Developing Countries.............................................................................................................. 7 Box 1.3: Fiscal Issues inNatural Resource Management............................................................... 8 Box 3.1: Results-Oriented Public Expenditure Management inCambodia................................... 25 Box 3.2: The Cambodian Experience with Contractingfor Service Delivery............................... 34 Box 3.3: Women and Service Provision I: 37 Box 3.4: Women and Service Provision11: Barriers to Access inEducation................................ Barriers to Access inHealth....................................... 47 Box 3.5: Policies and Strategiesfor Expandingthe Role of the Private Sector............................ 49 51 Box 3.7: Women and Service Provision111:Barriers to Access inAgriculture ........................... Box 3.6: The Fundfor the Repair and Maintenance of Roads (FRMR)....................................... 59 Box 4.1: Cambodia's PEFM System inComparative Perspective ................................................ 62 Box 4.2: Accountability Mechanisms inPractice: InternationalExperience with Participatory Public ExpenditureManagement .......................................................................................... 63 Box 4.3: LinkingPolicies and Budgets: The Case of the National PovertyReduction Strategy 64 Box 4.4: A Tradition of Rent-seeking?.......................................................................................... (NPRS) .................................................................................................................................. 85 Box 5.1: Tripling or Quadrupling?A Health Sector "Experiment" Shows that Public Sector Wages Needto be Increased by Multiplesof Current Levels............................................... 98 Box 6.1: Improving Geographical Targeting inHealth and Education....................................... 130 ANNEX A Table A1 Cambodia: Medium-TermMacroeconomic Framework............................................. 137 Table A2 Public ExpenditureCambodia and InternationalComparators.................................... 138 Table A4 Treasury-Executed Civil Administration Expenditureby Economic Category ...........138 Table A3 Treasury-Executed Spending Functions ...................................................................... Table AS Ratio of Operating Expenditures to Salaries and Allowances by Function .................139 139 Table A6 External Assistance as Percentageof Total Sector Expenditure.................................. Table A7 Treasury-Executed Civil Administration Expenditureby Economic Category ...........140 140 Table A8 Civil Servants by Ministryand Location ..................................................................... 141 Table A9 Country Comparisons of Education Indicators and Expenditure................................. 142 iv Table A11Government andPrivateRecurrentExpenditureby LevelofEducation...................143 Table A10 Total GovernmentandExternalExpenditureon Education...................................... Table A12ComparisonofESPProgramCost EstimatesandMTEFBudgetAllocations..........144 145 Table A13 Healthand Socio-economic Indicatorsfor SelectedCountries.................................. Table A14 Cross-country Comparisons of Coverage.................................................................. 146 147 Table A15 Major RevenueSources ofProvincia1Governments................................................. 148 Table A16 Major Expendituresof ProvincialGovernmentsas a Percentageof TotalExpenditure ..................................................................................................................................................... 149 Table A17 BudgetVariance......................................................................................................... 150 ANNEX B Table B1 Summary ofAchievements .......................................................................................... 151 ANNEX C CountryFinancialAccountabilityAssessment............................................................................ 161 APPENDIXES Appendix 5A................................................................................................................................ Appendix B Cambodia: CountryFinancialAssessmentAccountability..................................... 113 195 REFERENCES........................................................................................................................... 197 V EXECUTIVESUMMARY INTRODUCTION: THEAPPROACH AND ITS FINDINGS 1. Since 1999 Cambodia has made significant headway in reforming public expenditure policy and management, yet in order to implement its development agenda, Cambodia will have to make much more progress on four principal fiscal, fiduciary, and institutional challenges. First, Cambodia will have to improve resource mobilization to ensure aggregate fiscal sustainability. Second, to reduce the fiduciary risk to public funds, the Government will have to engage in comprehensive reform of budget execution, cash management, and public financial control systems. Third, the Government will have to rationalize public expenditure policy and management further to carry out both its Second Socioeconomic Development Plan (SEDP) and its National Poverty Reduction Strategy (NF'RS). Last, Cambodia will have need to undertake comprehensive civil service reform-focusing on pay and employment issues-in order to deliver poverty-reducing services. The following sections elaborate on these four core challenges and provide an outline of proposed solutions. Table 1and Box 1summarize the approach and its principalfindings. Table 1: Key IssuesinCambodianPublicExpenditure:PoliciesandInstitutions PublicPolicy,Expenditure, CivilService Reform Decentralization andFinancialManagement Aggregate Fiscal Medium term sustainability Tight control at the The decentralized Sustainability depends on improved domestic aggregate fiscal level, low communes are not yet an resourcemobilization and revenues, and expenditure aggregate macro-fiscal adequate, effective public sector policy result in a civil issue, due to the expenditure in economic services administration wage bill Government's careful (mainly road transport and that i s too low. approach,but own-source agriculture) to finance the poverty- revenuesare a medium- reduction and growth agenda. long term concern. Prioritizationof Weaknesses inboth the budget The share of expenditure The allocative efficiency Resource formulation and execution on civil administration of provincial Allocation processesundermine allocative humanresources i s too expenditures, interms of efficiency by resulting in the low, and there i s roomto targeting the poor, could misallocation of resources:across improve the sectoral and be improved; the and within sectors, across regions, geographic deployment of efficiency of commune to economic inputs,and over civil servants. spendingwill dependon time. future decisionsregarding the level and mode of intergovernmental transfers. Efficient and Major problems in Treasury Excessively low wages Progressis being madein Effective Use of operations and cash management, lead to myriad moving service delivery Resourcesfor andweaknessesinthe control dysfunctions, including closer to the poor, but ServiceDelivery environment, impede the delivery cofruption, and greatly muchwork i s neededon of public resources, making the underminecivil service the institutional and public expenditure andfinancial capacity. administrative support managementsystema source of structure. fiduciay risk. vii viii Cambodia: Integrated Fiduciaw AssessmentandPublic Expenditure Review Box 1: The CambodiaIntegratedFiduciaryAssessment andPublicExpenditureReview(IFAPER) The approach of the Cambodia Integrated Fiduciary Assessment and Public Expenditure Review (IFAPER) is fundamentally integrated and multidisciplinary in that it focuses on both policies and institutions in the delivery of poverty-reducing services. In the case o f Cambodia it is clear that both the reorientation of policies as well as the reform of institutions are required to improve the impact of services provided by the public sector. Building on the World Bank's 1999 Public Expenditure Review and the ADB's 2001 study entitled Enhancing Governancefor Sustainable Development, the IFAPER expanded the sectoral scope of the analysis by including other priority sectors-agriculture and road transport-in addition to education and health, and extended the institutional coverage to include the civil service and decentralization, in addition to public expenditure and financial management institutions. The augmented scope of the IFAPERi s necessary inthe context of the SEDP II/NPRS and for policy-based programmatic lending approaches. The approach is thus to focus on the three levels of public expenditure policy and management-aggregate fiscal sustainability, prioritization of resource allocation, and the efficient and effective use of resources for service delivery-in the institutional context of service delivery-public expenditure andfinancial management (PEFM), civil service reform, and decentralization. The analysis yields four principal findings: 9 Resource mobilization must be improved to ensure fiscal sustainability in light of the poverty > reduction and growth agenda; Weaknesses in the public expenditure and financial management system create unacceptably high levels o f fiduciary riskto public funds; 9 In order to meet the poverty reduction targets the Government will need to improve the effectiveness of public spending by more tightly linking it with priority outcomes and by reallocating resources from lower priority sectors, programs, and functions; 9 Given the serious problems afflicting the civil service-low pay, low skills, and thus low capacity-omprehensive civil service reform will have to be initiated inthe short term. IMPROVINGRESOURCEMOBILIZATION TO ENSURE AGGREGATEFISCAL SUSTAINABILITY 2. Despite the need for higher revenues to finance growth and poverty reduction, Cambodia's fiscal revenue ratios, especially tax revenue, remain among the lowest in the world. Cambodia's total revenue to GDP ratio stood at 11.7 percent in 2001, with tax revenues accounting for only 8.4 percent (see Table 2). Indeed, without significant increases in revenues, there are serious risks to bothmediumtermmacro-fiscal sustainability and the Royal Government of Cambodia's (RGC's) medium term poverty reduction program. To retain public debt sustainability and to finance higher levels of public spending, revenues will have to be greatly increased. 3. To be sure, Cambodia has made progress since the last Public Expenditure Review in 1999. As a result of the reformefforts, government revenue rose from 8.3 percent of GDP in 1998 to 11.7 percent in 2001, though the increase inrevenue was mostly due to growth inthe tax base. The replacement of the turnover tax and consumption tax on imports with a 10 percent value added tax (VAT) in 1999 also contributed to the revenue increases. At the same time, the VAT enhanced the efficiency of the tax system by simplifying the tax structure, widening coverage, and reducing cascading. 4. It is clear that continued improvement, on the order of what has already been accomplished, is necessary.The Government plans to raise revenues to 14 to 15 percent of GDP by 2007. Facilitating domestic resource mobilization i s likely to occur through improved tax administration in the short to medium terms and improved tax policy in the medium to long terms. As tax administration reform cannot depend on civil service reform in the short term, Executive Summary ix however, the Government should investigate options for rapid improvement through establishment of a semi-autonomous revenue authority. Table 2: Cambodia: Comparison of Tax Revenue Structure with other Selected Countries Fiscal Revenue 11 ~~ Total Tax Direct Taxes Per capita Revenue Revenue only GDP(US$) PRGFAsiancountries 16.2 12.8 3.1 427 Bangladesh 8.6 7.1 1.2 370 Lao P.D.R. 11.4 9.2 2.0 310 Nepal 11.2 9.2 2.1 240 Sri Lanka 17.0 14.7 2.4 830 Vietnam 20.6 15.6 5.5 410 Mongolia 28.5 20.8 5.3 400 PRGFSub-Saharan 17.6 13.7 3.9 323 African Countries --of which selected 14.4 11.4 1.8 277 agriculturalcountries Cameroon 17.8 12.2 3.3 570 GuineaBissau 19.5 10.8 0.0 160 Mali 14.7 14.0 2.6 210 Niger 8.9 8.3 1.5 170 Tanzania 11.8 10.6 2.7 280 Togo 13.8 12.3 0.4 270 Cambodia (2001) 11.7 8.4 1.o 259 Source: IMF staff country reports. REDUCINGFIDUCIARYRISKTO m L I CFUNDS 5. It has become increasingly apparent that weaknesses in the public expenditure and financial management system (PEFM) not only have high costs in terms of allocative and operational efficiency but also create unacceptably high levels of fiduciary risk to public funds. The cash-basedpayments system has emerged as a major constraint. Increasingly, budget execution has suffered from delays and an unpredictable release of funds (even for salary payments; see Figure l), due to cash constraints, undermining operational planning and resulting in the build-upof arrears. The system is plagued by gate-keeping and deficient accounting and reporting systems, thus leading ~~ to a weak control environment Figure 1.Monthly Salary PaymentsinMOH, 2001 and increasing opportunities 300,000 - for corruption (see Box 2). 250,000 Indeed, in comparative 200,000 perspective, Cambodia's 3- Provincialaverage PEFM system ranks below -: 150,000 100,000 -+- Central-monthly average (as compared to the 50,000 Provincial-monthly low income countries assessed 0 by a joint World Bank-IMF -50,000 diagnostic tool), indicating the Month need for substantial upgrading. ' , X Cambodia: Intearated Fiduciarv Assessment and Public Exvenditure Review Box 2: CorruptioninCambodia:EvidencefromCitizens,Firms,andPublic Officials Corruption has been repeatedly identified as a major constraint on Cambodia's development and a threat to its poverty reduction and growth agenda. A 1999 survey of citizens, firms and public officials, for example, revealed that corruption was identified by many as one of the most serious problems in Cambodia. Citizens living inrural areas identified corruption inthe public sector as the second most serious problem facing Cambodia, after the high cost of living, while urban citizens thought corruption was the single most important problem. Likewise, domestic firms cited corruption as the second most important constraint on business, while 42 percent of foreign firms said corruption was the single most important obstacle for the operation and development of their business. Source: "Cambodia Govemance and Corruption Diagnosis: Evidence from Citizen, Enterprise, and Public Oflcial Surveys," World Bank, 2000. 6. The impact of attempts to address these problems by implementing pilot initiatives-notably the Priority Action Program-has been muted by liquidity constraints andconcernsover the adequacy of control arrangements. It is noteworthy that execution rates for the priority sectors have trailed those for the civil administration as a whole, despite the channeling of a substantial share of sector expenditures through streamlined disbursement mechanisms (see Table 3). In2001, for example, although 92 percent of the budget provision for health was disbursed for the year as a whole, only 2 percent was disbursed in the first quarter of the year, 7 percent in the second, 11percent in the third, and 73 percent in the last quarter. The situation in2002 was largely unchanged, withjust 8 percent inthe allocation disbursed inthe first quarter. The pattern in education, and indeed across Government, i s much the same. These initiatives have, however, improved the alignment of resources with policy objectives and helped channel funds to operational units. If PAP, however, has failed to render significant improvement in budget execution, this is largely due to the fact that problems of cash management have persisted. Table3: The Back-loadingProblem:Percentageof ExpendituresPostedinDecember,2000 and2001 EconomidSector 2000 2001 Central Provincial Central Provincial Civil Recurrent 43.4 22.8 31.6 29.9 Ch. 10:Salaries 11.6 19.4 14.1 23.7 Ch. 11: Operating Costs 38.2 25.0 42.2 34.9 Ch. 13: Specific Program Activities 47.7 100.0 47.3 70.0 Capital Domestic Financing 10.8 18.2 Defense and Security 17.2 18.5 Education 63.2 21.1 34.7 30.2 Health 59.7 31.7 71.4 43.6 Agriculture 41.2 18.6 30.7 28.0 Rural Development 31.5 22.1 13.2 27.4 Source: Bank estimatesbased on M E F TOFE. 7. The challenge for the Government is to consolidaterecent initiativeswhile gradually addressing the structural constraints imposed by weak financial institutions and limited capacity. The Government will need to focus on improving budget execution, cash management, and the control environment. The analysis suggests that a gradual move toward a unified budget execution system i s needed. An important first step will be out posting financial controllers with adequate commitment authority to line ministries. The liquidity constraints that lie at the heart of the problems of poor budget execution can be traced back to the failure of agencies to comply Executive Summary xi with the budget law's requirements in key respects. A program of reform to strengthen cash management will have to focus on increased use of the banking sector, integration of local and foreign exchange operations under a unified Treasury system, improved planning and prioritization incash management, and improved information systems for tracking commitments. Lastly, in spite of a comprehensive legal framework and a multiplicity of controls, weak financial management practices undermine transparency and accountability, posing serious fiduciary risks. For example, anecdotal evidence suggests that "facilitation" of Treasury transactions- specifically the release of budgeted funds-is common. A sound internal and external control environment will require not only an appropriate legal framework but also the development of an enhanced public accounting system, a culture of compliance with the law, adequate ethical and professional standards, and appropriate oversight arrangements. In this context the National Assembly should be strengthened so that it can promote greater accountability regarding the use of public funds. RATIONALIZING PUBLICEXPENDITURE POLICY AND MANAGEMENT 8. Since 1998, the Government has significantly improved the alignment of resources with its developmental objectives by increasing allocations for priority sectors, notably education and health. Government-executed spending on the priority sectors increased from 1.4 percent of GDP in 1998 to 3.2 percent in 2001. In U S dollar terms, education spending has doubled from US$ 27 million in 1998 to US$ 56 million in 2001, with estimates indicating a further increase to US$ 75 million in 2002. Health expenditure has nearly tripled from US$ 11.8 million in 1998 to US$ 34.2 million in 2001, with a further increase to US$ 44.7 million estimated for 2002. Furthermore, the NPRS indicates that the RGC intends to continue this strategy, presenting ambitious targets for growth in priority sector spending to 2005. The reallocation to the priority sectors was financed through increased growth and revenues, and reducedexpenditures indefense and security. 9. However, to implement the SEDP II/NPRS and meet the localized Millennium Development Goals (MDGs) the Government will need to reallocate resources away from non-priority sectors and programs. While further cost refinement i s necessary, the estimate indicates that, in addition to increased revenues, official development assistance, and foreign direct investment, the Government will have to mobilize additional resourcesthrough reallocation (the preliminary cost of the NPRS action plan was estimated at US$ 1.5 billion over the period 2003-2006). Projected growth in revenue collection will not be sufficient to finance the NPRS. Resource scarcity means that expenditure rationalization i s a very high priority. Significant reallocations from low to high priority expenditures will need to occur. There are a number of possible avenues for expenditure rationalization: (a) further reduction of expenditures in the low priority sectors; (b) generation of additional savings from institutional rationalization, including consolidation of administrative structures and retrenchment; and (c) rationalization of operating expenditures by cutting administrativeoverhead while increasing maintenance spending. 10. Moreover, in order to reach the stated poverty reduction goals, it is necessary to improve the effectiveness of spending by linking it more closely to priority outcomes. Increased effectiveness can be attained by improving the pro-poor targeting of resources through more tightly linked sector plans and budgets. Ineducation and health planningprocesses have improved and greater linkages between planningand budgeting have been developed, resulting in improved prioritization of spending. Agriculture and road transport have yet to make progress on this front. Table 4 provides a detailed analysis of the linkages between sector policy, expenditure, and outputs/outcomes. xii Cambodia: IntePrated Fiducian,Assessmentand Public Exvenditure Review ~ Table 4: Linkages between Sector Policy, Expenditure, and Outputs/Outcomes Policy and Planning Expenditure Output/Outcome Health 9 The healthsector has 9 The allocativeefficiencyof 9 Therehavebeensome made importantprogressby spending couldbe increased(the significantachievementsinthe basingits strategyon wage share is too low and sector, especially regarding monitoringoutputs and maintenancespendingi s communicable diseases and outcomes. inadequate). expandedcoverageof the 9 Furtherworkis 9 The share of spendingatthe population. necessaryto cost and provincial level i s too low andthe 9 However,seriousproblems operationalize the strategy high inter-provincialvariance needto be addressed, including by linkingit effectivelyto results ininequalitiesinper capita low andunevenaccess to thebudget. spending across the country. servicesandthe inadequatepro- 9 The resultisinadequate poor incidenceof spending targetingofresources,which 9 Servicequalitycontinuesto reduces the pro-poorimpactof be aconcem. spending. Education 9 The sectorplanning 9 There hasbeenasignificant k Steadyprogresshasbeen process is well-developed increaseinGovemmentfunding made since 1999, inexpanding andeffectivelylinkedto the over the pastfew years. educational opportunitieswith budget andnascent MTEF. 9 The majorshareof spending total enrolment growing. i s allocatedto primary education. 9 Enrolmentgrowthhasbeen 9 The wage share ofrecurrent concentratedamong the poor. spendinghas droppedto Repetitionrateshave also fallen. unsustainably low levels. 9 The focusonprimary educationmeans that sector spending has become increasinglypro-poor. 9 However,netenrolment ratiosandthe completionrateat the primary level are the lowest inthe region. k Servicequality continuesto be aconcem. Road 9 Sectorpolicy goals 9 Totalexpenditureshave k Althoughastart hasbeen Transport requireClarification. increasedsignificantly,however made on reconstructionand 9 Sectorgoalsarenot allocationsfall short of the rehabilitation,the state of the operationalizedina strategy amountsneededto bring the roadnetworkremains poor, and statement,nor are they networkup to amaintainable poor roadquality, which results linked to the annualbudget. state. inhightransport costs, is one of 9 Routineandperiodic the major constraints on growth. maintenanceare not funded on an on-goingbasis. 9 The FRMRcould,asfirst priority, finance routineand periodicmaintenanceof the currently maintainablenetwork, plussomerehabilitation. Agriculture 9 Thoughthe NPRS 9 The sectoris under-funded. 9 Lowlevelsoffundingand makesprogressinsector 9 The shift fromwage the inefficiency of spending planning,the sector still spendingto researchhas reduced constrainbothbroad-based lacksacoherent pro-poor allocativeefficiency. economic growthandpoverty policy frameworklinked to 9 Sector spendingis not reduction. the budget. targetedto pro-poorprogramsand P A lackof sectormonitoring regionaldepartments are makes evaluationof impact resource-deprived. difficult. ~~~ ~ 11. Expenditures are targeted to the poor in education, but less so in health, and only marginally in agriculture. In education a significantly greater amount of public spending benefits the poor in 2002 as compared with 1997 (See Table 5). In 2002, it i s estimated that the poorest 40 percent of the populationwill receive 39 percent of total education expenditure, and 50 Executive Summary xiii ... percent at the primary level. It i s clear that increased subsidies have made education spending more pro-poor. The poorest, however, continue to benefit from only a small share of public subsidies on secondary and post-secondary education, though the poverty incidence of benefits at the lower secondary level also improved. Inhealth, on the other hand, the lowest income quintile consume 13 percent of the resources, while the top quintile uses almost 20 percent (see Figure 2). This disparity is even greater if the relative health need of the poor i s considered to be greater than that of the rich. Inequality in resource distribution i s most evident for the national facilities. By contrast, district hospitals appear to provide the greatest benefit to the poorest group. Table 5: Share of Education Subsidy Received by Figure 2: Health Benefits Incidence Income Quintile Consumption Group (RGC Spending) (RGC Spending) Quinrile: I II 111 IV V % of Public Subsidy Using I997 Enrolment Share Primary 21% 21% 21% 20% 17% 100% LowerSecondary 8% 12% 19% 23% 38% 100% UpperSecondary 4% 4% 14% 21% 56% 100% Post-Secondary 2% 0% 6% 15% 76% 100% Percentageof Government SubsidyReceived, 2002 estimate Total 19% 20% 19% 19% 23% 100% Primary 25% 25% 21% 17% 12% 100% LowerSecondary 11% 14% 19% 22% 34% 100% UpperSecondary 7% 7% 13% 26% 46% 100% Post-Secondary 1% 0% 12% 22% 65% 100% 12. Problems linking policy and expenditure have resulted in significant sectoral differences in the effectiveness of expenditures in improving social welfare outcomes. In education steady progress has been made since 1999, in expanding educational opportunities by growing total enrolment. However, net enrolment ratios and the completion rate at the primary level are the lowest in the region. Moreover, quality continues to be a concem. There have also been some significant achievements in the health sector, including a decline inthe level of some communicable diseases and expansion of physical coverage of the system. However, the sector needs to improve access to services, which remains low and uneven, and rectify the imbalance in the incidence of spending. Inthe roads sector, though a start has been made on reconstruction and rehabilitation, the state of the road network remains poor. Significant increases in maintenance expenditure are required. A critical first step i s to strengthen management of the Fund for the Repair and Maintenance of Roads (FRMR) by enabling it to carry out financial and performance audits of its expenditures and by requiring a formula-based maintenance expenditure program. This would ensure that expenditures from the FRMRfocus on prioritymaintenance programs and are made according to transparent criteria. Inthe agricultural sector the lack of both clear sector policy and output information makes evaluation of impact difficult. Gender analysis in each of the priority sectors also shows that there are systemic barriers to service access for women and girls. 13. Weaknesses in formal and informal accountability mechanisms have a negative impact on the efficiency and effectiveness of public sector spending, and thus on poverty reduction. To the extent that citizens do not have effective means to express their preferences both to politicians and frontline service providers the efficiency of spending may be reduced. To the extent that citizens do not have effective means to monitor the outputs and outcomes produced by the public sector, the effectiveness of spendingmay be reduced. One recent positive xiv Cambodia: Intenrated Fiducian, Assessment and Public Emenditwe Review development in this regard i s the relatively inclusive and transparent process that characterized the NPRS. Scaling up participatory mechanisms that promote inclusion and transparency would likely have significant positive effects on service delivery outcomes, thereby increasing the value for money of eachriel spent. 14. Improving the management of external assistance to ensure greater alignment with RGC priorities is also critical, given the sheer volume of external assistance in total Government expenditure. Reported external assistance flows are significantly higher than the volume of Treasury-executed expenditures (US$ 2,122 million) over the period 1996-2001. From 1996 to 2001, reported flows of external assistancetotalled US$ 2,672 million, of which roughly US$ 190 million has been provided as budgetary and balance of payments support, US$ 220 million as humanitarian assistance, and the remaining US$ 2,264 million through projects and program activities. Although the gap between external and domestic financing i s narrowing, the level of reported external assistance flows in 2001, at US$ 472 million, still exceeded domestic revenues, at US$390 million, by a significant margin. 15. The bulk of this assistance has been channeled off-budget, both in the sense that resource allocations are not reflected in the Government's budget documents and that funds are not disbursed through Treasury. Lack of information and the absence of effective instruments to guide the allocation of external financing seriously undermine the integrity and effectiveness of the budgetary system. At present, it i s impossible to assess the impact of external project financing on overall resource allocations, let alone its implications for future patterns of on-budget expenditures, as regards provision for operation and maintenance costs for new investments. There i s a considerable risk that the proliferation of donor-financed projects-by financing investments that are only tangentially related to ministries' development strategies- has ledto policy drift insome instances. It i s also notable that there has been a gradual shift inthe functional composition of external assistance, though there is no marked trend toward the priority sectors. 16. Improving the effectiveness of development assistance will require action by both Government and donors. There is a growing consensus that development assistance i s most effective when it i s channeled to countries with good policies and effective institutions. Infact, donors, both bilateral and multilateral, are moving toward systems that allocate aid based on country performance characteristics. The World Bank's Country Policy and Institutional Assessment (CPIA) i s used to allocate concessional resources to low income countries, while the U S Government has recently proposed a Millennium Challenge Account (MCA) for countries that "govern justly," "invest in people," and "promote economic freedom." Cambodia could increase the effectiveness of development assistance and possibly increase total assistance by making progress on good governance and improving administrative capacity. On the donor side, greater coordination with Government and among donors i s clearly needed. Donors will also be challenged to rationalize their salary supplementation practices as the Government moves forward with its civil service reform program. At the sectoral level, the solution lies inthe further extension of the sector wide approach adopted in health and education and the move toward budget support once adequate fiduciary safeguards are in place. In the education sector, the creation of coordination mechanisms involving agency management and external partners has helped forge a broad consensus on policy and improve the complementarity of donor inputs. PAY AND EMPLOYMENT REFORMINTHE CAMBODIAN CIVIL SERVICE 17. Given the serious problems afflicting the civil service-low pay, low skills, and thus low capacity-comprehensive civil service reform will have to be accelerated in the short term and carried out over the medium term if the Government's vision of poverty reduction is to become reality. The reformprogram must look at every possible avenue for improving civil Executive Summary xv service pay, to attract and retain skilled staff, especially for high level management and priority sector staff; rationalizing civil service employment, to ensure that human resources are wisely deployed in high priority sectors; and strengthening civil service management, to guarantee that human resource expenditures are subject to controls (viz., on hiring and promotion) and linked more closely to the budget formulation process. Indeed, one of the principal risks to the NPRS i s the capacity of the civil service to deliver. Moreover, it i s clear that low public sector wages provide a breeding ground for corrupt practices. At the same time it is apparent that low pay i s a leading cause of Cambodia's relatively poor standing on public sector performance. The World Bank's CPIA ranks Cambodia in the fourth lowest quintile among fellow low income countries on issues pertaining to public sector management and institutions, indicating significant roomfor improvement. 18. Though the Government has recently made some progress, including an increase in average remunerationin2002, carrying out of a civil service census, and the development of an automated payroll system, much more is required. In some of these areas, such as workforce control, key measures such as the introduction of an establishment register will allow the Government to buildproductively on its previous successes, while in other areas, notably pay and employment policy, the Government will be challenged to push itself farther and faster than currently envisioned. 19. The most pressing issue facing the Cambodian civil service is undoubtedly the low levelof pay for most civil servants, in relationnot only to wage levelsoutside the service, but also to the cost of living (see Tables 6 and 7). Comparing the average wage to per capita GDP finds that a Cambodian civil servant only makes slightly more than the annual per capita GDP. Cambodia's ratio of the average civil service wage to per capita GDP i s one of the lowest in the region. With the exception of a very limited number of highlevel staff (approximately 700 out of nearly 168,000 in 2003), most civil servants earn very little, in either absolute or relative terms. Moreover, the compression ratio (that is, the ratio of average pay of the highest category to the lowest category, which i s under two) i s very low by international standards. This will not be sufficient to attract the necessary quantity of qualified candidates to civil service positions. If the Government cannot attract qualified, experienced highand mid-level managers and professionals to spearhead program design and implementation, the RGC's ability to deliver even basic, good quality services will be impaired. Decompression of salaries coupled with average salary increasesmust thus become the key element of the Government's reformprogram. Table 6: MedianMonthlyRemuneration,2002 (US$') Category Level of Education US$/month Compression Ratio A Secondary school years 40 1.9 B Secondary school ++42 years 32 1.5 C Secondary school 26 1.2 D Other 21 -- Source: CAR. Table7: Private Sector Monthly Wage Data, 1999 Sector US$ Textiles 85.3 Hotels 63.7 Construction 68.6 Source: Cited in Abrahart (2000). xvi Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review 20. There is also room to reduce the overall size of the civil service, which has been growing since the mid-l990s, and to improve its deployment to priority sectors. According to the budget laws, public sector employment has increased by about 13 percent since 1994, from 148,353 to 167,778. The analysis suggests that: (a) reallocationof human resourcesto the priority sectors i s warranted, given the Government's commitment to poverty reduction; (b) some compensated departure is in order both in the non-priority sectors and of lesser skilled workers; and (c) redeployment from center to province is necessary in some sectors including agriculture and health. A functional review, covering the appropriate institutional arrangements, organization, processes, and staffing of government functions in order to improve efficiency and effectiveness in service delivery, would be necessary to ascertain what the overall level of employment ought to be, where exactly the sectoral inefficiencies are, and what composition of public sector employment makes most sense for Cambodia. CONCLUSION 21. The challenges to Cambodia over the medium and long terms, which include both political and administrative obstacles to reform, are great, and the reform agenda is daunting. However, Cambodia has undertaken a number of important reforms over the past several years, and reformers confronting the challenges ahead have a foundation on which to build. Since the current coalition Government assumed power in 1998 Cambodia has made important headway. Tax revenues were increased considerably in a high growth macroeconomic environment. The Government significantly improved the alignment of resources with its developmental objectives by reducing expenditure on defense and security and increasing allocations for priority sectors, notably education and health. In education and health planning processes have improved and greater linkages between planning and budgeting have been developed, resultingin improved prioritization of spending and a greater impact on poverty. The Government has also implemented important budget reforms, undertaken some measures to improve the management of the civil service, and launched a potentially important decentralization program. 22. However, in order to carry this program forward the Government will need to speed up the pace and extend the scope of reform in four key areas. First, Cambodia will have to increase revenue collections to ensure fiscal sustainability and to mobilize sufficient resources to fight poverty. Second, the Government will have to engage in comprehensive public expenditure and financial management reform to reduce the fiduciary risk to public funds. Third, the Government will have to rationalize public expenditure policy and management further by reallocating resources to its highest priorities and improving the efficiency and effectiveness of spending. Last, Cambodia will have need to undertake comprehensive civil service reform- focusing on pay and employment issues-in order to deliver poverty-reducing services. To be sure, some difficult decisions about resource allocation lie ahead, and in order to move forward the Government will have to deal with vested interests and other obstacles to reform. The analysis in this review aims to assist the Government in making those decisions as it goes forward in pursuitof its reformprogram. 23. The priority action matrix provides an indication of the reforms already underway as well as the next steps required to enhance service delivery through improved resource allocation and institutional reform. The matrix indicates the steps the RGC i s already undertaking or has committed to with donor support, in part as a result of the dialogue that developed around the IFAPER (the matrix only addresses priority issues identified in this report and does not cite other on-going reformefforts). These actions are regardedby both the RGC and donors as among the most critical, especially in the areas of public expenditure and financial management and civil service reform. The matrix also provides a set of short and medium term Executive Summary xvii reform options that require continued discussion with Government and other stakeholders. In many of these areas, the RGC is fully aware of the steps it musttake. What i s required i s building a coalition of central oversight agencies, line ministries, and other stakeholders, including donors, to move this important agenda forward. It i s hoped that this report will assist inthat process. I -3 A A P + A A 4 A A I + G A A A A x .3 x 4 A A + + 4 A A 4 A A \ A \ A A 4 A A .rl x x A A A 4 A A A A A 4 A I .-xx .e 4 A A A 4 A A A A 8 U Pv1 4 A 4 A 4 A A .->XX Pj M 35 5 4 A A A 4 A 1. THEMACROECONOMICFRAMEWORKAND MEDIUM-TERMFISCALSUSTAINABILITY: OPPORTUNITIESAND CHALLENGESFORPOVERTY REDUCTION "For the RGC growth is the most powerful weapon in combating poverty and it remains committed to pursuing policies that encourage macroeconomic stability, shift resources to more efficient sectors, and integration within the global economy."2 INTRODUCTION 1.1 Since the restoration of political stability inlate 1998, Cambodia has achieved high economic growth under a conducive macroeconomic environment. Fiscal revenue has improved, stability in the inflation and exchange rates has been maintained, and the international reserve position has strengthened. Progress has also been made on the structural front, particularly regarding the completion of the bank relicensing program, strengthening of tax and customs administration, and drafting of key commercial laws neededto meet the requirements for WTO accession. The Cambodian economy, however, remains vulnerable, and the task to rebuild a country shattered by three decades of civil strife in a fragile governance environment with severe capacity constraints i s da~nting.~ 1.2 Cambodia is well-positioned to ensure macro-fiscal sustainability over the medium term by (1) improving domestic resource mobilization to finance the growth and poverty- reduction agenda, and (2) increasing the level and effectiveness of public expenditures on economic services. In order to promote growth without reducing priority spending Figure 1.1: Cambodia:Contributions to Real GDP on the social sectors, revenues must be Growth, 1998-2002 10 10 greatly increased to finance greater levels of public spending. Indeed, without 8 8 significant increases in revenues, there are 6 6 serious risks to both medium term macro- fiscal sustainability and the RGC's 4 4 medium term poverty reduction program. 2 Greater revenues will need to be generated from tax administration as well as tax 0 policy measures. At the same time to -2I -2 reduce poverty Cambodia should be 1998 1999 2000 2001 2002 targeting spending to the economic sector, which has the potential to foment and facilitate broad-based growth. Indeed, without greater investment in road transport and greater value for money from spending in the transport and agricultural sectors (to reduce transport costs and improve agricultural productivity), the ability of "Cambodia: NationalPovertyReductionStrategy," December2002, p.43. All GDP data in this report are basedon the 2002 NationalAccounts (new National Accounts were releasedin mid- July 2003). 1 2 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review the economy to deliver adequate growth rates over the medium term will be weakened (current levels of expenditures and their effectiveness in these sectors i s analyzed in Chapters 2 and 3, respectively). The level and effectiveness of spending on transportation will also affect the Government's ability to attract foreign-direct in~estment.~ In sum, from the macro-fiscal perspective, Cambodia's medium term sustainability depends on improved domestic resource mobilization tofinance the growth and poverty-reduction agenda and adequate, effective public sector expenditure in the economic service sectors. OVERVIEW OFPERFORMANCE 1.3 From 1999 to 2001, economic growth averaged 7 percent annually (4.5 percent in per capita terms) and inflation was kept below 4 percent. Despite these achievements, growth has been narrowly based, largely relying on the garment sector, which employs only a small fraction of Cambodia's population, and on tourism-related activities that are concentrated in PhnomPenh and SiemReap. 1.4 Reform has proceeded on a number of fronts, both fiscal and macroeconomic. Between 1994-1999 a revampedbudget system was put inplace with donor support. Fundamental economic, fiscal, and financial rules of the game were revised with the Law on Investment in 1994, the Law on Taxation in 1997, and the Law on Corporate Accounting and Auditing in2001. The Government has made progress in establishing an environment attractive to private investment, but further reforms are needed. 1.5 Since 1999 fiscal policy and management has been central to the reform efforts. As recommended by the World Bank's 1999 PER, and in consultation with the IMF, the critical elements of the strategy were to increase revenue to meet expenditure needs (see Annex C for a complete discussion of progress made on domestic revenue mobilization since the 1999 PER).As a result, government revenue strengthened from 8.3 percent of GDP in 1998 to 11.7 percent in 2001. The increase in revenue was mostly due to growth in the tax base. The replacement of the turnover tax and consumption tax on imports with a 10 percent value added tax (VAT) in 1999 also contributed positively to the revenue increase^.^ At the same time, the VAT enhanced the efficiency of the tax systemby simplifying the tax structure, widening the coverage, and reducing cascading.6 1.6 However, the likely contribution from administrative measures was negative (with the exception of 2002), strongly suggesting that administrative capacity, both in terms of systems and human resources, is acting as the binding constraint on efforts to improve revenue collection? The core concern is inadequate number of professional tax and customs collectors and the incentive structure within which they operate. There has been some limited progress in tax and customs administration with assistance from the Technical Cooperation Action Program (TCAP) adopted in May 2001. Efforts focused on ensuring a more efficient use A sources of growth study, which will address issues of FDI,is currentlyunderpreparation. The 1997 Law on Taxation, which beganbeingimplementedin 1999,divided the tax systeminto areal regimeand an estimatedregime. Taxpayers in the real regime are obliged to pay the VAT in lieu of the tumover tax, submit a tax declarationonprofits basedontheir balance sheet andprofit statements,and are subject to the minimumtax. The taxable base of the VAT i s much wider than the tumover and consumption tax, as it covers all incorporated businesses that are registeredunder the "real regime." The only exemptions are: (a) public postal services; (b) hospital and medical services; (c) public transportation; (d) insurance and financial services; (e) imports for personal use exempted from customs duties; (f) non-profit activities in the public interest, and (8) imports of goods related to diplomaticandintemationalorganizations. Revenue from tax administration measures (TA) in time t i s derived as: TRt-e*TR,.lAB- PM; where TR is total revenue collected (actual); e measures the buoyancy of taxes and is assumedto be equal to 1; AB is the change in the proxy of the tax base (Le. GDP for domestictaxes andretainedimportsfor customs); andPMi s the estimatedimpact of policy measures. TheMacroeconomic Framework and Medium-Term Fiscal Sustainabilitv.. . 3 of pre-shipment inspection services, and increasedtransparency to reduce hidden costs incustoms procedures. Moreover, anti-smuggling operations were strengthened through enhanced inter- agency cooperation and the establishment of anti-smuggling units in key border provinces. Improvements in tax administration during 2001-02 included enhancing collaboration among government departments and strengthening tax auditing strategies. As a result of these initial steps, collection of tax arrears has begun. 1.7 In 2002, however, revenue was expected to fall 1.25 percent of GDP below the budget target. The deteriorating performance reflected a drop in trade taxes in spite of recent efforts to strengthen customs administration, an erosion of the tax base arising from clearing the backlog of applications for tax exemptions, and the signing of contracts that reduced transfers to the RGC. In particular, the revenue potential from customs duties, which currently represents 75 percent of total tax revenue, would have been much higher in Figure 1.2: CambodiaTotalTax Revenue(Central Government) the absenceduties large tax and of the customs exemptions. ~ n ~ ~ ~ ~ ~ e % of GDP ~ ~ ~ ; 160 9 Indeed, on average, between 1998 I 4 O 8 and 2001 total import duties paid 120 7 were almost the same as those so 6 exempted, most of which were 6o 5 granted under the 1994 40 4 Investment Law provisions (see u) 3 Table exemptions were for theofgarment 1.1). Many these -*O0 2 industry and reflect the large -40 1 .60 I , , , , , , , , , , quantity of imported material 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 used by the industry. Moreover, total tax exemptions, which can also usefully be considered as tax expenditures, dwarf actual expenditures on economic services, which stood at 5.2 percent of GDP in 2001. In addition, illegal logging activity was curtailed and several forestry concessions were cancelled, reducing collections. Table 1.1:ExemptedImportDutiesby ExemptionRegime (in% of GDP) 1998 1999 2000 2001 TotalExemptedImport Duties '' 4.7 5.2 6.3 7.3 Diplomaticmissionsand international 0.1 0.1 0.1 0.1 organization 3.6 3.9 4.9 5.8 InvestmentLaw provisions 0.6 0.3 0.3 0.3 International aid 0.1 0.2 0.1 0.1 NGOs 0.4 0.8 0.8 0.9 Other Memoitem 4.1 6.2 6.0 6.2 Total Import Duties '' Source: Data provided by the Cambodian authorities and Fund staff estimates. 1.8 With the useof traditional monetaryinstruments constrained by dollarization, fiscal discipline and the restructuring of the banking and payments system were key to the maintenance of monetary stability. Regained public confidence in the banking system led to financial deepening. Money velocity declined gradually, and most of the observed increase in broad money mirrored the accumulation of net foreign assets, with international reserves rising from US$ 390 million at end-1998 to about US$ 640 million at end-2002 (equivalent to 3.25 4 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review months of imports). Still, the use of the banking system for government transactions is very limited, and private credit remains weak. In addition, the spread between the loan and deposit rates continues to be highat around 13-14 percent, making it very costly for local firms to acquire loans (the spread i s due, inpart, to the cash-basednature of the economy, which results inexcess liquidity). 1.9 The RGC's reformefforts have been supported by favorable external developments. Exports grew rapidly from 29 percent of GDP in 1998 to about 40 percent of GDP in 2002, and integration into the regional and world economy also intensified since 1999 when Cambodia joined the ASEAN Free Trade Area (AFI'A). Export growth was driven by the garment sector, which was facilitated by the Most FavoredNation (MFN) status granted by the US in 1996 and helped to reduce the current account deficit to about 10.5 percent of GDP in 2002. Still, despite some advantages (for example, foreign investors do not need national partners, unlike in neighboring Vietnam), Cambodia's competitiveness has not improved. This reflects higher labor costs than in neighboring countries partly attributed to an erosion of competitiveness since the Asian crisis, higher operational expenses due to poor infrastructure, an uncertain business environment due to an inadequate legal and judicial system, and the low quality of public services. MEDIUMTERMMACROECONOMIC FRAMEWORKFISCALSUSTAINABILITY AND 1.10 Over the medium term, the Government's objective is ensure high growth and mobilize sufficient resources to guarantee fiscal sustainability. The Government intends to maintain inflation in the low single digits, while achieving an average GDP growth of 6-7 percent. The overall budget deficit (excluding grants) i s expected to be reduced to about 5.5 percent of GDP and fully financed by external concessional resources. While the growth strategy calls for an increase in government revenue to 14-15 percent of GDP by 2007, the economy would also have to rely on foreign savings of about 3-4 percent of GDP with gross international reservesrepresenting about 3.25 percent of imports of goods and services (see Table A1 inAnnex A). 1.11 Output growth will require rising levels of private savings and investment and diversification of the sources of growth. Indeed, the expansion experienced inthe garment and tourist sectors during 1999-2002 may not be sustainable over the medium term. The agricultural sector has the potential to play a major role through improved output inboth growth and poverty reduction. Ensuringthat spending inthe sector i s adequate (and effective) to boost production i s a high priority for expenditure policy from the growth perspective. Greater access to financial services by local investors and farmers and lower intermediation costs will also be important. 1.12 Foreign direct investment (FDI), an improved investment environment, and continued donor support will be key to achieving GDP growth targets inthe medium term. Inthe shorttermitwillbenecessarytoimproveroadinfrastructuretofacilitate market accessand reduce transportation costs, and to enhance port management and electricity and water connections to reduce production costs. High transport costs on the national road network are a major obstacle to increased growth. It will also be important to improve the civil service and legal and judicial system to improve governance and enhance the environment for private investment. Indeed, many consider corruption a major drag on Cambodia's development (see Box 1.1). Moreover, improving tax administration, reducing the discretionary aspects of the tax system, and adopting international accounting standards would also promote FDI.Thus, in addition to public sector reforms, greater public expenditure and greater effectiveness in spending in the economic sector, principally agriculture and road transport, must be key elements of any growth strategy. The Macroeconomic Framework and Medium-Term Fiscal Sustainability.. . 5 Box 1.1: CorruptioninCambodia: EvidencefromCitizens,Firms,andPublicOfficials Recent empirical work suggests "a strong positive causal relationship from improved governance to better development outcomes (Kaufman, Kraay, and Zoido-Lobath, 1999: 15)." Corruption has been repeatedly identified as a major constraint on Cambodia's development and a threat to its poverty reduction and growth agenda. A 1999 survey of citizens, firms, and public officials, for example, revealed that corruption was identified by many as one of the most serious problems in Cambodia. Citizens living in rural areas identified corruption in the public sector as the second most serious problem facing Cambodia, after the high cost of living, while urban citizens thought corruption was the most important problem. Likewise, domestic firms cited corruption as the second most important constraint on business, while 42 percent of foreign firms said corruption was the single most important obstacle for the operation and development of their businesses. One of the most widespread manifestations of corruption in Cambodia i s unofficial payments, or bribes, to public officials. Illicit payments are often required to get public officials either to do their jobs or to provide special benefits to the payer. Rural households, urban households, and foreign firms reported having to pay bribes for public services in 43 percent, 53 percent 68 percent of their encounters, respectively, with public officials. Moreover, the survey found that bribes are regressive and thus disproportionately affect the poor, further increasing the cost of service delivery and decreasing demand for public services from the poor. Incomparison to the frequency of bribe-paying by firms in other countries, Cambodia ranks inthe bottom quintile, suggesting the gravity of the problem. The literature on corruption argues that its impact on development is manifested through lower economic growth and reduced efficacy o f public services. Given Cambodia's situation, the survey data suggest that Cambodia can illafford the negative economic and administrative externalities associated with corruption, though two-thirds of survey respondents reported in 1999 that corruption was "much worse" as compared with three years ago. Reducing corruption could yield significant economic and poverty-reducing benefits, but a lack of progress on this front only compounds Cambodia's already difficult development challenge. Source: "Cambodia Govemance and Corruption Diagnostic: Evidence from Citizen, Enterprise, and Public OfJicial Surveys," World Bank, 2000. 1.13 Despite significant progress achieved in the last 4 years, Cambodia's fiscal revenue ratios, especially tax revenue, remain among the lowest in the world. To maintain a sustainable fiscal framework, the RGC intends to increase revenuesto about 14.5 percent of GDP by 2007. Such an increase would be the minimumnecessary for meeting the expenditure needs underlying the current development strategy, including higher debt service payments arising from the expected completion of external debt rescheduling agreements with the US and the Russian Federation, though it would not allow for increased expenditures on economic services beyond current plans.* This strategy would enable the RGC to maintain a current budget surplus of about 1 to 1.5 percent of GDP over the medium term, with an overall deficit (excluding grants) declining over time and financed by external concessional resources. 1.14 The anticipated cumulative increase of two percentage points of GDP in fiscal revenue over the medium term is based on containing further erosion of the revenue base and strengthening administrative capacity. Broadening the tax base is expected to focus on tightening and avoiding ad hoc tax and duty exemptions. Inthis regard, a simple and transparent investment regime with lower tax rates (as compared with regional competitors) could be more attractive to potential investors than a system that provides for large exemptions. On the administrative side, efforts are expected to focus on addressing staff integrity, improving procedures and the information system, and strengthening collection enforcement. The automation of customs procedures over the medium term i s expected to improve capacity and to * Interest payments and amortization figures will vary depending on assumptions made on the consolidation period, andor treatment of the pre- and post-cut-off date for Russian debt. Figures are based on Scenario A, presented in: Adachi, M. (2002), "Cambodia-Outstanding External Debt and Rescheduling Scenarios", IMF Staff Report, EBSlO2113. 6 Cambodia: Intearated Fiduciarv Assessment and Public Emenditure Review reduce hidden costs. On tax administration, collection of tax arrears i s expected to be stepped-up through several enforcement measures, including freezing bank accounts. Given the urgent need to increase revenue collections, even beyond current plans, a more comprehensive reform model, such as a semi-autonomous revenue authority, merits serious attention (see Box 1.2). Otherwise, the Government might find itself ina conundrum, given that civil service reformi s necessary for improving tax administration, yet improved revenue performance i s necessary for carrying out civil service reform (see Chapter 5). Moreover, experience in the health sector (see Box 5.3) shows that paying higher salaries and improving management can have powerful impacts on performance, and the decision to grant the National Audit Authority autonomy to develop its own salary scale provides a precedent for autonomization under the right circumstances. The Government should review the possibility of adopting the revenue authority model for Cambodia. Table 1.2: Cambodia: Comparison of Tax Revenue Structure with Other Selected Countries FiscalRevenue Tax Revenue Total Ofwhich Total Direct Indirect Trade Other Per capita Revenue taxes taxes taxes revenue2/ GDP (US$) (Inpercentof GDP:unless otherwisenoted) PRGFAsian countries 16.2 12.8 3.1 7.3 2.3 3.8 427 Bangladesh 8.6 7.1 1.2 5.8 2.0 1.5 370 Lao P.D.R. 11.4 9.2 2.0 3.8 1.o 4.6 310 Nepal 11.2 9.2 2.1 4.1 3.0 2.0 240 Sri Lanka 17.0 14.7 2.4 10.2 2.1 2.3 830 Vietnam 20.6 15.6 5.5 6.3 3.4 5.0 410 Mongolia 3' 28.5 20.8 5.3 13.4 2.1 7.7 400 PRGFSub-Saharan African Countries 17.6 13.7 3.9 5.5 4.6 2.0 323 of which: selected agriculture countries 14.4 11.4 1.8 3.8 3.3 2.1 277 Cameroon 17.8 12.2 3.3 6.5 2.5 0.7 570 Guinea Bissau 19.5 10.8 0.0 0.0 0.0 8.7 160 Mali 14.7 14.0 2.6 5.0 5.4 0.7 210 Niger 8.9 8.3 1.5 1.9 4.4 0.6 170 Tanzania 11.8 10.6 2.7 6.7 1.2 1.2 280 Togo 13.8 12.3 0.4 2.8 6.1 0.8 270 Cambodia 11.7 8.4 1.o 4.6 2.8 3.3 259 Source: 1MF staff country reports. Data refer to average of 1999-2001 Includes nontax and capital revenue 3/ Other revenues includesonly nontax 1.15 Greater transparency over the terms governing the use of state assets will also be necessaryto boost revenues. To improve transparency, all future contracts involving state assets are expected to be reviewed by the MEF and approved by the minister. This measure could also The Macroeconomic Framework and Medium-Term Fiscal Sustainability... 7 help offset future declines in some non-tax revenue items. Indeed, revenue from the auction of garment quotas would end if the quota system i s phased out after 2004, as scheduled; and export licensing requirements could eventually be eliminated. Nonetheless, non-tax collections are expected to be maintained at their current levels over the medium term. Revenue from forestry could increase once a forestry concession system based on sustainable practice i s established (see Box 1.3), and, if the number of foreign visitors continues to increase, tourism-related revenue would rise intandem. Inaddition, revenue from leases i s slated to increase owing to the improved monitoring of the use of state assets and the inventory of states assets underway Box 1.2: RevampingFiscalBureaucracy: The Semi-Autonomous Revenue Authority Model in Developing Countries During the past decade or so diverse countries in urgent fiscal straits have turned to radical reform in order to improve the way their fiscal bureaucracies conduct one of the most pressing national tasks: the collection of taxes. The semi-autonomous revenue authority (ARA) reform, an early version of which originated in the developing world in Bolivia and Ghana in the late 198Os, has now been adopted by more than fifteen countries, including Malaysia, New Zealand, Singapore, Ghana, Kenya, Malawi, Rwanda, South Africa, Tanzania, Uganda, Zambia, Bolivia, Guatemala, Guyana, Mexico, Peru, and Venezuela. The pattern in each of these countries is that traditional line departments (income tax, value-added tax, and sometimes customs) are separated from the ministry of finance and granted the legal status of semi- autonomous authorities. The success of the first wave of reforms, in which the ARA model, bolstered by good tax policy reforms, boosted revenues significantly and improved service delivery, has encouraged imitation. ARAs are defined as tax administrations that have greater than typical autonomy along several organizational design dimensions, including legal character, corporate governance, financing and budgeting, personnel policy, procurement policy, and accountability relationships. Some specific autonomy-enhancing features include self-financing mechanisms, boards of directors with high ranking public and private sector representatives, and sui generis personnel systems authorized to pay the salaries necessary to attract professional accountants and tax lawyers. The board of directors model promises benefits in terms of agency management and accountability. Boards, if properly designed, can play an important role inadvising the revenue authority on key issues such as personnel policy and overall business strategy. Boards can also insulate ARAs from political pressure and can serve as a check on their autonomy. The record of ARAs in Latin America, Asia, and Africa demonstrates that the ARA model permits managers to take greater control over their organizations and manage them more effectively. ARA managers have repeatedly stressed the importance of being able to make decisions more quickly and manage their organizations more responsively as a result of their autonomy. For instance, some ARA managers report that their autonomy has enabled them to pay attractive salaries and take more effective control over personnel management by confidently resisting pressure to make staff decisions based on political calculations. Autonomy has also enhanced the effectiveness of other reforms, such as large taxpayers departments. Source: Taliercio (2003). 1.16 The success of the medium term fiscal strategy thus faces several serious risks. First, a slowdown in the world economy could reduce the number of foreign visitors dampening tourism-related activities and reducing foreign exchange earnings. Second, failure to create a predictable business environment will continue to discourage new investments ina context where the scheduled elimination of the US textile and apparel quotas on all WTO members in 2005 could lead to a sharp drop in Cambodia's exports. Third, inadequate and inefficient spending on agriculture and road transport in the medium term would undermine long term growth. Fourth, the eventual rescheduling of external debt agreements will increase fiscal pressureby resulting in higher debt service obligations, making increased revenues even more important. Failure to increase revenue to adequate levels would thus imply either cuts in government expenditure or recourse to domestic bank financing. Cutting government expenditures would constrain social spending, civil service wages, and non-wage current spending, further exacerbating the low 8 Cambodia: InteRrated Fiduciary Assessment and Public ExDenditure Review quality of public services. Recourse to domestic bank financing could lead to inflation, worsening income inequality, and could weaken confidence in financial institutions, reversing recent financial deepening and exchange rate instability. Box 1.3: FiscalIssues inNaturalResource Management The overall scarcity of resources i s compounded by under-collection and under-reporting of own source revenues (OSRs) in the natural resources sector. Table 1.3 shows that only 52 percent of total estimated revenues, as forecast in the budget law, were collected in 2001. Revenue estimates are also thought to be under-estimated. The fees generated by fishing lot owners (through sub-leasing) may be five times higher than the formal payment made for the lot.' The public revenue performance of Cambodia's forestry sector has also been quite poor. Over the last decade, forest revenues have seldom exceeded US$ 10 million per year and in 2001 were only US$ 7.7 million (in 2002 revenues from forestry concessions were insignificant as operations were suspended). In contrast, in 1996 the World Bank estimated that forestry could sustainably generate as much as US$ 100million annually inroyalty revenues. Given current prices and quantities, however, it i s unlikely that revenues will rise to this level. Table 1.3: MAFF RevenueCollections: EstimatedandActual, 2001-2002 (BnCR) 2001 2002 a/ Estimated Actual % Estimated Actual % Forestry 56,270 28,550 51% 53,113 1,534 3% Fisheries 12,475 5,660 45% 9,887 10,508 106% Others 1,435 2,300 160% 1,345 678 50% Total 70,180 36,510 52% 64,345 12,720 20% Source: MAFF. dProvisional datafor 2002. Several factors drive poor revenue performance. While the RGC has made some limited advances incontrolling illegal logging, there hasbeena substantial, butas yet unquantified, erosioninthe sustainable potential harvest. Beyond this fundamental issue, policy and administrative weaknesses depress the revenue performanceof the sector. The most important is the failure of the physical resource management system to deliver sustainable operations. In2002 and 2003, the failure of concessionaires to present satisfactory forest management plans has meant that logging has been effectively suspended. Facilitation fees, "friendship" money, and other corrupt and illicit payments are encouraged by the absence of tight physical controls and transparency inresource management. These charges can amount to US$30 per cubic meter or more. Deficiencies in revenue management account for a substantial share of revenue shortfall. Under Cambodian law, forestry revenues should be transferred intheir entirety to the National Treasury (NT) after collection by the Department of Forestry and Wildlife (DFW). As revealed by the 2002 MEF audit, "offsetting" transactions have been authorized whereby forest revenues are diverted to both legitimate and fraudulent claims against Government. CONCLUSION 1.17 The challenging economic times ahead mean that the Government should focus on: (a) strengthening its revenue-raising institutions and (b) improving the prospects for long term growth and poverty reduction by investing effectively in the economic service sectors and improving the investment climate. Greater revenues are likely to come from improved tax administration in the short term and improved tax policy in the medium term. At the same time, having recently undertaken significant reallocations to the social sectors, the Government must now turn to improving outcomes in economic services-namely, reducing transport costs and increasing agricultural productivity-to promote growth and poverty-reduction (Chapter 2), and improving the efficiency and effectiveness of spending in those sectors (Chapter 3). Moreover, TheMacroeconomic Framework and Medium-Term Fiscal Sustainabilitv.. . 9 improving the investment climate will require a stronger public sector characterized by less corruption and greater capacity to delivery highquality public services (Chapters 4-6). 2. RESOURCEALLOCATION: TRENDS, OPPORTUNITIES,AND CHALLENGES "Cambodia has therefore rigorously embarked on expenditure rationalization. The thrust of this policy is to implement a `New Social Policy Agenda' for Cambodia. The thrust of this social policy agenda is that growth must take apro-poor path. Having implemented this `New Social Policy Agenda' for Cambodia, the RGC has increased budget allocations for the priority sectors such as education, health, agriculture and rural development, which are thefoundations offuture growth, equity andprogre~s."~ INTRODUCTION 2.1 Since 1998, Government has sought to improve the alignment of resources with its developmental objectives by reducing expenditure on defense and security and increasing allocations for priority sectors, notably education and health. Thus far the impressive shift in the functional and institutional composition of Government spending since the mid-1990s toward the priority social sectors has been achieved by allocating additional resources mobilized by Government and, to a lesser extent, by cutting allocations to lower-priority sectors. The National Poverty Reduction Strategy (NPRS) indicates that the Government intends to continue this strategy, presenting ambitious targets for growth inpriority sector spending to 2005. 2.2 Going forward, however, it is clear that additional resources mobilized will not be sufficient to finance the poverty reduction strategy, which means that reallocations from low priority and inefficient expenditures will be necessary. Additional resources will be needed not only in health and education, but also in other priority sectors. Moreover, if the Government is to allocate greater resources to other critical functions, such as the judiciary and road transport, there is an even stronger case for further cutting base expenditures in low priority sectors. In addition to further reduction of expenditures in the low priority sectors, additional savings can be generated from: (a) institutional rationalization (this i s possible because Government-executed spending on general administrative functions has increased from 1.9 percent of GDP in 1998 to 3.1 percent of GDP in 2001), including possible retrenchment; and (b) rationalization of operating expenditures through reductions in administrative overhead and increasedmaintenance spending. TRENDSINTHEFUNCTIONAL COMPOSITION OF SPENDING 2.3 Broadly defined, the overall level of public expenditure has increased from 23 percent of GDP in 1998 to just over 25 percent in 2001. The broad definition comprises external assistance to projects, humanitarian aid, and Treasury-executed expenditures (see Table 2.1). Reported external financing accounted for 49 percent of total expenditures in 2001. Although this data shouldbe interpreted with caution, owing to shortcomings inthe coverage and classification of the available data on external assistance, it does provide a more accurate picture of the volume of resources consumed in the delivery of services than i s gained from using "Cambodia: NationalPovertyReductionStrategy," December2002, p. 133. 11 12 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review Government expenditures alone." This i s a much higher level of public expenditure than found in the region and other low income countries (see Annex A Table A.2). Obviously, international comparisons are much less favorable when only Government expenditures are considered. Table2.1: Governmentand ExternallyFinancedExpenditureby SummaryFunction,1996 to 2001 (% GDP) Summary Function 1996 1997 1998 1999 2000 2001 Core Government 9.5 9.2 9.4 6.8 8.0 7.6 Economic Services 6.4 4.7 4.5 4.5 4.9 5.2 Agriculture 2.3 0.8 0.6 1.o 1.6 1.5 Transport 2.2 1.6 1.7 1.3 2.0 2.6 Other EconomicServices 1.8 2.1 2.1 2.1 1.2 1.o EnvironmentalProtection 0.1 0.2 0.1 0.1 0.1 0.1 RuralDevelopment 2.5 2.2 2.1 1.9 2.1 2.0 SocialServices 5.2 5.4 6.8 6.9 7.2 8.9 Health 1.9 1.6 2.5 3.1 2.8 2.9 Education 2.1 2.5 2.8 2.4 2.5 3.0 Community & SocialServices 1.3 1.2 1.6 1.3 2.0 3.0 HumanitarianAid & Relief 0.4 0.5 0.3 1.4 2.0 1.2 Other 0.7 0.1 0.2 0.4 0.3 0.3 Total 24.8 22.2 23.4 21.9 24.6 25.3 Of which Treasury-Executed 10.8 10.2 9.3 10.6 11.8 12.8 Source: Bank estimatesbased on CDCDatabaseand MEF TOFE. 2.4 Externalfinancing will remain critical for investment ininfrastructure and capacity building well into the future. However, over the medium term, the resources needed to finance improved service delivery for poverty reduction will have to be mobilized through the Government's budget, primarily from domestic sources. The remainder of this section examines how the Government has mobilized additional resources for priority poverty reduction sectors in recent years and options for reallocationover the mediumterm. 2.5 There has been a significant shift in the functional and institutionalcomposition of Government spending since the mid-l990s, particularly since 1998. This has been achieved by allocating additional resources mobilized by Government and, to a lesser extent, by cutting allocations to lower-priority sectors. Treasury-executed expenditures" increased from 9.3 percent of GDP in 1998 to 12.8 percent in 2001-a 54 percent increase in dollar terms, from US$ 283 million to US$435 million. Comparing institutional allocations in 1998 and 2001, net allocations increased byjust over US$ 140 million, while a further US$ 32 million was mobilized by cutting expenditures in five of the thirty-three agencies (see Table 2.2). Cuts in low priority sectors assume greater importance as a source of financing year-on-year. Nevertheless, it i s clear that increasesinthe budget base have driven expenditure restructuringinrecent years. 2.6 The RGC's strategy for financing expansion of sectoral allocations for priority poverty reduction sectorsis similarly predicatedon the mobilization of additional resources from economic growth and improved fiscal performance. Domestically financed expenditures are programmed to increase from 12.8 percent of GDP in 2001 to 14.3 percent of GDP in 2005. Reallocations are to be made incrementally from additional resources rather than by restructuring institutions' base expenditures. Expenditures in low priority sectors such as defense and security and general administration are programmed to decline as a share of GDP, but will stabilize in nominal terms. This raises two concerns. First,the scope for resource reallocation to the priority loIt should also be noted that there are concerns about the reliability of Government expenditure data due to weaknesses in its financial reporting systems (see Chapter 4). ''Treasury-executed expenditure includes donor assistance channeled through the NationalTreasury. ResourcesAllocations: Trends Omortunities and Challenges... 13 sectors will be limited if programmed improvements in revenue performance and growth rates fail to materialize. Second, even if programmed increases in aggregate expenditure are achieved, the additional resources mobilized will not be sufficient to finance the poverty reduction strategy laid out for priority sectors, particularly in agriculture and rural development where the expenditure base is currently very low, let alone the requirements of critical functions, such as the judiciary and roadtransport, that are given less prominence inthe MTEF. Table 2.2: Financing Reallocation, 1998 to 2001 Indicator 1998-99 1999-00 2000-01 1999 to 2003 N o Institutions 33 33 33 33 N o Institutions with Expenditure Cuts 7 6 3 5 N o Institutions with Expenditure Increases 26 27 30 28 Cuts inInstitutional Allocations $m (30.7) (31.2) (49.1) (32.5) Increases inInstitutional Allocations $m 85.6 78.7 87.9 173.8 Increase inAllocated Budget Base $m 54.9 47.5 38.9 141.3 Priority Sectors Net Increment $m 42.6 9.1 41.7 93.4 Defense and Security Net Increment $m 1.2 (6.4) (8.4) (13.7) Source: WorldBank estimatesbased onMEF TOFE. 2.7 Government's priority development sectors-education, health, agriculture and rural development-have absorbed just over two-thirds of the increase in the base budget from 1998 to 2001. Allocations to the priority sectors have increased substantially-from 15.9 percent of Treasury-executed spending in 1998 (1.4 percent GDP) to 26.5 percent in 2001 (3.2 percent GDP) as recommended by the Bank's 1999 PER (see Table 2.3 and Table A3 in Annex A). Indollar terms, spending on the priority sectors has more than doubledfrom US$ 45 million in 1998to US$ 115million in2001. Table 2.3: Treasury-Executed Spending by Function, 1996-2003 (% of GDP) Function 1996 1997 1998 1999 2000 2001 2002E 2003B Core Government 6.7 6.4 6.5 5.8 7.0 6.5 5.3 5.1 General Administration 1.8 1.6 2.4 2.0 3.4 3.3 2.4 2.4 Defense 3.5 3.3 2.7 2.7 2.4 2.1 1.8 1.7 Security 1.4 1.5 1.3 1.1 1.1 1.1 1.o 0.9 Judiciary 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 Economic Services 1.2 1.4 0.9 1.5 1.5 2.2 1.1 1.1 Agriculture 0.2 0.3 0.2 0.2 0.3 0.5 0.4 0.4 Transport 0.3 0.4 0.2 0.3 0.6 0.8 0.1 0.1 Public Works 0.3 0.3 0.1 0.2 0.6 0.7 0.1 0.1 Other transport 0.1 0.2 0.0 0.0 0.0 0.1 0.0 0.0 Other Economic Services 0.6 0.7 0.5 1.o 0.6 0.8 0.5 0.5 Environmental Protection 0.0 0.0 0.0 0.0 0.0 0.1 0.1 0.1 Area Development 0.1 0.0 0.1 0.1 0.1 0.2 0.1 0.1 Social Services 2.1 2.2 1.8 2.8 2.9 3.6 4.2 4.1 Health 0.5 0.6 0.4 1.o 0.8 1.0 1.2 1.3 Recreation, Culture & Religion 0.1 0.1 0.1 0.1 0.2 0.2 0.2 0.2 Education 1.o 0.9 0.9 1.2 1.3 1.7 2.1 2.0 Social Protection 0.5 0.6 0.4 0.5 0.7 0.7 0.7 0.6 Other 0.7 0.1 0.2 0.4 0.3 0.3 2.3 3.2 Debt 0.7 0.1 0.2 0.3 0.2 0.2 0.2 0.3 Other not classified 0.1 0.1 0.1 2.1 2.8 Total 10.8 10.2 9.3 10.6 11.8 12.8 13.0 13.6 Source: Bank estimates based on CDCDatabase andMEF TOFE. 14 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review 2.8 Education and healthhave benefitedfrom muchof this increase. Education's share of Treasury-executed spending has risen from 9.4 percent in 1998 (0.9 percent of GDP) to 12.9 percent in 2001(1.7 percent of GDP), and health's share from 4.2 percent (0.4 percent of GDP) to 7.9 percent (1 percent of GDP). In dollar terms, education spending has doubled from US$ 27 million in 1998 to US$ 56 million in2001, with estimates indicating a further increase to US$75 million in2002. Health expenditure has nearly tripled from US$ 11.8 million in 1998to US$34.2 million in 2001, with a further increase to US$ 44.7 million in estimates for 2002. Nevertheless, levels of per capita spending are still modest, at US$ 4.07 for health-far below the cost of a minimum package of services-and US$ 6.81 for education in 2002. Significant increases in education and health sector allocations are programmed in the medium term expenditure framework, with their combined share programmed to reach 4.2 percent of GDP by 2005. Even so per capita spending will remainwell below the averagefor low income countries. 2.9 Allocations to the other priority sectors have also increased sharply since 1998:rural development's share of Treasury-executed expenditure tripled from 0.6 percent to 1.8 percent, while agriculture's share has increased from 2 to 3.9 percent. However, the level of govemment spending on both these sectors remains extremely low indollar terms, leaving the implementation of sector programs heavily dependent on donor financing. In 2001, Treasury-executed expenditures on agriculture amounted to just US$ 17 million and rural development US$ 7.8 million. Preliminary estimates presentedinthe NPRS provide for only marginal increasesinthese sectors' allocations for recurrent expenditures from 0.4 percent of GDP in 2001tojust 0.6 percent in2005, suggestinglimitedscopefor increasedlevels andquality of service provision. 2.10 The allocations for other critical functions, such as the judiciary and public works, are also of concern. The judiciary is currently allocated just 0.4 percent of Treasury-executed expenditures, about the same as environmental protection and less than 5 percent of the amount spent on security. Road maintenance i s another area of concern. Recurrent spending on public works amounted to just 1.3 percent of total Treasury-executed expenditure in2001(0.2 percent of GDP), about US$ 5.5 million. While the MTEF provides for modest increases in domestically- financed capital expenditures, much of which are allocated to road rehabilitation, from 1.4 percent of GDP in 2001 to 2 percent in 2005, increases in allocations for recurrent spending are less than 0.8 percent of GDP. This i s clearly insufficient to meet the requirements of road maintenance programs (see Chapter 3). 2.11 Given that demands for increased spending in priority poverty reduction sectors exceed the resources that can be mobilized from incremental budget growth, there is a strong case for cutting base expenditures inthe low priority sectors.Inrecent years additional resources have been mobilized by cutting defense and security expenditures from a peak of 47.3 percent of Treasury-executed expenditures (4.8 percent of GDP) in 1997 to 25 percent in 2001 (3.2 percent of GDP). This entailed a 36 percent reduction of defense spending in dollar terms and a 16percent cut in security spending (see Table 2.4). 2.12 These cuts have been achieved by reducing the size of armed forces and cutting operational expenditures in both defense and security. Defense personnel costs were reduced by one quarter from 1996 to 2001, providing savings of US$ 17 million, through first round demobilizations up to 1998 and a second round starting in 2000. Defense operating costs have been cut even more drastically, by more than half from 1996to 2001, generating savings of US$ 23 million. As a result, the personnel costs as a share of total defense spending have increased from 62 percent in 1996 to 72 percent in 2001. Incontrast, security personnel costs increased by 45 percent over this period, as the security services have expanded, inpart to accommodate some of the demobilized military. Consequently, the burdenof cuts again fell on operating costs, which were cut by half over the six year period. ResourcesAllocations: Trends Omortunities and Challenges.. . 15 Table 2.4: Treasury-ExecutedSpendingby Function,1996-2003 (US$. millions) Function 1996 1997 1998 1999 2000 2001 2002E 2003B Core Government 214.1 200.1 196.5 190.8 233.7 221.8 194.6 194.6 General Administration 58.2 49.7 73.7 66.2 115.2 111.6 89.0 91.2 Defense 111.2 103.4 83.4 88.1 80.1 71.7 67.2 65.5 Security 43.8 46.1 38.8 35.3 36.9 36.9 36.2 35.5 Judiciary 1.o 0.9 0.6 1.2 1.5 1.6 2.3 2.4 Economic Services 37.0 44.6 26.3 49.5 49.8 73.4 39.6 41.5 Agriculture 7.1 8.2 4.9 6.5 9.2 17.1 14.2 14.1 Transport 10.5 12.9 5.0 9.0 20.4 27.9 5.1 5.5 Public Works 8.9 8.2 3.5 7.7 18.8 23.4 3.7 3.7 Other transport 1.6 4.7 1.5 1.4 1.5 4.5 1.3 1.8 Other Economic Services 18.7 23.0 16.0 33.3 19.2 26.3 18.3 19.8 Environmental Protection 0.8 0.5 0.4 0.6 1.0 2.0 1.9 2.1 Rural Development 1.9 1.3 1.6 4.1 3.2 7.8 5.1 5.1 Social Services 66.5 67.2 53.3 92.1 98.6 122.4 151.1 157.0 Health 16.5 18.6 11.8 33.0 26.5 34.2 44.7 49.3 Recreation, Culture & Religion 2.0 2.4 1.8 2.6 7.2 8.4 6.5 7.5 Education 31.0 28.7 26.7 39.9 42.3 56.3 74.9 75.8 Social Protection 17.1 17.5 12.9 16.6 22.5 23.5 25.0 24.3 Other 22.3 3.2 5.3 12.3 11.4 9.7 84.3 122.0 Debt 22.3 3.2 5.3 8.6 8.0 7.7 6.3 12.9 Other not classified - 3.7 3.4 2.0 78.0 109.0 Total 341.9 316.4 283.0 348.8 396.7 435.1 474.6 520.1 Source:MEF TOFERepons and Bank estimates. 2.13 Further reductions in defense and security's share of spending are programmed to 2.6 percent of GDPin the 2003 budget and, inthe NPRS to 2.4 percent of GDP in 2005. In dollar terms, defense and security expenditure i s programmed to increase from US$ 100 in 2003 to US$ 112 million in 2005. An increase inthe share of defense and security budget allocated to operational expenditures i s already programmed in the 2003 budget. Consequently, while the scale and pace of demobilization will be critical in determining the volume of resources that can be mobilized from defense, close attention should also be paid to containment of increases in security services payroll and increases in operational expenditures in both sectors. Moreover, a review of the accuracy of defense personnel databases, which are regarded as inflated by both the Government and donors, would yield additional payroll savings. 2.14 Spending on general administration functions-including legislative and core executive functions, such as planning, finance, foreign affairs and information-has increased rapidly in recent years, from 18 percent of Treasury-executed spending in 1998 (1.9 percent of GDP) to 24.3 percent in 2001 (3.1 percent of GDP), a larger share than education and health combined. Forward estimates of recurrent expenditures presented in the NPRS, using a narrower definition excluding inter-governmental transfers and cross-governmental expenditures, indicate that general administration's share of expenditure will drop sharply from 2.4 percent of GDP in 2001 to 1.5 percent in 2003 and 1.35 percent in 2005. This entails holding expenditures roughly constant in dollar terms. Since there i s pressure to increase spending on legislative and oversight functions, including the newly created National Audit Authority, the burden of efforts to contain expenditures i s likely to fall on executive functions. 2.15 Reduction in administrative overhead offers an important source of savings. As in many countries with coalition governments, Cambodia has a large number of central government institutions with fragmented and sometimes overlapping responsibilities. There are, for example, ten ministries with portfolios ineconomic services and six with social service portfolios. Eightof 16 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review the twenty-four ministerial portfolios had budgets of less than US$ 2 million in 2001, a large proportion of which are allocated to administrative functions rather than the provision of services. The large number of institutions is exacerbated by incremental budgeting, which leads to the dispersal of the resources mobilized from increases in the budget base. From 1998 to 2001, allocations for 28 of the 32 budget entities increased in dollar terms and a similar pattern i s seen in the year-on-year increments. Institutional rationalization would release resources for service delivery inpriority sectors. This would entail the redeployment, and possibly the retrenchment, of staff (see Chapter 5). 2.16 The MTEF demonstrates the Government's commitment to increased expenditure on priority poverty reduction services, yet the scope for resource reallocation is limited using an incremental approach. While spending on education and health i s programmed to increase significantly, increases in allocations for agriculture and rural development from their present low base are modest. Moreover, there i s little scope for increasing allocations to other critical functions, such as the judiciary and road transport. Integration of these sectors into the next round of the MTEF would enable Government to program further increases in resources to support the strategy laid out in the NPRS. At the same time there i s a need to identify opportunities for resource reallocation from the base budget. Further demobilization and capping of increases in defense and security services operational expenditures offer one source. Rationalization of administrative structures offers another. TRENDSINTHEECONOMIC COMPOSITION OF EXPENDITURE 2.17 Changes in the functional composition of expenditures have been accompanied by a restructuring of their economic composition. Whereas the share of salaries and allowances in defense and security has increased significantly since 1996, the share of salaries and allowances in civil administrationexpenditures has fallen. Conversely, the share of spending on operations, special program activities and capital expenditures has increased. This would be considered a positive trend were it not that the current levels of pay are insufficient to support a motivated civil service. 2.18 Spending on civil administration salaries and allowances has increased from 1.3 percent of GDP in 1998 to 1.7 percent in 2001, yet growth trailed increases in total expenditures. Increases in civil administration expenditures on salaries and allowances have trailed growth in the budget base so that the share of total expenditures has fallen from a peak of 27 percent in 1997 to just 18percent in2001. Budget data suggestthat the increase inthe nominal expenditures on salaries and allowances was driven by an increase in staffing levels from 1996 to 2000. Educationhas absorbed much of this increase, its share of civil administration expenditures on wages increasing from48 percent in 1996 to 57 percent in2001. 2.19 Recognizing that low levels of pay are a major constraint to improved public sector performance, the Government recently implemented the first round of its pay reform strategy (average monthly remunerationincreased by 44.3 percent to US$28 inMay 2002). The RGC has previously constrained the total wage bill, including defense and security, to be less than 40 percent of recurrent spending (see Table 2.5). However, while the 40 percent limit does serve the purpose of highlightingthe tradeoffs between levels of staffing and remuneration, it i s artificial. A more strategic approach would address the tradeoffs between base and incremental expenditures on public sector pay and those on operations and capital. 2.20 Civil administration operational expenditures increased by 140 percent from 1998 to 2001, as recommended by the 1999 PER from US$ 58 million to US$ 139 million. The priority sectors have benefited from over half this increase, health and education alone from over 40 percent. By 2001, health and education accounted for 34 percent of total civil administration Resources Allocations: Trends Opportunities and Challenges... 17 operational expenditures, up from 23 percent in 1998. Over the same period, general administration's share dropped from 55 percent to 43 percent. As aresult of this shift inresources to operational expenditures, the ratio of operational expenditures to spending on salaries and allowances has increased significantly across civil administration, from CR 1.42 for each Riel spent on salaries and allowances in 1998 to CR 2.36 in 2001, with the greatest increase seen in the priority sectors (see Annex A, Table A5). The combination of the 40 percent indicative ceiling on the wage bill (as a percentage of recurrent expenditures), combined with the programmed spending shift toward the priority sectors, produced a pincer-like effect on the economic composition of spending, in effect, reducing the share of spending on salaries and allowances significantly from 1998-2002. Table 2.5: Treasury-ExecutedCivil AdministrationExpenditureby EconomicCategory, 1996-2003 @) Budget Chapter 1996 1997 1998 1999 2000 2001 2002E 2003B Sub-Total Current Expenditure 78.1 80.7 79.8 73.0 70.9 77.4 80.4 78.7 10: Salaries and allowances 24.8 27.2 25.4 22.2 19.8 18.1 21.4 19.6 11:Operating costs 32.0 33.5 31.4 31.3 33.2 36.9 31.5 27.1 12: Subsidiesfor provincial admin. 3.4 3.4 2.7 1.7 1.8 1.9 3.5 6.2 13: Specialprogram agreements 0.1 0.2 4.8 0.6 1.0 5.8 9.5 9.2 20: Interest on loans 2.6 1.9 2.4 2.6 2.0 1.7 1.7 1.6 30: Economic transfer payments 3.1 1.3 2.4 1.4 2.9 2.7 2.9 2.0 31: Social transferpayments 10.9 12.5 10.4 10.9 8.4 9.0 7.9 6.3 32: Transfer to international org. 1.2 0.2 0.4 0.6 0.6 0.7 0.6 0.3 40: Miscellaneous - 0.4 1.6 1.2 0.6 1.4 5.6 41: Contingencies 0.6 Sub-total Capital Expenditure 21.9 19.3 20.2 27.0 29.1 22.6 19.6 21.3 50.1: Construction & Equipment 10.6 14.8 13.9 22.5 20.6 15.0 15.0 8.4 50.2: Counterpartfunds 2.0 3.3 5.4 3.2 7.5 4.2 4.1 4.1 50.3: Investmentby foreign funds 2.8 0.5 7.3 52: Financial Operations - 0.0 53: Debt amortization 9.3 1.2 0.9 1.2 0.9 0.7 1.5 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 ~ Source: MEF TOFE reports. 2.21 Analysis of the structure of operating and maintenance expenditures (Chapter 11) in the four priority sectors confirms that administrative overhead tends to predominate when sector allocations for operational expenditures are modest. Operational expenditures channeled through the conventional budget execution mechanism, cover both administrative costs, including office costs, vehicle costs, expenditure on meetings and training, and travel, as well as a range of program expenditures. In the case of rural development, for instance, administrative overhead accounted for 99 percent of Chapter 11 expenditures in 2001. In agriculture, administrative overheads amounted to about 60 percent of Chapter 11expenditures. In education the share of administrative expenditures has dropped steadily as Chapter 11 (and Chapter 13) allocations increased inrecent years, from 32 percent of total operating expenditures in 1999 to 26 percent in 2001. However, some operating expenditures might not be efficiently allocated. For example, of total MOEYS Chapter 11spending in2001, 69.7 percent was spent on sports ("matches and utilities"), and fully 79 percent of the total increase in Chapter 11 (1999- 2001) had been allocated to sports. It may be that in some sectors, as Chapter 13 i s "priority" expenditure, Chapter 11has become "non-priority" spending. In the health sector, on the other hand, administrative overheads have dropped from 35 percent of operating expenditures to 22 percent over the same period. Medicines accounted for 62 percent of health Chapter 11operating expenditures in2001, up from 48 percent in 1999, and medical equipment a further 15 percent. 18 Cambodia: Intearated Fiducian, Assessment and Public Exvenditure Review 2.22 The RGC has sought to increase allocations for capital expenditure by running a current budget surplus.'2Project-related expenditures have increased from just 6.9 percent of Treasury-executed expenditures (0.7 percent of GDP) in 1996 to 19.8 percent (2.3 percent of GDP) in 2000, before dropping back to 16.6 percent (2.1 percent of GDP) in 2001. While there has been some increase in counterpart funds, the bulk of these resources has been allocated to Government projects. In2001, spending on Government projects amounted to US$50 million. 2.23 The structure of spending on government projects is erratic at the sector level, reflecting the impact of a few major construction projects on a relative small government project portfolio. (See Table 2.6) In2001, the economic sectors accounted for 61 percent of total government project expenditures, with public works accounting for about one-third, amounting to about US$ 15 million. Allocations for other priority sectors tend to be much lower (rural development 8.6 percent in 2001, agriculture 7 percent , education 5.3 percent and health 0.2 percent). Despite the modest levels of Treasury-executed expenditure on Government projects in public works, projects still account for about two-thirds of sector spending (63 percent in 2001). This is equally true of the rural development sector, where project expenditures of US$ 4.2 million accounted for about half of sector spending in2001 (See Table 2.7). Table 2.6: Expenditure on Government Projects by Function, 1996-2001 (%) Function 1996 1997 1998 1999 2000 2001 Core Government 48.2 24.2 55.8 14.3 48.3 15.8 General Administration 48.1 2.5 50.5 13.0 48.2 14.9 Defense - 0.9 Security - 21.7 4.8 1.2 Judiciary 0.2 - 0.4 0.1 0.0 Economic Services 43.5 62.5 38.5 50.1 38.9 61.1 Agriculture 4.5 9.3 3.4 0.3 2.3 7.0 Transport 32.8 25.7 10.9 11.1 24.4 35.9 Public Works 29.3 13.7 7.6 10.2 23.5 30.0 Other transport 3.6 12.0 3.3 1.0 0.9 5.8 Other Economic Services 5.9 27.4 24.2 38.6 12.2 17.9 EnvironmentalProtection 0.3 0.1 - 0.0 - 0.3 Rural Development 5.4 1.4 4.3 6.4 2.2 8.6 Social Services 2.9 11.9 1.4 29.3 10.7 14.5 Health 0.7 8.4 0.5 25.6 0.2 0.2 Recreation, Culture & Religion 0.1 0.3 0.2 0.0 5.4 7.6 Education 1.9 1.2 0.4 3.0 1.8 5.5 Social Protection 0.2 2.0 0.3 0.6 3.3 1.3 Total 100.0 100.0 100.0 100.0 100.0 100.0 Source: MEF TOFEReports. 2.24 Maintenance expenditures are inadequate in relation to the current levels of capital spending and the capital stock, giving rise to concerns regarding the sustainability of the investment program. In health, for instance, reported maintenance expenditures in 2001 amounted to just 5.7 percent of cumulative Government project expenditures in the sector over the 1996 to 2001 period. This proportion drops even further when one considers that Govemment l2Treasury-executedcapital expenditure comprises debt amortization, counterpart contributions to extemally financed projects, Govemment-financedconstruction and equipment projects and, from 2001, extemal financing for projects channeled through the Treasury. Counterpart contributions are allocatedas requiredby the MEF and are, for the most part, postedagainstMEF rather than the implementingagency. Similarly, allocationsfor Govemment projectsare held as areserve for allocationby the MEF on the advice of the Office of the PrimeMinister (OPM) duringthe course of the year. Giventhese information constraints,trends inthe structure of capitalspendingmustbeinterpretedwith caution. ResourcesAllocations: Trends Ovvortunities and Challenges... 19 projects account for less than one-tenth of total investment in the health sector, most of which is financed by external partners, and less than half of maintenance spending i s on infrastructure, most being used for equipment and vehicles. Levels of maintenance spending are even lower in other priority sectors: in education, total maintenance expenditures in 2001 amount to just 2.9 percent of cumulative expenditures on government projects for the period 1996 to 2001; inrural development, just 2.5 percent; and in agriculture less than 0.1percent. Inthe absence of sufficient funding for routine maintenance, a large part of the capital spending i s directed at rehabilitation of the infrastructure rather than increasesinthe capital stock. Table 2.7: Government Projects as Share of Total Sector Expenditure, 1996-2001 f%) Function 1996 1997 1998 1999 2000 2001 Core Government 4.3 3.8 6.6 3.8 11.9 3.5 General Administration 15.8 1.6 16.1 10.1 24.2 6.6 Defense - 0.6 Security - 14.8 2.9 1.8 Judiciary 3.2 - 17.4 3.5 1.2 Economic Services 22.4 44.2 34.3 52.1 45.1 41.2 Agriculture 12.3 35.9 16.4 2.2 14.2 20.3 Transport 59.5 62.8 51.4 63.3 69.2 63.5 Public Works 62.6 52.7 51.2 68.1 72.0 63.3 Other transport 42.7 80.1 51.8 36.0 35.2 64.3 Other Economic Services 6.0 37.7 35.3 59.7 36.6 33.6 Environmental Protection 6.4 4.0 - 2.7 - 7.4 RuralDevelopment 54.5 35.8 63.1 19.3 38.9 54.5 Social Services 0.8 5.6 0.6 16.4 6.3 5.9 Health 0.8 14.2 1.0 40.0 0.4 0.2 Recreation, Culture & Religion 0.5 4.2 3.0 0.2 43.4 44.8 Education 1.2 1.3 0.3 3.9 2.5 4.8 Social Protection 0.2 3.6 0.5 2.0 8.4 2.8 Total 6.0 10.1 8.4 15.3 15.0 11.6 Source: MEF TOFEReports. EXTERNALLY-FINANCEDEXPENDITURE 2.25 Reported external assistance flows are significantly higher than the volume of Treasury-executed expenditures (US$ 2,122 million) over the period 1996-2001. From 1996 to 2001, reported flows of external assistance totaled US$ 2,672 million, of which roughly US$ 190million has been provided as budgetary and balance of payments support, US$220 million as humanitarian assistance, and the remaining US$ 2,264 million through projects and program activities. Although the gap between external and domestic financing i s narrowing, the level of reported external assistance flows in 2001, at US$472 million, still exceeded domestic revenues, at US$390million, by a significant margin. 2.26 The bulk of this assistance has been channeled off-budget, both in the sense that resource allocations are not reflected in the Government's budget documents and that funds are not disbursed through Treasury. Consequently, it is not possible to present a consolidated picture of total expenditures from budget documents and budget execution reports. For the purposes of this analysis, data on external financing i s based on data collected directly from donors by the Council for the Development of Cambodia (CDC). However, coverage i s incompleteand data consolidation i s hampered by information gaps and a non-standard functional classification. 20 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review 2.27 There has been a gradual shift in the functional composition of external assistance, though there is no marked trend toward the priority sectors (see Table 2.8, 2.9, and Table A6 in Annex A). The data set developed for the IFAPER allows rough estimates of external financing by function since 1996 and a more comprehensive analysis of externally financed expenditure in 2001.13The share of external financing allocated to core government functions dropped sharply after the 1998 elections. There has also been a gradual reductionin financing for economic development, particularly in communications and more recently energy, though allocations for transport, mostly for road rehabilitation, have increased sharply from US$ 34 million in 1999 to US$ 60 million in 2001. External financing for social development has gradually increased, accounting for over 40 percent of total external financing in2001. However, there is no marked trend indicating a reallocation of external financing toward the Government's long-termpriority sectors. Indeed, financing for health and education has declined from a peak of US$ 120 million in 1998 to US$ 111million in 2001. Most of the increase in external financing of social development has been absorbed by community and social services, including social protection and community amenities such as water supply, which has doubled over the same period to US$69millionin2001. Table 2.8: Functional Structure of ExternalAssistance, 1996 to 2000 (% of Total Spending) Summary Function 1996 1997 1998 1999 2000 2001 Core Government 29.2 30.2 31.1 15.2 12.5 14.7 Economic Development 81.9 69.7 80.7 66.4 69.9 58.8 Agriculture 90.1 68.8 71.7 79.7 82.7 67.4 Transport 85.1 74.2 90.5 79.0 69.8 68.1 Other Economic Services 67.6 64.4 74.7 51.5 53.6 24.4 Environmental Protection 80.5 91.7 90.0 82.8 68.1 33.4 Area/Rural Development 97.7 98.2 97.5 93.4 95.4 88.8 Social Development 59.9 59.6 74.3 59.6 59.5 59.6 Health 72.6 63.3 84.2 68.3 71.9 65.9 Education 52.9 62.7 68.6 50.3 48.9 44.4 Community & Social services 52.2 48.6 69.1 56.3 55.1 68.6 Humanitarian Aid & Relief 100.0 100.0 100.0 100.0 100.0 100.0 Total 56.5 54.1 60.1 51.7 51.9 49.4 Source: Bank staflestimatesbased on CDCdatabase)and MEF TOEF reports. 2.28 While the share of external financing in estimated total expenditures has gradually fallen, key sectors remain heavily dependent on external financing. External financing still accounts for around two-thirds of the sector spending in health, agriculture, rural development, and transport. In the case of rural development, the external assistance portfolio completely dominates the sector. 2.29 Cambodia's external assistance portfolio is heavily geared toward technical assistance.Spending on technical assistance, comprising bothpersonnel charges and training and associatedoperational expenditures i s estimated at US$ 162million in2001, around 45 percent of the external assistance project portfolio, excluding humanitarian assistance, and 21 percent of combined domestic and externally-financed expenditures (see Table 2.10). On this basis spending on technical assistance exceeds the Government's total wage bill and i s nearly three times Government spending on civil administration wages. Technical assistance inputs constitute a l3 CDC data has beenconsolidatedinto a summary functional classification, and adjustmentshave been made to avoid double counting budgetary and balance-of-payments support (included in govemment expenditures) and, for 2001, bilateralfinancing of projectsexecutedthrough multilateral agencies. Dueto these adjustments, totals differ from those giveninrecent CDC reports. Resources Allocations: Trends Opportunities and Challenges.. . 21 particularly large share of total expenditure in the priority sectors, ranging from nearly half of total spending in area development and agriculture, to just over one fifth of total expenditure in education. While technical assistance inputs provide an important contribution in terms of capacity building and institutional reform, over time there i s a need to shift more resources toward investment, and, under the right conditions, toward financing operational costs through budgetary support in health and education. In the medium term, however, there i s a case to be made for creative thinking about the possibility of rechanneling some technical assistancefunds, including salary supplements, toward wage expenditures (see Chapter 5). Table 2.9: TotalExternalAssistanceby SummaryFunction, 1996-2001 (US$, millions) Summary Function 1996 1997 1998 1999 2000 2001 Core Government 88.2 86.7 88.7 34.2 33.5 38.3 Economic Development 167.1 102.7 110.3 97.6 115.5 104.6 Agriculture 64.6 18.0 12.4 25.6 44.1 35.4 Transport 60.2 37.2 47.1 33.9 47.1 59.7 OtherEconomic Services 38.9 41.6 47.3 35.3 22.1 8.5 Environmental Protection 313 5.8 3.5 2.8 2.1 1.0 Area Development 78.1 67.9 63.3 58.1 67.3 61.9 Social Development 99.3 99.1 154.3 136.1 144.6 180.7 Health 43.7 32.0 63.0 70.9 67.7 66.1 Education 34.7 48.3 58.3 40.4 40.5 45.0 Community & Social services 20.8 18.8 33.1 24.7 36.4 69.6 Humanitarian Aid & Relief 12.3 15.8 9.8 47.2 66.9 40.1 Total 444.9 372.2 426.4 373.3 427.9 425.5 Source: Bank estimatesbased on CDCdatabaseand MEF TOFEreports. Table 2.10: Expenditureby Source of FinancingandInput,2001a/ % DomesticFinancing ExternalProjectFinancing Summary Total Salaries Other Capital Total FTC & IPA Total Function Recurrent ITC Core Government 86.3 34.5 43.7 8.1 13.7 11.6 2.0 100.0 EconomicServices 40.0 3.2 16.6 20.3 60.0 19.1 40.8 100.0 Agriculture 32.4 3.5 15.8 13.0 67.6 47.6 20.0 100.0 Transport 31.1 2.0 5.7 23.4 68.9 2.7 66.2 100.0 Other Econ. Serv. 72.7 5.3 42.3 25.0 27.3 14.4 12.9 100.0 Env.Protection 45.7 7.4 34.7 3.6 54.3 50.6 3.7 100.0 Aremural Dev. 12.9 0.6 4.7 7.6 87.1 54.8 32.2 100.0 Social Services 43.1 13.8 26.4 3.4 56.3 23.0 33.3 100.0 Health 38.5 4.0 33.4 1.1 61.5 30.0 31.5 100.0 Education 58.6 34.9 19.4 4.3 41.4 22.9 18.5 100.0 CommunityServ. - 100.0 2.2 97.8 100.0 Other SocialServ. 65.2 3.5 52.7 9.1 34.8 30.5 4.3 100.0 Total 54.5 17.1 28.1 9.2 45.5 20.8 24.7 100.0 Constructedbasedon classijication of individual projects. Source:Bank staff estimatesbased on CDCdatabaseandMEF TOEF reports. 2.30 Externally financed project expenditures at the provinciallevel account for about 16 percent of total expenditures, while domestic financing at the provincial level accounts for about 12 percent of the total (Table 2.11). Decentralized external project financing is thus a particularly important in the ruraVarea development and transport sectors. Externally financed national projects and programs, with activities at both central and provincial levels, comprise a 22 Cambodia: Integrated Fiduciary Assessmentand Public Expenditure Review substantial share of total expenditures inthe priority sectors, accounting for 47 percent of the total in health, 27 percent in agriculture, and 22 percent in education. This tends to point to greater deconcentration of external as compared to Treasury-executed expenditures (see Chapter 6). Table 2.11: Expenditure by Source of Financingand Administrative Level, 2001 (US$,millions) Domestic Financing ExternalFinancing Summary Function Sub- Central Pro- Sub- Central National Phnom Prov- Total vincial total Adm. &Pro- Penh vinces & vincial Policy Core Government 221.8 213.2 8.7 35.1 26.8 7.2 0.3 0.8 Economic Services 73.4 66.2 7.2 109.9 12.6 23.3 5.2 68.8 Agriculture 17.1 13.7 3.4 35.8 5.7 14.5 - 15.6 Transport 27.9 26.4 1.6 61.8 2.8 6.0 1.7 51.2 Other Econ. Serv. 26.3 24.5 1.8 9.9 3.1 2.0 3.5 1.3 Env. Protection 2.0 1.6 0.4 2.3 0.8 0.7 0.8 AreaDevelopment 7.8 6.3 1.5 52.4 1.8 11.3 - 39.4 Social Development 122.4 49.0 73.4 157.9 37.3 74.9 30.5 15.1 Health 34.2 23.5 10.6 54.6 9.3 41.5 2.2 1.6 Education 56.3 15.1 41.2 39.7 12.2 21.3 6.3 Community Services - 46.6 0.5 18.3 26.9 0.8 Other Social Services 31.9 10.4 21.5 17.0 6.2 2.9 1.5 6.5 Total 425.3 334.6 90.7 355.3 78.5 116.6 36.1 124.1 a/Constructedbased on classificationof individual projects. Source: Bank staff estimatesbased on CDCdatabase andMEF TOEFreports. CONCLUSION 2.3 1 Since 1998 the Government has made significant progress in reallocatingresources to its priority sectors, and the NPRS demonstrates the Government's continued commitment. The Government's priority development sectors-education, health, agriculture and rural development-have absorbed just over two-thirds of the increase in the Treasury- executed spending from 1998 to 2001. The MTEF demonstrates the Government's commitment to continue increasing expenditure on priority poverty reduction sectors, yet the scope for reallocationi s limited usingan incrementalapproach. 2.32 Generating additional resourcesfrom savings elsewhere is thus a top priority, given existing commitments to increase allocations further to health and education, and the need to provide additional funding to the economic services sector over the medium to long term-including road transport and, under the right conditions, agriculture and rural development. There i s thus a clear need to identify opportunities for resource reallocation away from low priority sectors and programs. Rationalization of administrative structures, including retrenchment, offers another avenue. Reductions in indirect costs, primarily administrative overhead, would also afford the Government additional resources. In sum, it i s clear that expenditure reallocation-from low priority sectors, programs, and activities-will need to form part of the Government's medium term expenditure plan in order to meet the objectives of the NPRS. 3. POLICIES,EXPENDITURES,AND OUTCOMESIN PRIORITYSECTORS:HEALTH,EDUCATION, AGRICULTUREAND ROADTRANSPORT "Itis now widely recognizedthatpoverty is a multidimensional problem and should be viewed in a broader context than merely in terms of low levels of consumption and income. Lack of food, uncertainties about access to natural resources, powerlessness and hopelessness, social exclusion, lack of education, etc. are all dimensionsof po~erty.'"~ INTRODUCTION 3.1 Cambodia's NPRS provides the road map for moving forward with poverty reduction policies over the medium term. This section examines the links between policy, expenditures, and outputs/outcomes in four priority sectors: health, education, agriculture/rural development, and road trans~ort.'~ The objective of this chapter i s to assess the impact of public expenditure policy and managementon different dimensions of poverty. 3.2 As the NPRS states, "poverty is a multidimensional problem." From agricultural production and roadtransport to education and health, affecting menand women differently along the way, and taking on different manifestations in rural as opposed to urban areas, poverty i s multi-faceted. The NPRS, building on the SEDP 11, represents the Government's bold attempt to reduce poverty in its many dimensions simultaneously, recognizing the inter-related nature of poverty characteristics, for men and women, and for rural and urban families. The effectiveness of resource use depends critically on line agencies, and individual sectors' decisions about policy priorities and sectoral expenditure allocations, which translate into expected outputs and outcomes,all of which determine whether poverty decreases or not, and, if so, how. 3.3 The effectiveness of public expenditure policy and management in delivering services efficiently will depend, fundamentally, on three overarching factors: (a) the nature of poverty, (b) the characteristics of the service delivery productionprocessesineach sector, and (c) the strength of accountability mechanisms. This section briefly addresses each of these overarching issues to set the context for analysis of the impact of sector programs on poverty reduction. The nature of poverty16 3.4 Poverty in Cambodia is widespread, with over one-third of the population living below the poverty line, with significant regional variation. Poverty is much more concentrated in rural areas as compared to urban areas, especially Phnom Penh. The data suggest that the reduction in poverty occurred principally in urban areas, indicating the necessity of delivering services to rural areas if services are going to have an impact on poverty. The data also suggest that progress in reducing poverty, as measured by the headcount index, was slow in the late 1990s, raising questions about the efficiency and effectiveness of public service delivery, though l4"Cambodia: NationalPovertyReductionStrategy," December2002, p. 15. l5Bothoutcomeandoutput indicatorsare utilizedto the extent that they are available. l6 The povertyanalysisi s based onthe NPRS (2002),Chapters 3 and4. 23 24 Cambodia: Integrated Fiduciary Assessmentand Public Expenditure Review trends in social indicators are more positive (see the sections on health and education below)." Poverty i s manifested comparatively in Cambodia's low rank in the Human Development Index-130-just below Myanmar and India, its low rank in the Gender Development Index- the lowest inAsia-and a life expectancy of only 56.4 years. Table 3.1: HeadcountIndices,OverallPovertyLine, 1997 and 1999 Regions 1997 (Adjusted) 1999 (Round2) PhnomPenh 11.1 9.7 Other UrbanAreas 29.9 25.2 RuralAreas 40.1 40.1 Cambodia 36.1 35.9 Source: A Poverty Profile of Cambodia (1997and I999),MOP. 3.5 Poverty in Cambodia is closely associated with agriculture. The overwhelming majority of the poor, 79 percent, are those with heads of households employed in the agricultural sector. Land distribution inCambodia i s inequitable, and the large majority of the poor own small parcels of land. Inthe period 1998-2000 the average growth rate of agricultural employment was only 1.6 percent. 3.6 Lack of opportunities, in the form of access to public services such as health care, education, agricultural extension, and transportation, impedes movement out of poverty. Moreover, the constraints that impede movement out of poverty are intrinsically cross-sectoral. Poor health i s the major cause of impoverishment and other forms of social deprivation (e.g., loss of educational and employment opportunities). The cycle of poverty, illhealth, and high health care costs can economically cripple a typical Cambodian family. An Oxfam study of landlessness found that 40 percent of new landlessness was related to ill health. Poor nutrition reduces children's ability to absorb education, reducing the effectiveness of educational spending. Inturn, poverty rates are higher for households in which the head has had either no formal education or only some primary schooling. Children also lose schooling opportunities because they must support their families with their labor. And access to markets i s impeded by inadequate infrastructure, especially roads, which results inhightransport costs. 3.7 The poverty context in which public services are produced and delivered, briefly assessed here, indicates the rural yet multidimensional nature of poverty in Cambodia. Given its multidimensionality, service delivery interventions, to be effective, would need to address a wide range of issues simultaneously across the priority sectors-education, health, agriculture, and transport-both interms of increasing equityandgrowth. Inter- andintra-sectoral policy formulation is thus the startingpoint for the production of poverty-reducing services, but it i s only the starting point. Implementing those policies efficiently and effectively i s necessary, yet inmany ways muchmorechallenging. The Service Delivery Production Process and Accountability: A Simple Framework 3.8 This chapter proposes a simplified framework for analysis of the service delivery production chain. Ideally, policies, based on agreement about priorities, determine expenditures, which are translated into outputs that in turn generate improved outcomes consistent with policy intentions. A simple productionchain highlightsthe key linkages that, if not functioning properly, could lead to a breakdown in service delivery. Broadly speaking, there are four links.The first i s the development of appropriate sector policy based on government priorities, informed by data, past performance, users' concerns, etc. A lack of clarity or focus in sector policy making can seriously hamper the production of priority services. Linking policy to expenditures i s the next l7Problemswith the validity andreliability of the 1999 data set are discussedinthe NPRS,p. 33. Policies, Expenditures and Outcomes in Priority Sectors: Health, Education, Agriculture.. . 25 step. This requires a strategy, or operational plan, complete with costings of programmatic areas linked to the annual budget and medium term resource planning. This also requires a budget system that facilitates linking programs with expenditures. The link between expenditure allocations and outputs depends on eflciency and egectiveness in organizational production processes. If expenditures are inefficiently allocated, services will be inefficiently and inadequately produced. The lack of a results-oriented management perspective in all but two sectors has made it difficult to link expenditures and outputs, and to monitor progress indoing so (see Box 3.1). Lastly, the link between outputs and social welfare outcomes will determine the extent to which poverty i s actually reduced. This last link i s the most difficult to analyze, in part becauseof the causal complexity of aprocess such as poverty reduction, and given available data, this linkis necessarily probed inapreliminary way. Box 3.1: Results-OrientedPublicExpenditureManagementinCambodia A recent study found that Cambodia has made some progress in developing a result-oriented perspective inpursuitof its poverty-reduction agenda, but has far to go. The study concludes that: & In broad terms Cambodia has already adopted a number of performance targets relative to its poverty reduction goals. & The extent to which ministries have translated performance targets into medium term sectoral > performanceobjectives varies-from considerably to not at all-by sector. While it i s generally the case that most ministries have neither costed nor operationalized targets, thus resulting ininappropriateresource allocations, there are other reasonsthat better explain slow progress inthis regard. & Some sectors inCambodiapursue policies inconsistent with or inimical to statedpovertyreduction > objectives. Management gives little or no guidance to front line provides on the results expected from them. & It is broadly true that Cambodia gives rather superficial political and parliamentary scrutiny to the causes of performancebelow expectation and is reluctant to hold those responsible to account. Cambodia has considerable progress to make in monitoring implementation and outputs of poverty-reducing expenditure programs. Source: Dom, Ensor, and Suy (ODI,2003). 3.9 Each of these links in the service delivery production process is underpinned by accountability relationships between service providers and beneficiaries. One of the reasons for shortcomings in service delivery performance-manifested through ineffective links between policies and outcomes and inefficiencies in expenditure-is often the lack of effective accountability mechanisms. Practitioners and academics are increasingly of the belief that to improve service delivery outcomes, citizens, civil servants, and politicians need to work better together. The central idea is that outcomes will improve when service delivery beneficiaries are able to holdpoliticians and civil servants accountable for results. 3.10 Accountability depends on the strength of the institutions-both formal and informal-that mediate the relationships between: citizens and politicians, politicians and civil servants, and citizens and civil servants." Citizens hold politicians accountable through formal political institutions, including political parties and elections, as well as through informal mechanisms, including NGOs and advocacy campaigns. The relationship between politicians and civil servants depends on both political and institutional arrangements. These include the incentives faced by civil servants and the resources available to them: Are civil servants rewarded for good technical performance or for political reasons (e.g., patronage)? Does public sector remuneration promote professionalism? I s sufficient information available to judge bureaucratic performance?The relationship betweencitizens and civil servants also plays a role indetermining '*Adapted from the WorldDevelopmentReport 2004, forthcoming. 26 Cambodia: Intearated Fiducian,Assessment and Public Exvenditure Review service delivery outcomes. It i s critical for citizens to have mechanisms to express their preferences to frontline service providers and to have the ability to monitor them on a continuous basis. 3.11 Weaknesses in these formal and informal accountability mechanisms in Cambodia have a negative impact on the efficiency and effectiveness of public sector spending, and thus on poverty reduction. To the extent that citizens do not have effective means to express their preferences both to politicians and frontline service providers the efficiency of spending may be reduced. To the extent that citizens do not have effective means to monitor the outputs and outcomes produced by the public sector, the effectiveness of spending may be reduced. One recent positive development in this regard i s the relatively inclusive and transparent process that characterized the NPRS. Scaling up participatory mechanisms that promote inclusion and transparency would likely have significant positive effects on service delivery outcomes, thereby increasing the value for money of each riel spent (see Box 4. on participatory approaches in public expenditure management). 3.12 The approach is to examine the simple model of service delivery in depth ineach of the following sectors: education, health, agriculture/rural development, and road transport (with the exception of the last, all are considered by the RGC as priority sectors). Overall, the chapter finds that the RGC has made progress in improving the efficiency and effectiveness of priority sector spending, especially in health and education. This i s even more significant when considered in the context of the findings of Chapter 2-that the RGC has also made impressive progress in increasing funding for the priority social sectors. The analysis also finds, however, two important weaknesses. The first is that sub-sectoral expenditure allocation could be greatly improved in all the sectors examined. The second i s that the agriculture and road transport sectors are lagging behind, raising serious concerns about the efficiency and effectiveness of expenditures in these sectors. This problem i s complicated by the fact that overall levels of spending on economic services are inadequate for growth and poverty reduction objectives. 3.13 The analysis finds that the development of appropriate policy and the linking of policy to expenditures has been quite uneven across sectors. Ineducation and health planning processes have improved and greater linkages between planning and budgeting have been developed, resulting in improved prioritization of spending. Agriculture and road transport have yet to make progress on this front. Lack of development in lagging sectors i s becoming a binding constraint on better resource use and poverty reduction. 3.14 Moreover, at the sub-sectoral composition of spending level, there is much room to improve overall allocative efficiency: the linkage between expenditures and the production of servicesis weakened by misallocation of resources. Some key trends are apparent: economic composition-the wage share of spending has decreased across the board and i s unsustainably low; spending on maintenance is too low in some sectors (health and road transport); inefficiencies are noted generally in high levels of spending on indirect costs (viz., administrative overhead); geographic allocation-spending on provincial level facilities i s too low in all sectors, relative to central ministry expenditures, given that the provinces are where the majority of the poor reside, which also suggests inefficient spending at the central level inthe form of highindirect administrative costs; and programmatic allocation-expenditures are targeted to the poor ineducation, but less so in health, and only marginally in agriculture. There has also been progress in improving the coverage of services in both education and health, though not in agriculture. Policies, Expenditures and Outcomesin Priority Sectors: Health, Education, Agriculture...27 3.15 Gender analysis in each of the priority sectors also shows that there are systemic barriers to access for women and girls across Cambodia. This suggests an important role for the incorporation of gender analysis at the policy development stage (see Chapter 4 as well). The chapter thus shows that breakdowns in service delivery production occur at all three levels: policy development, linking policies and expenditures, and linking expenditures and outputs/outcomes. Given the weaknesses in policy development and linkages between policies and expenditures, with the exception of education, there i s much room for improvement. Indeed, transforming resources into poverty-reducing services will require substantial reform at the sectoral level, from policy making, to expenditure planning, to service delivery management. 3.16 The analysis also argues for moving beyond a rigid understanding of "priorities" to a more nuanced approach. The concept of the "priority sector" served the RGC well as it undertook a major shift in resource allocation over the past few years. In some sectors, such as health and education, high priority programs have been identified, which naturally implies the existence of lower priority programs. The analysis shows, however, that low priority programs, even in the high priority sectors, continue to receive disproportionate funding. In other sectors, such as agriculture and rural development, the priority concept has not trickled down to the sector program level. At the same time the gender analysis shows that some gender-related programs, which ought to be high priority, are not. And some sectors, such as road transportation and the judiciary, might deserve higher priority status. Thus the analysis suggests a review of the status and definition of priority sectors and a more nuanced approach that identifies priority programs at the sectoral and cross-sectoral level in order to improve the targeting of resources to pro-poor programs. HEALTH 9 Policy and Planning: The health sector has made important progress by basing its strategy on monitoring outputs and outcomes, though further work i s necessary to cost and operationalize the strategy by linkingit effectively to the budget. P Expenditures: Overall government spending has increased significantly over the past few years, though the allocative efficiency of spending could be increased. In economic terms: the wage share i s too low as a proportion of recurrent spending; and maintenance spending i s inadequate, though non-wage recurrent spending has been increasing. The share of spending at the provincial level is too low and the highinter-provincial variance results in inequalities in per capita spending across the country. The result i s inadequate targeting of resources, which reduces the pro-poor impact of spending. > Outputs and Outcomes: There have been some significant achievements in the sector, including a decline in the level of some communicable diseases, and expansion of physical coverage of the system, and piloting of innovative management systems. However, serious problems need to be addressed: low and uneven access to services, and the fact that the incidence of spending i s skewed away from the poor and toward middle income and wealthy groups. Sector policy and planning 3.17 The health sector strategic plan (HSSP) identifies the goals and objectives of the MOH and the wider health sector, putting forward eight key strategies for their achievement. The HSSP, completed in August 2002, was written by a core group of MOH officials, but is based on wide consultations within the government, as well as the NGO and donor communities (see Table 3.2). Closely linkedto the HSSP i s the health section of the NPRS, 28 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review which contains the main elements of the HSSP that affect the poor (e.g., the NPRS places considerable emphasis on the financial and physical barriers to accessing services by the poor). Table 3.2: The Health Sector Strategic Plan,2003-2007 Goal Outcomes Enhance health sector development inorder to 9 Reduced infant mortalityrate improve the health of the people of Cambodia, 9 Reducedchildmortalityrate especially mothers and children, thereby 9 Reducedmaternal mortalityration contributing to poverty alleviation and socio- > Improvednutritional status among children and economic development. women 9 Reduced total fertility rate 9 Reducedhouseholdhealthexpenditure, especially among the poor 9 Moreeffective andefficient health system Strategies Health service delivery Quality improvement 1. Further improve coverage and access to 5. Introduce and develop a culture of quality in health services especially for the poor and other public health, service delivery and their management vulnerable groups through planning the location through the use of Ministry of Health quality of health facilities. standards. 2. Strengthen the delivery o f quality basic Human resource development health services through health centers basedupon 6. Increase the number of midwives through basic minimumpackage of activities. training and strengthen the capacity and skills of 3. Strengthen the delivery of quality care, midwives already trained through continuing especially for obstetric and pediatric care, inall education. hospitals through measures such as the Healthfinancing complementary package of activities. 7. Ensure regular and adequateflow of funds to the Behavioral change health sector especially for service delivery through 4 Change for the better the attitudes of health advocacy to increase resources and strengthening providers sector wide to effectively communicate financial management. with consumers especially regarding the needs of Institutional development the poor through sensitization and building 8. Organizational and management reformo f interpersonal communication skills structures, systems and procedures inthe Ministry of Health to respond effectively to change. 3.18 The central innovation of the HSSP is to supplant a strategy based on coverage by facilities with one based on outputs and outcomes. The indicators proposed for monitoring each strategy are intendedto be basedlargely on existing monitoring systems, and include a range of both general activity indicators, such as contacts per capita, as well as more specific indicators. Some indicators are broken down by income group, and the financing indicators emphasize the proportion of the poor given exemptions from user charges. Targets for achievement have also been developed for each strategy. Moreover, the HSSP forms the basis for the sector wide management (SWiM) approach. 3.19 Further work needs to be undertaken, however, to cost and operationalize the HSSP. First,M O Hneeds to link the strategies inthe planto the program areas used inthe MTEF (these are based mainly on budget management centers, or BMCs). Both indicators and costs must be related to sector programs. Second, operational plans for each program area need to be developed, based on the new planning manual. This i s a substantial and challenging undertaking since it requires that program area managers identify key activities, costs, and performance indicators. Third, over the medium to long term, the central budget process will have to be revised so that it supports the disbursement and monitoring requirements of a program budget approach (PAP, discussed in Chapter 4, has given MOH some experience in program-based Policies. ExDenditures and Outcomes in Prioritv Sectors: Health, Education, Agriculture.. . 29 budgeting). The HSSP i s a major step forward in that it provides a strategic framework for the sector and underpins boththe nascentMTEFand the SWiM. Health sector expenditures 3.20 The health sector is financed from three main sources: government revenue, donors and private, mainly out-of-pocket, contributions. There i s a significant degree of uncertainty about the total level of fundingderived from each source.19The majority of health care spending i s made through out-of-pocket payments in public and private sector facilities. Most of the household surveys indicate a figure of between 75 and 85 percent (US$ 20-30 per capita) of total funding(see Figure3.1). Figure 3.1: Health Sector Fundingby Source of Finance is86 (HDS) 1987 (SES) lSSS (SES) 2000 (DHS) 3.21 Total Government funding increasedfrom CR 26.7 billion in1995 to a budgeted CR 147.7 billion in the 2002 budget, a 450 percent increase (280 percent in real terms). This implies a threefold increase in dollars per capita, from just over one to a budget of around three dollars in2002 (see Table 3.3). 3.22 There are three key concerns from an economic perspective: (1) allocations for maintenance are inadequate; (2) the share of wages to recurrent spending has decreased significantly; and (3) the share of operating costs (excluding drugs and supplies) has risen consistently (see Table 3.3). The problem on the capital-recurrent side i s that the capital and recurrent funding processes are totally separate. Capital spending i s treated as a "free-good" in the sense that no obligation i s placed on MOHto maintain and eventually replace items. In2001 salaries accounted for about 10 percent of recurrent spending, one of the lowest shares in the world. By comparison, in Vietnam the share i s 23 percent, while in Bangladesh the share i s 45 percent (World Bank, 2001, HEU and MAU, 2001). At the same time the rising share of (non- drug) operating costs raises concerns about the efficiency of spending. l9 Sources of information on private spending: HDS - Household Demand Survey (1996), CSES - Socio-economic survey (1997 & 1999),andDHS -Demographic and Health Survey (2000). 9 e4 e4 8 ?c!r7l'c!9c! m o m m o o o m 9 N m N m g 3 9 Q e4 9 ch ch 3 9 00 ch 3 Reforming Institutions to Improve ServiceDelivery I.. . 31 3.23 The allocation of spending, as proportion of total government spending, to the provincial level was only 56.7 percent in 2001.20The remaining allocation to the central level (43.3 percent) was divided between national programs and institutions (19.6 percent), national hospitals (13.7 percent), and the central ministry (10 percent).21 Though the percentage allocated to the central ministry itself is on the low side, the amount allocated to national programs, institutions, and hospitals, at 33.3 percent of total spending, seems somewhat high (especially in light o fthese facilities' better offclientele). 3.24 External sources, including development banks, multilateral and bilateral agencies, and international NGOs, provide around two-thirds of public funding. According to CDC data external funding for the health sector peaked in 1999 at around US$ 70 million. In 2001 it was US$ 66 million, o f which 46 percent was provided by bilateral donors (the US, France, and Japan account for 72 percent ), 22 percent through ADB and World Bank loans and credits, 10 percent by the UN and EU, and the remaining 22 percent by international NGOs. The main concern i s to what extent external financing is consistent with the Government's own health policy goals. For example, in 2001 fully 31.8 percent o f external financing was allocated to national programs and hospitals, and over 12 percent to national hospitals alone, though the HSSP gives priority to locally provided basic services. This i s not to suggest that all funding should necessarily be allocated to provincial facilities, but that provincial populations should be helped to access specialist facilities when required. 3.25 The geographical allocation of resources presents a serious concern about the poverty-targeting of resources. Figure 3.2 shows that per capita allocations in fact vary seven- fold across the country, largely the result o f a resource allocation formula based on capacity and number o f facilities. Figure3.2: ProvincialResourceAllocation, RGCActual Health Care Spending (CR per capita, 2001) 5867 6490 8030 9148 21620 3.26 I n 2001, minimum expenditure per capita as a proportion of the maximum was 12 percent, and this had fallen from 19 percent in 1998. Similarly, the overall coefficient of variation has increased. In per capita terms, therefore, inequalities have increased. When donor Expenditures can be divided into allocations to provincial facilities and funding to national level facilities and programs. This allocation cannot be directly derived from the expenditure book, since the budget for medicines, which i s allocated through Central Medical Stores (CMS), i s shown as a lump sum allocation under MOH. In 2001 there were 11 national hospitals and 12 national programs and institutions that received a budget from government. 32 CambodiaIntegratedFiduciaiy Assessment Public Expenditure Review (DHS, 2000) and donor allocations, 25000 - indicating some positive targeting o f R ' = W provinces with poorer health status 2oo oo (Figure 3.3). It should be observed, d A ,5000 . * however, that only around 25 percent o f Z ' 0- . . : donor spending could be directly ioooo - attributed to individual provinces from existing records so the descriptive 50 00 conclusions can only be tentative. 1998 1999 2001 Central level 2,238 2,641 3,396 Provincial level 1,206 1,071 1,932 Total 3,444 3,712 5,327 Percentgovemment spending 7.6% 3.O% 4.5% Percent ofprivate expenditure 0.6% 0.9% Percentprivate spendinginpublic facilities 2.8% Source: Health Economics TaskForce, Department of Planning & Information, MOH 3.29 The regulations regarding user fee expenditure allow facilities to retain 99 percent of collections (49 percent is allocated to staff remuneration and 50 percent for operational use). The remaining 1 percent i s deposited in the Treasury, which i s supposed to ensure that a report o f fees i s received at the central level. User charges are appreciated by local authorities who view them as an important way o f ensuring supplies and attendance o f health staff. Facilities generally report that user fees have led to a general increase in activity.22 Of considerable 22 InTakeo, for example, outpatient consultations fell although specialized consultations increased (De Loof, Bonnet et al. 2001). Inpatient admissions also increased. In Sotnikum, where there is strong appreciation by patients for Reforming Institutions to Improve Service Delivery I... 33 importance is that a previously unofficial practice is made official and regular. Price lists posted in facilities increase transparency and reduce the need for patients to shop around for supplies needed for their treatment. 3.30 The concern is that, unless adequate provision is made, user fees impact disproportionatelyon the poor and undermine their ability to access care. The concerns are not principally about whether charges should exist, since the alternative seems to be an unregulated unofficial system (though it should be noted that many public sector employees are involved in the private provision o f health services, which results, in part, from the civil service pay problems discussed in Chapter 5). Inhealth centers the very modest charges and willingness to grant exemptions may minimize the negative impact o f charges. Even in areas with exemption mechanisms, however, there is evidence that user charges cause destitution. Poor people are often forced to sell possessions or even land to pay for public health care, particularly in rural areas (Wilkinson, Holloway et al. 2001). This finding i s supported by the DHS which found that more than 26 percent o f patients paying for health care obtained the money by borrowing or selling assets. 3.31 Health financing schemes must make provision for the poorest, by establishing equity funds, but the precise design is left up to the implementing locality and actual practice varies widely (Bitran, Espinosa, et al. 2002). In Takeo, for example, an asset questionnaire, that identifies key factors that differentiate the poor from the non-poor, i s used.23 Exemptions are financed from user fee income and a maximum level o f exemptions required to sustain the scheme are computed to ensure affordability. Provision for exemptions range from 10 to 22 percent in the evaluations recorded. Given that the poverty rate, according to the SES 1997, exceeds 36 percent, this provision is probably insufficient even in the more generous schemes (Bitran, Espinosa et al. 2002). 3.32 A second concern with user charges is that they may distort the prioritiesof health staff toward healthcare that makes money rather thanwhat is most effective for patients. It i s unclear from the evaluations whether user charges are leading to a reduced emphasis on prevention. It could certainly be argued, however, that unofficial charges also distort priorities, and ifofficial systems are established correctly then the distortion can be lessened. Health sector outputs and outcomes 3.33 During the last five years there have been important achievements in the health sector.Polio has been eradicated and the levels o f other communicable diseases, while still high, have shown significant decline. The number o f malaria-related deaths has fallen while the numbers o f people (15 to 49) living with HIV has declined consistently for the last three years from a peak o f 4.04 percent to the current estimate o f 2.8 percent (see Table 3.5). This downward trend may be at least partly due to the significant increase in condom use achieved within high risk groups. Rates o f condom use among sex worker, for example, now exceed 90 percent, up from only 16 percent in 1996. 3.34 Physical coverage of the populationby the public health system has also expanded. Delivery care at health centers or hospitals has increasedbut is still below 10 percent of expected deliveries (see Annex A Table A13). At the same time MOH experimented with different modalities o f service delivery and its instructive experience has gained international attention (see Box 3.2). ~ ~~ guaranteed opening hours and drugs supplies, utilization shows no decline and in some cases a positive trend is reported (Van Damme and Meessen 2001). 23 Includes house roof of grass or leaves, no furniture, no television, no transport, fewer than 3 farm animals. 34 CambodiaIntearated Fiduciarv Assessment Public Exaenditure Review Table 3.5: Trends in Selected Utilization, Access and Health Status Indicators Provincial 1997 1998 1999 2000 2001 variation ~ ~~ Number o f contacts per year (public facilities) 0.33 0.31 0.29 0.31 0.35 0.09 to 0.9 Bed occupancy (not all facilities) 45% 45% 47% 51% 51% 9.8Yoto 122% Prenatal coverage 33% 30% 36% 41% 43.0% 5.6% to 65.2% Delivery coverage (health centerhospital as percent o f expected deliveries) 6.9% 5.7% 7.7% 9.2% 8.6% 0.01% to 9.2% Measles coverage 68% 63% 63% 69% 72% 26.6% to 89.9% Malaria - cases per 100,000 population 1,549 1,246 1,199 1,104 1,114 Malaria case fatality rate 0.51% 0.44% 0.64% 0.47% 0.19% Adult HIV prevalence 4.55% 3.46% 3.2% 2.8% Source: National Health Statistics Report, Ministry of Health, 2001; Report of National Center for Parisitolology Entomology and Malaria Control, 2001. 3.35 Despite these achievements, progress in other areas has lagged, and health conditions in Cambodia continue to be poor relative to other Asian countries. Infant mortality has fallen to 86 in2001 from95 in 1997, slightly lower than inLao PDRbutthree times the rate in Vietnam. One of the reasons is that access to services remains low and uneven, as considerable provincial variations are apparent. The number of contacts with public facilities varies from 0.06 (Kok Kong) to 0.79 (Mondulkiri).Based on MOH data for 2002, only about 54 percent of the population lives within 10 kilometers of basic public health services Box 3.2: The Cambodian Experience with Contracting for ServiceDelivery Some attempt to harness the management capacity and incentives present in the non-government sector was made through a project supported by ADB in 1998 (Keller and Schwartz 2001). This project envisaged two approaches to contracting: (a) a contracting-out approach (in 2 districts), in which the management o f all district health services, including procurement, equipment purchase and staff employment, was contracted to a non-government agency; and (b) a contracting-in approach (in3 districts), in which non-government contractors were hired to provide management expertise while responsibility over management was kept with civil servants. Contractors were given a budget to provide a full range o f services. In addition, four similar "control" areas were selected to provide a basis for on-going comparison with the intervention areas. In order to provide a fair basis for comparison, budgets in control areas were also boosted to a comparable level. An evaluation was undertaken in 2001at a relatively early stage in implementation. A baseline survey (conducted prior to implementation) was compared with indicators from a post-implementation survey. The evaluation found that while the cost per capita in districts that contracted out and, to a lesser extent, that contracted in, was considerably higher than in control areas, total costs, including those incurred by patients, were reduced in the contracting areas (see also B o x 5.1). In addition, activities as measured by, for example, total number o f patients, number o f immunizations and attended deliveries, were found to be considerably higher. There was, however, also much variation between districts with similar interventions. So, for example, while the average increase inthe number o f facility deliveries was higher in contracting-in and -out districts, in two out o f five contracting districts there was no significant difference (while one o f the control areas registered the third highest increase). A key finding o f the evaluation, however, was that in contracting districts health care spending by the poorest declined suggesting that the improvement in access helped to reduce bothphysical andfinancial barriers to obtaining health services. The experience is o f considerable interest in assessing future scope for non-bureaucratic and non- govemment methods o f contracting. Some more work may be required in assessing what features contributed to the greater success o f contracting in certain districts. An element o f self-selection is inevitable, however carefully districts are chosen, where districts that have structures that are initially stronger than others are chosen to participate. Extending the experience requires both greater resources and a greater appreciation o f the pre-requisites that are required for the contracting to be successful. Source: "Contracting and Similar Experiences in Health in Cambodia," Peter Feenstra/Medicam, December 2001. Reforming Institutions to Improve Service Delivery I... 35 3.36 The 1996 health coverage plan sought to concentrate funding on two levels: health centers and district referral hospitals, which are based in 67 operationaldistricts (OD). The plan envisaged strengthening the current system o f health services so that all health centers could provide the minimum package o f activities (MPA) while all district hospitals would provide a complementary package o f activities (CPA).24 However, only a minority o f hospitals are able to provide the CPA. According to the 2002 budget book, only 40 percent o f 48 provincial and district hospitals were able to provide the CPA while 82 percent o f 991 health centers were able to provide the MPA. Coverage o f the minimumstandard is thus incomplete. 3.37 The major outcome-related concern is that the incidence of benefits is skewed away from the poor and toward middle income and wealthy groups. It is clear from Table 3.6, for example, that use o f services, as reflected in immunization rates, trained delivery, and medical attention for disease, is higher for rich rural and urban residents compared to the rural poor by a wide margin. Given that many diseases such as ARI and diarrhea are likely to impact more heavily on the poor, the actual difference in use relative to need i s even larger. Likewise, health status indicators are much worse for the poorest groups. Infant mortality, for example, o f the poorest group i s more than double that o f the richest. Table 3.6: Use of Health Facilities, Knowledgeaiof Health and Health Status by Socio- Economic Status Rural Rural Urban Urban Poorest poorest rich poor rich women Immunization -measles 45% 85% 78% 41% Diarrhea - ORT use 11% 44% 41% 12% AFU seen medically - 3 1% 52% 53% 25% Severely underweight 16% 6% 6% 18% Antenatal visits -trained person 14% 56 Yo 71% Delivery - attendedbymedicallytrained 14% 75% 88% Use of modern contraceptive 10% 22% 24% Knowledge of HIViAIDS 59% 89% 93% Infantmortality rate 108.70 58.30 (-1 41.10 94.10 Source: Adaptedfrom DHS, 2000, Macro International by Livia Montana (..)Indicates large sampling error due to small number of cases, 3.38 The analysis indicatesthat whereas the lowest income quintile consumes 13 percent of the resources, the top quintile uses almost 20 percent (Figure 3.3). Put differently, the bottom 40 percent o f the population consumes only about 30 percent o f public health spending. This disparity is even greater if the relative health need o f the poor is considered to be greater than that o f the rich. Inequality in resource distribution is most evident for the national facilities, where relatively wealthy urban Cambodians make most use o f these services. By contrast, district hospitals appear to provide the greatest benefit to the poorest group. 3.39 These results, however, should be placed in context. Although they indicate some pro- rich (or at least pro-middle income) bias, this has been a common finding in a number o f developing countries. In fact the concentration index (which indicates the overall bias to the rich or poor in distribution o f benefits) indicates that the distribution in Cambodia i s less pro-rich than 24 The MPA comprises most essential first level services including initial consultation and primary diagnosis, emergency first aid, chronic disease care, routine child health, maternal care including normal delivery and birth- spacing advice, health education and onward referral. The CPA comprises consultation and advanced diagnosis for referred patients, medical and surgical emergencies, hospitalization and dentistry (Ministry of Health, Health Center Manual, Phnom Penh, 1997). 36 CambodiaIntegratedFiduciaw AssessmentPublic Exuenditure Review that found in Vietnam, Bulgaria, and Ghana, and similar to the distribution in Bangladesh, but more pro-rich than in Sri Lanka (Demery 2000).25 Figure3.3: BenefitsIncidenceConsumptionGroup (RGC Spending) 300% 7 0National hospital 25 OP 20 0% 10 0 1 5 0% 0 0% Source: Computedfrom Governmentdata and CSESdata on utilization(nationa1Institute of Statistics,1999) 3.40 Resources appear to be more or less equally distributed between men and women (Figure 3.4). The small difference almost exactly mirrors the differences in the relative size o f each group. It is, however, usual for the health needs o f women to be greater than those o f men. If reproductive health care, which is predominantly used by women, i s subtracted from utilization and spending it is likely that the benefits incidence would be shown to favor men. Repeatingthe analysis for men and women outside the 15-44 age group, benefits favor men over women (26 to 24 percent o f spending). N o difference between men and women o f different income groups was found (see Box 3.3 for a discussion o f barriers to access for women). Figure 3.4: Gender BenefitsIncidence of Government Spending Male Female Source: Computedfrom Governmentdataand CSESdata on utilization (national Institute of Statistics 1999). 3.41 This analysis reinforces the need not only to improve health indicators on average, by improving the overall management of the sector, but also to target interventionsto the 25 Defined as the area between the benefits concentration curve and the 45 degree line of equality. A positive index indicates a pro-rich bias whereas a negative index indicates a pro-poor bias. The figures for Vietnam, Bulgaria, Ghana, Bangladeshand Sri Lanka are respectively 0.16, 0.126, 0.192, 0.073, and -0.006. The comparable figure in Cambodia i s 0.074. Reforming Institutions to Improve Service Delivery I... 37 poorest groups where low cost interventionscould have mostimpact.Fullyunderstandingthe reasons for this distribution requires an analysis o f the multiple determinants o f health sector use, which remains an important areas for future investigation. A preliminary multivariate investigation, based on the 1999 data, suggests that location o f facility i s extremely important both in determining how long people wait before consulting a practitioner in the event o f illness and in determining which facility i s chosen for treatment. It is likely, therefore, that choice o f health sector or hospital is strongly determined by population concentration. Box 3.3: Women and Service Provision I:Barriers to Access in Health Though MOH's MPA and CPA both have a strong focus on the health care needs of mothers and children, according to the Cambodia Demographic and Health survey, more than 90 percent of women reported having one or more problems inaccessinghealth care for themselves.The most common problem identified was inability to pay for health care (88 percent of women surveyed). Distance to health facility and difficulty with transportation were each identified by more than 40 percent of women surveyed. As in the case of education, international evidence shows that utilization of health care services by girls and women i s more price sensitivethan utilization by boys and men.' For this reason, changes in out-of-pocket expenses, through user fees or health equity funds, are likely to have different effects on women than on men. Problems in access result in serious health problems for women and children and have a major impact on the capacity of the health system to deliver services effectively. Some of the more serious problems underline the importance of ensuring access: Iodine deficiency, causing miscarriage, premature birth, and increased child mortality, is high. Sixty-six percent of pregnant women and 87 percent of children under one year are anemic.Nationwide, only 2.4 percent of women had iron supplementationfrom any source during pregnancy. Only 38 percent of pregnanciesreceived antenatal care fiom trained health personnel. Only 10 percent of births take place inhealth facilities. Thirty-two percent of birthsare assisted by trained medical staff (midwife, nurse or doctor). Most commonly, births are assisted by Traditional BirthAttendants(66 percentofbirths). Another important problem, access to contraceptive methods, i s an important determinant of the HIViAIDS epidemic. There is a large unmet demand for contraceptivemethods. 57 percentof women wish to use contraception, but only 24 percent of women currently using contraceptive methods. Of these, only 19 percent are using modern methods of contraception. This takes on increasing importance as the epidemic moves increasingly beyond sex workers andtheir clients and into the generalpopulation. The MOHprovincial health care system (health clinics and referral hospitals) is a minor players in HIViAIDS activities, accounting for only 10 percent of spending on HIV/AIDS and doing very little of the condom distribution (less than 5 percent). Going forward, there would seem to be considerable scope for promotion of condoms as a means of contraception.Emphasison condoms for contraception may be useful within the context of marriage and other long term relationships as a way to avoid directly discussing HIViAIDS transmission with its implication of distrust and infidelity. To date, there has been little movement inthis direction, however. The principalconstraint to improved service provision inthe area of HIViAIDS will not be budget but rather will be the ability ofNCHADS andthe provincialhealth systems to work together and for each to change their approach: NCHADS will have to develop messages relevant to married couples and the MOH provincial health system will have to be willing to emphasize condoms, rather than other means of contraception. 3.42 One possible response is to craft a pro-poor strategy that requires resources to be channeled to modes of service delivery used relatively more by the poor. This suggests greater emphasis o f funding on peripheral rather than national level facilities. This appears to be envisaged by the HSSP and community delivery in particular could help to reach some o f the groups currently not accessing the health system. The incidence analysis suggests that the poor utilize health centers and clinics less than middle income groups, thus monitoring the effect o f improved targeting will be particularly important. It is, however, unlikely to be efficient to attempt to provide all services through operational districts or health centers. 38 CambodiaIntegrated Fiduciary Assessment Public ExpenditureReview 3.43 A second, complementary,approachwould be to help the poor obtainessential care when necessary inlargerfacilities inthe provincialor nationalcapital.Demand side funding, targeted not only at reducing the burden o f facility charges on the poor, but also o f other expenses, such as transport, may have a role to play inreducing the barriers to accessing services. This is likely to be a particularly important issue in more remote provinces with less developed transport networks. There may be a role here for targeting equity funds at these areas in particular and permitting their use not only for user charge exemptions but to contribute toward other costs o f receiving services. EDUCATION 9 Policy and Planning: There has been major progress in sector planning, which is well- developed and effectively linked to the budget and nascent MTEF. 9 Expenditures: There has been a significant increase in Government funding over the past few years, doubling as a share o f GDP, and the major share o f spending i s allocated to primary education. In economic terms there has been a shift in favor o f recurrent spending vis-&vis capital expenditures, but the wage share o f recurrent spending has dropped to unsustainably low levels, despite some correction in2002. P Outputs and Outcomes: Since 1999 steady progress has been made in expanding educational opportunities with total enrolment growing. Moreover, enrolment growth has been concentrated among the poor. Repetition rates also fell. The focus on primary education means that sector spending has become increasingly pro-poor. At the same time the subsidies provided by the RGC are pro-poor, and, if implemented as planned, will reduce out-of-pocket costs significantly. However, net enrolment ratios and the completion rate at the primary level are the lowest in the region. Moreover, quality continues to be a concern. Sector policy and planning 3.44 The education sector has developed a high quality planning process, composed of two stages, that is linked to the budget process and the nascent MTEF.The 2001 Education Strategy Paper (ESP) puts forth the sector's main goals, while the Education Sector Strategic Plan (ESSP) describes annual plans and activities for operationalizing the ESP (see Table 3.7). The ESSP i s a rolling plan that is reviewed annually by the MOEYS together with development partners. 3.45 The strategies enumerated in the ESP/ESSP give rise to twelve priority programs funded through the Priority Action Program (PAP).26Moreover, these programs have been costed by MOEYS and linked to the MTEF. The program cost i s estimated to require RGC recurrent spending volumes to triple between 2000-2005 (from 0.9 percent to 2.9 percent o f GDP, and from 10.0 percent to 21.8 percent o f total recurrent spending). The capital investment cost o f the ESSP is estimated at US$ 120 million over the five year period 2001-2005. The indicative MTEF ceilings, while they do not provide full coverage o f costs as estimated by MOEYS, do indicate provision for over 90 percent o f ESP costs. Thus, the education sector has made enormous headway in improving its planning process, linking it effectively to the annual budget and the MTEF, and introducing elements o f program-based budgeting. All o f this should continue to improve the linkage between plans and expenditures and thereby strengthen the poverty focus on sectoral spending. See Table A12 in Annex A for a comparison of ESP program costs and indicative MTEF allocations. Two additional priority programs are fundedfrom the capital budget. ReformingInstitutions to Imvrove Service Deliverv I... 39 Table 3.7: The Education Sector Strategy Paper and Strategic Plan, 2001-2005 Goals Outcomes 1. Achieve the Education For All (EFA) objective EquitableAccess to Services of nine years of quality basic education(grades 1-9) 9 Achieve a95 percent netenrolmentrateinbasic for all children by 2015, delivered in the most education (primary and lower secondary) by efficient manner possible 2015 with a 2005 target of 95 percent in 2. Expand opportunities for post-basic education primary, 50 percent in lower secondary, and 20 through public/private partnerships percent in upper secondary education, with girls Achieve greater equity through appropriate and rural students increasing their share of total financing tools enrolment at all levels of education 9 Double the enrolment in post-secondary education and increase the private share to 85 percent . Improved Quality and EfJiciency of Service Delivery 9 Increase the pass rate fkom 50 percent to 95 percent Strategies Equitable Access to Services 1. Reduce physical constraints and the direct and indirect costs of education to parents (Le., lower their share of total recurrent expenditures to 18 percent at the primary level, or fkom roughly CR 40,000 to 26,000 per child) by prioritizing capital investments accordingly, eliminating formal and informal fees in basic education, providing waivers and scholarships for the poorest in upper levels of education, and providing operational resources to all public education establishments to compensate for the loss of fee income, cover maintenancecosts, andprovide materials and suppliesfor all students. Improved Quality and EfJiciency of Service Delivery 2. Reduce repetition, enhance capacity of lower levels to implement school operational budgets transparently, and maintain sustained increasesinrecurrent spending. Institutional Development and Capacity Building 3 . Provide continuous teacher education, develop and improve core instruction materials, and strengthen the monitoringsystem. Education sector expenditures 3.46 The education sector is financed from three main sources: the Government budget, donors, and out-of-pocket, contributions. Government expenditure in education only accounts for about half o f total sectoral spending, while 35 percent i s sourced from parental contributions, and the remaining 15 percent from external project financing (see Table 3.8).27,28Total planned expenditure on education in 2002, all sources combined, i s estimated at approximately 4 percent o f GDP, or about CR 45,000 per capita (US$ 11per capita). 3.47 Significant shifts in the level and composition of spending took place over the past few years. These shifts reflect commitments to increase domestic recurrent education expenditure, and particularly the non-wage share. Between 1997 and 2000, RGC expenditure on education doubled in nominal terms (from CR 86 billion to 170 billion), and i s projected to 27 As is the case in Cambodia generally, data caveats are in order. Government expenditure by level of education is difficult to estimate and depends on the analyst's assumptions, The MOEYS budget i s structured into four sections: General Education, Tertiary Education, Youth and Sports, and Provincial. Each section has four main chapters: Salaries (Chapter lo), Operations and Maintenance (Chapter 1l), PAP (Chapter 13), and subsidies for scholarships (Chapter 31). 28 Capital expenditure reported inbudget returnsby the MEF captures only RGC expenditure made through the budget. Ineducation, this typically represents RGC counterpart contributions to externally financed projects. See Chapter 4 for further discussionofthe problems with external financing. 40 Cambodia Integrated Fiduciaiy AssessmentPublic ExpenditureReview increase to 290 billion in2002 and 337 billion in2003. As a share o f GDP, education expenditure more than doubled from 0.9 percent to 2 percent between 1997-2002. Table 3.8: Estimated Total Expenditure on Education, 1997-2002 a' 1997 1999 2000 2001 2002 b/ Total Education, Capital + Recurrent(CR millions) 401,251 496,111 503,826 526,932 576,338 As apercentageof GDP 4.4% 3.9% 3.9% 3.9% 4.0% Per Capita (inCR) 36,551 41,371 40,199 40,233 42,937 Per Capita (inUS$) b' 12 11 10 10 11 Percentage Government budget 21% 31% 34% 44% 50% PercentagePrivate 37% 38% 35% 35% 34% PercentageExternal (project support) 41% 31% 31% 22% 16% Total RecurrentExpenditure (Govt. + Private inCR millions) '' 231,908 337,435 341,441 394,948 481,638 Total Govt. Exp.(Capital + RecurrentinCR millions) 86,246 55,732 169,860 229,653 290,200 Percentage GDP 0.9% 1.2% 1.3% 1.7% 2.0% Education share oftotal Governmentexpenditure 7% 8% 8% 9% 12% Education expenditureper capita 7,856 12,986 13,553 17,531 21,620 Government RecurrentExpenditure on Education 83,403 49,781 165,815 212,305 286,200 As a pct. of Governmenteducationexpenditure 58% 68% 68% 75% 84% As a share oftotal Governmentrecurrent expenditure 10% 14% 14% 15% 20% Sources: Expenditure information from 1999 Poverty Assessmentfor 1997 data and private expenditure, TOFE for 1998-2001 RGC expenditure, Budget Lawfor 2002for government expenditure, UNDP and ESSP estimates of external education (project)jkancing for 1999-2002. Populationfigures arefrom the UNFPAprojections based on the 1999 population census. a/ Total expenditure excludesprivate investment expenditure due to lack of information. h/ Expenditurefor 2002 is budgeted. Other years are actuals asper the TOFE. c/ The estimated expenditure reduces private expenditure by the increased Government subsidy for fees, book, materials and supplies. 3.48 Externally-financed project expenditure shows a marked downward trend, while RGC recurrent spending shows a sustained upward trend. Between 2000-2002, external project financing declined by 40 percent in nominal terms and i s expected to decline further in 2003. Its share o f total expenditure fell from 41 percent in 1997 (including private) to 16 percent by 2002. This contributed to a shift in the recurrent-capital ratio from 60:40 in 1997 to an expected ratio o f 85: 15 in 2002. External financing also played a critical role inthis shift through greater budgetary support. While a higher recurrent-capital ratio was essential for the sustainability o f investment, and i s now more comparable with other countries in the region, the ratio may have grown too much. Planning should be undertaken to ensure an adequate level o f capital investment to support sectoral planning for improved coverage. 3.49 Due to increasing allocations and stagnant wages, the share of wages in recurrent spending has dropped considerably. Since the scaling up o f the PAP, budget allocations and expenditure on wages have fallen considerably as a share of education recurrent expenditure, from 78 percent in 1997 to only 67 percent in 2002 (see Table 3.9).29This is low incomparisonto developing economies where the wage share ranges between 70-80 percent. Accordingly, the budgeted non-wage share increased from 22 percent to 33 percent o f recurrent expenditure, which i s high by international standards. 29 Accordingto preliminaryexpenditureestimatesfor 2002. Reforming Institutions to Improve Service Delivery I... 41 Table 3.9: Distribution of Government Recurrent Expenditure, 1997 - 2002 1997 1999 2000 2001 Total (CR millions) 83,403 149,781 165,815 212,305 Per Student (CR) 36,8107 61,840 64,286 75,115 Wage (CR millions) 64,766 111,972 121,028 132,088 Percentage of RecurrentExpenditure 78% 75% 73% 62% Non-Wage(CR millions) 18,637 37,809 44,787 80,2 17 Chapter 11: Operations& Maintenance 19% 23% 23% 23% Chapter 13: PAP 0% 0Yo 2% 13% Chapters3 1,32: Subsidies 3% 2% 3yo 2% Note: The wage/remuneration expenditure under the PAP is classiJiedas wage expenditure. Sources: Expenditure datafor 1997 arefrom World Bank, 1999 Public Expenditure Review. Datafor 1999- 2001 arefrom end-ofyear MEF outturns. Enrolment data arefrom the MOEYS annual statistics. 3.50 Between 1997-2002, non-wage expenditure increased seven fold in nominal terms, with increases above 70 percent in 1999,2001, and 2002. In 1999, this was due to a boost in operations and maintenance (O&M) funding, particularly at the central level. In2001 and 2002, it was due to the budget increase channeled through PAP, while O&M also continued to increase. By 2002, the composition o f non-wage expenditure was 51 percent for PAP, 44 percent for central O&M, and 5 percent for subsidies. The increase was also significant on a per student basis where recurrent expenditure almost doubled in nominal terms, increasing from C R 37,000 to 89,000 per student (US$ 12-22) with more than halfthe increase channeled through the PAP. 3.51 Chapter 11 (O&M) continues to represent a large share of the budget given that O&M for schools is financed through the PAP. It is likely that savings can be generated from these expenditures. For example, o f total Chapter 11 spending in 2001, 69.7 percent was for sports (matches and utilities), and fully 78.7 percent o f the total increase in Chapter 11 (1999- 2001) has been allocated to sports. If these data are accurate, these resources should be reallocated to higher priority programs. 3.52 RGC recurrent expenditure by level of education reflects the ESSP priority goal of achieving EFA (see Figure 3.5).30Over the past five years, on average, 64 percent o f education expenditure was for primary education. The share increased over this period from 60 percent to 67 percent in 2001 and fell back to 62 percent in 2002. The share o f expenditure allocated to lower secondary education fluctuated with changes in the share allocated to primary education. Together, primary and lower secondary education (or basic education) received between 80-84 percent o f RGC recurrent expenditure. The share allocated to basic education i s high compared with other countries inthe region, and that allocated to upper secondary andtertiary educationare quite low. This is illustrative o f the priorities set by the RGC, but also o f tight resource constraints. However, the MOEYS also spends a relatively high share on "other" education. It would be advisable for MOEYS to reassess the contribution o f this category o f expenditure to the achievement o f ESSP goals. 30It was not possible to attribute capital expenditure to the different levels of education because most capital expenditure is externally financed through projects. 42 Cambodia Integrated Fiduciary AssessmentPublic Expenditure Review Figure3.5 Shareof Government RecurrentExpenditure by Levelof Education, 1997-2002 100% +Primary I 80% ,-Lower Secondary 60% Upper Secondary 40% 20% Secondary +Other 0% 1997 1998 1999 2000 2001 2002 1- Basic (grades 1-9)~~ Source: Government budget data and MOEYSstatistics. 3.53 In 1997, private expenditure on educationwas on the order of CR 150 billion, with 62 percent of expenditure at the primary level (see Table 3.10). On average parents spent the equivalent o f about CR 48,200 (US$16 at 1997 exchange rates) for a primary student. Expenditure on students in lower secondary, upper secondary and tertiary education were 2.8, 5.5 and 4.4 times higher. Parental expenditure ranged between 65 percent o f total recurrent expenditure on primary education to 80 percent in upper secondary education. Tertiary education was the most highly subsidized sector, where parental expenditure covered only 50 percent o f the total cost o f education. Table 3.10: PrivateRecurrent Expenditure, 1997 Per Student (in CR) ~ Private Average Pet. of Total Expenditure Poorest Richest National US$ (1997 Ratio to Recurrent (CR million) 20% 20% Average exch. rate)Primary Expenditure Total 148,505 25,581 154,046 65,555 16 64 Primary 92,484 23,904 107,500 48,194 12 1 65 Lower Secondary 35,851 51,900 224,337 134,833 34 2.8 67 Upper Secondary 16,214 75,870 341,228 262,911 67 5.5 80 Post Secondary 3,956 0 255,478 210,454 53 4.4 50 Source: Poverty Assessment estimates, 1997. Educationsector outputs and outcomes 3.54 Faced with difficult initial conditions, Cambodia has made significant progress over the last decade in expanding educational opp~rtunities.~~ Steady progress has been made, 3'The main source of regular data on the education system is the school census conducted annually by the MOEYS, which is fairly consistent over time at broad levels of aggregation. However, trends in breakdowns are sometimes difficult to interpret. Household surveys (in 1997 and 1999) and the population census (in 1998) provide a basis for comparisonandprovide more reliable information on geographicand other variances. Reforming Institutions to Improve Service Delivery I.. . 43 most particularly since 1999. By 2002, the coverage o f primary education reached the equivalent o f 123 percent o f 6-11 year olds, the highest gross enrolment rate in the region. However, along with Lao PDR, Cambodia has the lowest percentage o f 6-11 year olds enrolled in school in the East Asia region, and the lowest percentage o f primary students that eventually graduate (See Annex table A14). Coverage at the secondary and tertiary level i s far from that o f other countries inthe region. The gross enrolment insecondary education is only 24 percent, slightly higher than that o f Lao PDR, and less than half to one-third that o f other countries in the region. The gap i s even wider intertiary education, which enrolls only 1.4 percent o f adults aged 19-22. 3.55 As education in Cambodia is predominatelypublicly provided, the recent increase in public schoolenrolment is criticallyimportant?' Since 1997, all levels o f general education grew at rates more than triple the rate o f population growth (see Figure 3.6). On average, primary enrolment grew at 7 percent a year, lower secondary at a slightly lower pace, and upper and post secondary at highest rate o f 11 and 13 percent respectively although the starting base is much lower for these levels o f education. Most o f the increase took place between 1999/00 and 2001/02 coinciding with the scaling up of the ESSP/PAP. Compared with ESP targets, total enrolment by 2002 grew slightly less than planned. Primary grew slightly more, tertiary education significantly more, and secondary education significantly less.33 Figure3.6: Enrolment1996/97-2001/02 3500000 1 +Re-School 3000000 +Rimary 2500000 Lower 2000000 -Upper Secondary 1500000 I ~ ~ 1000000 Secondary 500000 -o- Total Secondary 0 -t-Rural -Urban Source: MOEYSannual statistics. 3.56 Cambodia now has the highest gross enrolment ratio in the region. By 2001/02, enrolment represented an estimated 124 percent o f the school-aged population, up from 105 percent in 1999/00. However, the net enrolment ratio remains among the lowest. At most, an estimated 86 percent o f 6-11 year olds attend primary school (although household surveys suggest a much lower share). The difference between the net and gross enrolment ratios, and the widening gap between them over time, i s due to an increasing enrollment o f over-aged youths in primary school. According to MOEYS data, over-aged youths accounted for an estimated 80 percent o f the enrolment increase inprimary education since 1999/00. As a share o ftotal primary enrolment, overage youth increased from 15 percent to 30 percent o f the total enrolment. This may suggest an important backlog o f overage youths who will continue to enter for several years, resulting inhigh levels o f coverage needed over a transitional period o f time o f probably about 10 years. Or, it may result from inflated enrolments to receive higher operational budgets. It would 32The private sector accounts for an estimated 1 percent o f enrolment (Poverty Assessment, 1999). 33Infrastructure expansion has not kept pace with increased enrolment. To accommodate the enrolment increase, the percentage o f schools double shifiing increasedfrom 52% to 76% and teacher work loads have intensified. 34 Cambodia Integrated Fiduciary Assessment Public Expenditure Review be advisable for the MOEYS to assess the source o f this growth to enable better planning to accommodate over-aged youths. 3.57 Enrolment growth was concentrated among the poor, reducing the enrolment gap betweenthe richand the poor (see Figure3.7). Nearly allthe growth inprimary enrolment, and for a significant share in lower secondary, was in rural areas with the poorest districts registering the highest enrolment increases inprimary education, particularly for girls. Figure 3.7: Grade 1 Enrolment by CommunityWealth Quintile, 1999-2001 180000 I 160000 140 000 120000 fi~ - ~ 100000 80 000 60 000 40 000 20 000 I Il o 1Total Female1Male 1 I Total ~ 85 (Wealthiest) Wealth Quintile Source: Royal Government of Cambodia MOEYS,Poverty Impact Analysis, 2002. Note: Quintiles refer to communities ranked according to WFPpoverty criteria. 3.58 However,the availabilityof lowerand upper secondary educationgenerally remains limited and out of the reach of the poorest communes. Coverage increased only slightly. Between 1997-2002, the gross enrolment ratio in lower secondary education increased from 27 percent to 34 percent, and from 19 percent to 23 percent in upper secondary. This level o f coverage, along with Lao PDR, i s among the lowest in the world and only half the coverage found inneighboring countries such as Vietnam. 3.59 A comparison of ESP targets versus actuals shows that targets for primary enrolments have been exceeded, although the net enrolment ratio is lower than expected (see Table 3.11). Double shifting increased much faster than planned, and student-teacher ratios are far above the maximum norm. The rural share o f secondary enrolment increased significantly and student-teacher ratios are increasing marginally, although secondary enrolment grew less than expected, and student-teacher ratios are still very low. 3.60 While the Government has made substantial progresson certainaspects of coverage and efficiency, the most impressive result is the increased pro-poor focus of spending. A significantly greater amount o f public spending benefits the poor in 2002 as compared with 1997 (see Table 3.12), as recommended by the 1999 PER (see Annex C for a complete discussion of progress made on expenditure policy since the 1999 PER). In 1997, the poorest 40 percent o f the population received 42 percent o f expenditure on primary education, but only 20 percent, 8 percent and 2 percent o f public expenditure on lower secondary, upper secondary and post- secondary education, respectively. Increased subsidies have made education spending evidently more pro-poor. In 2002, if the ESSP has disbursed as planned, the poorest 40 percent o f the population would receive 39 percent o f total education expenditure, and 50 percent at the primary level. The poorest, however, continue to benefit from only a small share of public subsidies on secondary and post-secondary education, though the incidence o f benefits at the lower secondary level also improved from a pro-poor perspective. Reforming Institutions to Improve Service Delivery I... 45 Table 3.11: ESP Targets Comparedwith Actuals ESP Targets Actual ~~ 2001 2002 2005 1997 2001 2002 Total Enrolment 2,78 1,151 2,989,898 3,528,934 2,3 10,161 2,746,473 2,805,091 Primary 2,289,007 2,385,703 2,388,554 2,3 10,161 2,408,109 2,408,109 Net Enrolment rate 95% 83% 99% 86% Progression .62-.8 0.9 Teachingpostsrequired from 1999for S:T ratio 50:l 1761 2014 -1333 3494 6455 Pct. SchoolsDouble Shifting 20% 20% 52% 69% 76% Pupil teacher ratio 48 48 44 53 57 Lower Secondary 373,842 468,837 846,913 265,895 233,278 283,578 Net Enrolment rate 25% 50% 26% 27% 26% Pct. Ruralparticipation 16% 20% 45% 42% 67% 69% Percentage girls 45% 37% 37% 39% Double shift 30% 50% 29% 35% 43% Student: Teacher Ratio 20 26 19 20 22 Upper Secondary 118,302 135,358 293,467 61,671 105,086 113,404 Net 9% 10% 20% 7% 10% 8% PercentageRural 13% 44% 47% Percentage Girls 40% 35% 32% 32% Student-TeacherRatio 23 26 28 TVET. PTTC. HE.RTTC 22.910 24.193 27.520 18.796 29.612 33.712 Source: Targets ESP 2001. Actuals -MOEYSannual education statistics. - Table 3.12: Share of EducationSubsidyReceived by Income Quintile Quintile: I I1 I11 IV V Percentage of Public Subsidy Using 1997 Enrolment Share Primary 21% 21% 21% 20% 17% 100% Lower Secondary 8% 12% 19% 23% 38% 100% Upper Secondary 4% 4% 14% 21% 56% 100% Post-Secondary 2% 0% 6% 15% 76% 100% Percentage of Government Subsidy Received, 2002 estimate Total 19% 20% 19% 19% 23% 100% Primary 25% 25% 21% 17% 12% 100% Lower Secondary 11% 14% 19% 22% 34% 100% Upper Secondary 7% 7% 13% 26% 46% 100% Post-Secondary 1% 0% 12% 22% 65% 100% Note: The estimate of the subsidy capture for 2002 assumes that the enrolment trend observed in 2001 (i.e. the enrolment increase primarily reflected increased enrolment of childrenfvom the poorest households) would continue into 2002. Enrolment of other income quintiles is capped at 100percent net. 3.61 These results are largely a function of the share of enrolmentby income quintile, as the poorest quintiles represent a larger share of enrolment at the primary level than the wealthier quintiles. However, the wealthiest 40 percent o f the population comprises 61 percent o f enrolment in lower secondary education, 77 percent o f enrolment in upper secondary education and 91 percent o f enrolment intertiary education. Given these distributions, subsidies to primary education, which are a fundamental element of the Government's strategy, and particularly for 46 CambodiaIntegratedFiduciaw AssessmentPublic Exuenditure Review fees, supplies, books and materials, would be strongly pro-poor. Increased subsidies in secondary and tertiary education would increase the amount o f public funding captured by the wealthiest, and particularly Phnom Penh residents, and thus would be largely undesirable, unless they were specifically targeted to the poor. 3.62 The development of more effective financing mechanisms to lower the costs of education by providingsubsidies is likely to have a major impact(coupled with the waiver o f the school enrolment fee and the banning o f informal fees).34According to current estimates, the private share of the recurrent cost per student would fall from 66 percent to 41 percent (and Government's share would increase accordingly). The impact o f the increased subsidy would be greatest in the case o f primary education, where private costs would fall from 65 percent to just 37 percent o f the total. The families o f primary school children would continue to support the costs o f uniforms, food and transport. The per student PAP allocation, estimated to be C R 13,500 in 2002 for primary students, should be sufficient to substitute for what the poor are now spending for fees, books, materials and supplies. 3.63 The increased subsidy would be less significant at other levels of education-with the exceptionof higher education, which, due to a combinationof PAP and other subsidies, i s more highly subsidized than other levels. The planned new subsidies in higher education would further reduce the private share o f per-student costs from 50 percent to just 24 percent. Cambodia's poorest families benefit little from public subsidies in higher education, given that their enrolment at this level is so low. However, total enrolment intertiary education in Cambodia i s among the lowest anywhere in the world, and expenditure on tertiary education in Cambodia is only a small share o f total Treasury-executed spending on education. Under these circumstances, increased subsidies, as currently proposed under the ESP, will disproportionately benefit Cambodia's wealthiest families. A more comprehensive and equitable strategy for public expenditure at this level is needed-such a strategy would target subsidies to poor but promising students, encourage private sector development, and invest selectively inthose public institutions that meet (or are taking steps toward meeting) the new standards for accreditation. 3.64 The other major advance is the decline in repetition rates. According to MOEYS data, between 1999/00 and 2001/02, repetition rates in primary schools fell by more than half (from 22 percent to 10 percent). The decline was most significant in grades one and two where repetition rates fell from 37 to 17 percent in grade one, and from 24 percent to 10 percent in grade two. The dramatic drop in repetition is unprecedented and results in large benefits in the form of reduced public and private cost o f education. It would be useful to further explore to what extent the drop in repetition i s attributable to higher learning achievement as a result o f the remedialprogram. 3.65 I t is not readily apparent, however, that increased RGC expenditure on education over the past five years has improved the quality of education. The Government does not yet have a system to assess learning outcomes across the system relative to defined achievement standards. This i s anticipated in the ESP so that the impact o f expenditure on learning outcomes and its cost-effectiveness may be estimated. 34 For 2002, the estimates assume that: (a) parents maintain the same level of per student private spending as was observed in 1997, adjustedfor inflation, and (b) parentsreducetheir out-of-pocketexpenditureon fees, books, supplies and materials by the same amount of the RGC subsidy provided through PAP for school operating budgets and textbooks. Reforming Institutions to Improve Service Delivery I... 47 Table 3.13: Per StudentRecurrent Expenditure, Public and Private before and after PAP Subsidy 1997 2002 Government PAP Government Per Student Private Per % Subsidy Per Student Private Per % Expenditure Student Private per Expenditure Student Private (in CR) Expenditure Share Student (in CR) Expenditure Share Average Total 36,817 65,555 66% 13,546 89,320 61,531 41% Primary 26,050 48,194 65% 16,043 66,024 37,954 37% Lower Secondary 66,493 134,833 73% 3,535 145,588 147,534 50% Upper Secondary 66,493 262,911 80% 368 145,588 294,201 67% Post Secondary 213,396 210,454 50% 12,920 711,913 222,876 24% Poorest 20% Primary 23,904 50% 10,739 14% Lower Secondary 51,900 44% 54,614 27% Upper Secondary 75,870 53% 84,638 37% Post Secondary 0 0% 0 0% 3.66 An area of further challenge for MOEYS is movingtoward greater gender equity in the provisionof services. MOEYS proposes to pay particular attention to the causes of unequal access to educational services for girls and women and devise measures to remedy the problems (see Box 3.4) Box 3.4: Women and ServiceProvision11: Barriers to Access in Education The problem of unequal access to education services is manifested principally at the lower secondary school (LSS) level, where boys enrollments are 20 percent, and girls enrollments only 14 percent. MOEYS assessments focus on three reasons for the gender disparities in LSS. First 39 percent of grade six students (Le., students eligible to move onto LSS) live at least 3km, and 10 percent live at least 10 km. from the nearest LSS. Given the poor quality o f rural roads, such distances present significant constraints for the majority who lack even a bicycle for transportation. Though there is a well-established system o f boys living inreligious institutions (temples) while attending LSS, no comparable system exists for girls. Even where LSSs are within commuting distance, parents are concerned for the safety o f their daughters duringthe long walk to school. Second, although direct costs o f education are the same for boys and girls, parents generally perceive higher opportunity costs and lower benefits from educating daughters. Under these circumstances, poor families, who are not able to afford to educate all their children, naturally prefer to educate sons. In most families, the opportunity cost o f girls' education i s labor in rice fields as well as care for younger siblings and other household tasks. Third, few schools have toilet facilities and even fewer have separate facilities for boys and girls. This situation creates obvious problems o f modesty, particularly for adolescent girls. These reasons for low girls enrollment are also found inmany other countries. Household demand for girls' education has been found to be more sensitive than that for boys' distance to school in Africa, Asia and Latin America. On price sensitivity, a study using data from nearly 90 countries over three decades found that price elasticities o f demand for primary and secondary enrollment are between 12 and 21percent higher for girls than for boys. Finally, the importance o f toilet facilities as a determinant o f girls school enrollment has been discussed in a variety o f contexts. New and existing programs will address the first two o f these concerns, though to varying extents. Construction programs to increase physical access to LSSs and scholarship programs aimed at reducing the direct cost of education to bringthe total costs o f education below its benefit to families are on the agenda, though total funding is uncertain. Although all newly constructed schools are expected to have toilets, there are no major efforts currently underway to provide toilets in existing school facilities. 48 CambodiaIntegrated Fiduciary AssessmentPublic Expenditure Review ROAD TRANSPORT 9 Policy and Planning: The major constraint is that sector policy goals require clarification. In addition sector goals need to be operationalized in a strategy statement and linked to the annual budget. P Expenditures: Total expenditures have increased significantly, however allocations fall short o f the amounts needed for maintenance and rehabilitation. If used as intended, the Fund for the Repair and Maintenance of Roads (FRMR) could, as first priority, finance routine and periodic maintenance o f the currently maintainable network, plus some rehabilitation each year to restore additional segments o f the network. 9 Outputs and Outcomes: Although a start has been made on reconstruction and rehabilitation, the state o f the road network remains poor, and poor road quality, which results inhightransport costs, remains one o fthe major constraints on growth. Sector policy and planning 3.67 The RGC's goals for the transport sector, in the period 2001-2005 emphasize efficiency, competition and the contributionof transport services to economic growth and regional cooperation. The Cambodian road network consists o f 35,500 kilometers (km) including approximately 4,000 km o f national roads and 3,500 km o f provincial roads under the responsibility o f the Ministryo f Public Works and Transport (MPWT), and 28,000 km o f tertiary roads under the responsibility o f the Ministry o f Rural Development (MRD).35 As Cambodia is in a catch-up phase, MPWT's focus is on rehabilitating and reconstructing the main national roads, building road links to neighboring countries and developing a sustainable road maintenance program. The MRD's draft Policy for Rural Roads emphasizes the establishment o f a strategic tertiary network linkingdistrict centers and improving access to isolated communities. 3.68 However,as pointed out inthe NPRS,while the ministries have strategies, especially for road investments, "the lack of a clear policy direction results in inadequate attention to maintenance and inefficient allocation of the transport budget (local and The NPRS highlights critical issues facing the sector, including "the lack o f a reliable management and financing mechanism for road maintenance, the fragmentation o f public expenditures in the sector, lack o f transparency, and need for clarity in the functions o f ministries and institutions concerned with the management and financing o f roads." The NPRS goes on to point out: "Currently investment decisions are made and contracts awarded directly by authorities other than the transport agencies. Also, transport agencies frequently receive only a fraction o f their official budget allocation. This affects their ability to plan and manage the country's road assets." 3.69 Based on this assessment,the NPRS identifies the followingactions as priorities for the road sector:(a) a reliable, steady, and adequate source o f funding for road maintenance; (b) transparent management o f all sectoral hnds by the official transport agencies; (c) upgrading and repair o f rural, provincial, and national roads; (d) competitive bidding for all government and donor-funded contracts, coupled with public outreach to the local private sector and civil society; (e) institutional strengthening and capacity building; (f) adoption o f clear sector policies and strategies in support o f the NPRS; (g) encouragement o f public transit services using low cost 35As in many countries, the roads sector is housedin two ministries. In Cambodia MF'WT is formally responsible for the overall management and development of the sector, including roads (except for rural roads), road transport, and other transport modes. MRD is responsible for rural development, including rural/tertiary roads (defined as those carrying fewer than 50 vehicles per day). 36National Poverty Reduction Strategy, 2003, p. 69. Reforming Institutions to Imurove Service Delivew I... 49 methods' and (h) development and strengthening of regional links to facilitate the border trade with Vietnam and Thailand. 3.70 Although the NPRS action matrix incorporates many key road sector objectives, there is an urgent need to develop a consensus around a strategy that will guide and facilitatethe restructuringand developmentof the sector.37This strategy would, among other things, lay out the principles for the planning and management o f the road sector. Key principles would likely emphasize: (a) rational economic criteria as the basis for assessing road expenditure needs and determining priorities from among competing activities, projects and regions; and (b) cost recovery from user charges, at least for the primary and secondary road networks. The strategy would also establish the principles o f contracting out most works and use o f open and transparent tendering procedures. In this regard, both MPWT and MRD have expressed their commitment to contracting out road maintenance and rehabilitationto the private sector (see Box 3.5). However there are government-wide issues o f control over and transparency o f tendering decisions and procedures, especially for roads financed by the Government's own funds. Box3.5: Policiesand Strategies for Expandingthe Role of the PrivateSector The RGC's policy is to encourage greater private-sector participation (PSP) in the provision of road and road transport services and infrastructure. There is growing competition in freight and passenger transport services, although state-owned enterprises are still active in both. There also i s a growing capability in the local road construction industry, fostered both by the volume of joint-venture and subcontractedwork under donor-fundedprojects andMRD's programto strengthenthe capabilities of local groups in labor-based appropriate technology (LBAT) methods of construction. These efforts at strengthening local capabilities will be given a boost by the proposed IDA-financed Provincial Rural Infrastructure Project (PRIP) and an upcoming project proposed for funding under the Japan Fund for Poverty Reduction (JFPR) in partnership with the ILO. Both projects will build on ILO activities to promote LBAT, apro-poor initiative designedto maximize the input of local entrepreneurs, materials, labor and ingenuity inthe rehabilitation andmaintenanceprogramsofbothMPWT andMRD. Opportunities for direct private investment inroad infrastructure are limited, however, by political risk, security concems, relatively low levels oftraffic, the lack ofanadequate legal framework for PSPthat recognizes and allocates risk and responsibilities, an inadequate financial system, and a shortage of professionally-managedcompanies capable of handling such projects. 3.71 In the absenceof a coherent sector plan,it is impossible to assess total requirements and arrive at a prioritised,balancedprogramof maintenance and rehabilitationof the road network. At present there is no systematic, unified planning and budgeting process. The analytical tools (e.g., pavement management systems) are in the process o f being established in the sector ministries with the help o f donors, but to date have provided little inputto the planning and budgeting process. The lack o f a coherent plan and budget are partly explained by two factors: (a) the country has been in a period o f emergency reconstruction and i s only now establishing systems and procedures; and (b) approved programs and budgets have not governed actual expenditures. Many decisions are made without the knowledge o f the sector ministries responsible for roads, and there is a lack o f transparency in investment and procurement decisions. 37 A draft policy statement was prepared as part of the recently completed Transport Sector Study (TSS), June 2002. The policy is in its fourth revision, but still under consideration. 50 CambodiaIntemated Fiduciaw AssessmentPublic ExuenditureReview Table 3.14: Road Transport Sector -NPRS Action Plan Matrix Objectives 1. Improve accessibility ofroadnetworks, and ensure cost-effective and safe transportation 2. Ensurethat transport policy supports poverty reduction objectives. 3 . Improve rural transportthrough maintaining existing road assets and enhancingrural access and mobility Strategies ImprovedAccess 1. Restructureintercity public transportto enable better, more cost-effective and safer service 2. Reconstructroads aligned with international standards 3. Rehabilitate roads for connectionbetweenSouthEast andNorthWest regions 4. Maintainroads for a sustainable and efficient roadnetwork 5. Review road code, including road safety initiatives for preparing anational road safety plan Support of Poverty Reduction 1. Formulate (inaparticipatory manner), adopt, and implement anational transport policy, including rural and urbantransport, and a strategyfor private sector participation Improvedrural transport 1. Preparerural roads inventories and mapping 2. Sustainand improve living standardofpeopleby maintaining existing rural roadsusing Labor-Based Appropriate Technology (LBAT) 3. Maintain andrehabilitate rural transport infrastructureusing LBAT to increase rural access and generaterural employment ina sustainable and environmentally friendly way 4. Review availability, affordability, and safety ofruraltransport services outputs ImprovedAccess 9 50 kmlyear ofroadconstructed 9 500kdyear ofroadsrehabilitated 9 2,000 kmlyear ofroadmaintained Support of Poverty Reduction 9 Policy andstrategyformulated in2003-2004andadoptedin2005 Improved Rural Transport 9 60 percent oftotal (3 1,000 km)roadlengthmappedby2005 9 4,500 kmofruralroadsmaintainedingoodconditionby2005 9 3,000 kmoftertiary andsub-tertiaryroadsimprovedby2005 and5.7millionworking days of employment generated (45 percent female) 9 Roadaccidentsinruralareasreducedby25 percent by2005 Road transport sector expenditures 3.72 Total expenditureon roads has increasedsignificantly in the past several years, and are likely to increase further as resources become available from the Fund for the Repair and Maintenance of Roads (FRMR) (See Box 3.6). There has been dramatic growth in capital expenditures for MPWT, which have increasedby a factor o f three (innominal terms) from 1998- 2001, and MRD, which saw a tenfold increase from 1997-2002 (in nominal terms) and much more if PAP allocations from the current budget are included as shown in Table 3.15. This resource shift reflects the high priority given to road rehabilitation in recent years (though not all MPWT and MRD expenditures are for roads).38 It is estimated that total annual Treasury- 38 Consolidated information on planned and actual expenditures on roads is unavailable, as road spending is not separately itemized in the budget and accounts preparedby MEF, MPWT and MRD, nor is it reportedfor the FRMR. It is assumed, based on discussions with Government officials, that 80 percent of MPWT's current expenditure and 90 percent of its capital expenditure, and 70 percent of MRD's current expenditure and 90 percent of its capital expenditure, are for road-relatedactivities. Reforming Institutions to ImproveService Delivery I., . 51 executed expenditure (that is, excluding external project-related assistance) on road rehabilitation and maintenance (excluding expenditures funded from the FRMR) was on the order o f CR 106 million in 2001, equivalent to about US$ 28 million. Of this some US$21 million (76 percent) was for MPWT's national and provincial network39and US$ 7 million (24 percent) for MRD's tertiary network. Capital expenditures made up about 77 percent o fthe total (78 percent ifMRD's PAP expenditures, used almost entirely for road rehabilitation and repair, are included). Table 3.16 summarizes these estimates o f road expenditure in 2001 (assuming that expenditures from the road fund were at the same rate as for 2002). External assistance for the transport sector is estimated at US$ 59.7 million in2001. Box 3.6: The Fund for the Repair and Maintenance of Roads (FRMR) The FRMR was established in June 2000 (Prakas #399) to provide funding for road repair and maintenance. Initially financed from loans from MPWT-owned enterprises, the fund has been supplemented since January 2002 by a surcharge of U S 2 cents per liter on gasoline andU S 4 cents per liter on diesel, levied by the Customs Department at the point of import. In May 2002, with the accumulated revenue reportedly amounting to an estimated US$ 3.5 million, the management of the FRMR was transferred to MEF, under the direct control of the OPM (Sub-decree #39). Target levels of revenue have been estimatedby MEF at US$10 millionto US$25 million annually-though TOFE reports indicate that revenuestotalling about $26 millionhadbeenearmarkedto the fund in2002. So far, most of the funds have been used for road rehabilitation projects (as well as a few rehabilitation projects in other sectors such as water resources), inmuchthe same way as the MEF's capital budget. Rather than provide a predictable basis for maintenance planning and the development of maintenance systems, finds are allocatedbythe OPM to projects proposedby provincial administrations or directly by the MCE, without reference to transparent criteria and processes for determining priorities. Since technical oversight agencies are not routinely involved in the financing decision or in monitoring implementation, there is little scope to monitor performanceto ensure value-for-money. If the FRMR is to achieve its objectives in providing adequate, predictable funding for road maintenance, its institutional framework will have to be strengthenedto clarify its function and oversight structure. Important first steps could include: (a) creation o f oversight arrangements that would monitor the use of funds to be used for roads (such as technical and financial audits); and (b) adoption of a formula- based allocation mechanism that would allocate resources to institutions and provinces for road maintenance on the basis of the length of road maintenance network to be maintained, allowing them to program maintenance activities well in advance and develop the necessary maintenance systems. In many countries, oversight arrangements include the creation of a management board, comprising private and public sector representatives,responsible for applying allocation criteria and contracting out the audit and inspection ofroad maintenanceworks. Table 3.15: Total Current and CapitalExpenditures, MPWT and MRD, 1997-2001 (CR millions) Ministry Expenditure 1997A 1998A 1999A 2000A 2001B 2001A 2002B MPWT Current 6,944 6,373 9,322 20,316 23,050 21,456 22,940 Capital 17,190 6680 19,926 52,267 0 70,682 0 Total 24,134 13,053 29,248 72,583 23,050 92,138 22,940 MRD Current a/ 2,037 2,215 3,256 7,558 13,730 12,513 20,000 Capital 1,649 3,783 12,474 4,810 0 18,163 0 Total 3,666 5,998 15,730 12,368 13,730 30,676 20,000 Grand total 27,820 19,051 44,978 84,951 36,780 122,814 42.940 Pct. of Govt. Expenditure 22% 12% 25% 4.1% 1.5% 5.1% 1.6% IncludesPAP. A=actual; B=budgeted. Source: TOFEand annual budgets. 39 The August Memorandum of Understanding (MOU) in preparation for the Asian Development Bank supported GMS: CambodiaRoad Improvement Project puts the figure for primary and secondary roads much higher at US$ 79 million per year, with the Government providing about 20 percent and development partners the rest. This significant discrepancy needs clarification by MEF, MPWT andMRD. 52 Cambodia IntegratedFidiiciar). Assessment Public Expendiriire Review 3.73 During the reconstruction period, approximately 1,500 km of national roads under MPWT and approximately 11,000 km of rural roads under MRD have been rehabilitated. This has been achieved by donor funded support, government own funded works and includes approximately 900 km implemented by the Military Corps o f Engineers (MCE). Table3.16: EstimatedExpenditureson RoadsOnly, MPWT and MRD,2001 Ministry Expenditure CR millions US$ millions MPWT Current 17,000 5 Capital 64,000 17 Total 81,000 21 MRD Current 9,000 2 Capital 16,000 4 Total 25,000 7 MEFiFRMR Capital 70,000 18 b' Grand Total 176,000 46 Includes PAP. b' MEFIFRMR includes approximately US$3.5 million from the FRMR and the remainder from MEF allocated funds. Source: WorldBank estimates based on discussionswith officials, TOFE,and annual budgets. 3.74 Both MPWT and MRD have ambitious plans to rehabilitate their respective networks. For MPWT's national and provincial road network, projections o f annual road rehabilitation and maintenance expenditures rise to a higho f US$ 176 million as the rehabilitation effort scales up before falling to a long-term sustainable level (see Figure 3.9).40,41To put the cost o f rehabilitation incontext, ifM P W T rehabilitates a further 2,500 km o f its national network at an average cost o f US$ 100,000 per km, the cost will rise to about US$250 million over a five year period (about US$ 50 million per year), and increase the routine maintenance responsibility for the resulting 4,000 km national network to US$2-4 million per year. For MRD's tertiary road network over the 2001-2005 period, MRD envisages the rehabilitation and maintenance o f an ambitious 21,195 km at an estimated cost o f US$ 57 million. This estimate o f needed expenditure i s understood to reflect economical standards appropriate to tertiary roads, however there has been no comprehensive survey o f road conditions nor economic evaluation o f the proposed expenditures. While issues o f funding and maintenance management are beginning to be addressed for the primary and secondary network, they have received little attention as yet for the more extensive tertiary network. 3.75 A comparison of the projected costs of rehabilitating and maintaining the entire road network with current resource availability reveals a shortcoming of enormous magnitude. Assuming an average cost o f US$ 11 million per year to rehabilitate and sustain the tertiary network, and expenditures on the order o f US$ 120-130 million per year for the primary and secondary network, Government resources are able to cover less than one third o f the total annual cost. External financing, though it will help close the gap, looks to be insufficient, at least over the medium term, to meet expenditure needs. 3.76 A principal concern is the lack of expenditures for road maintenance. Almost all rehabilitation works and periodic maintenance on the national and provincial network are carried out on a project basis, whether local or donor-funded. Routine maintenance, to the limited degree it is carried out, and some small-scale emergency repairs, are funded from the recurrent budget. For the tertiary network, most road works-usually regraveling o f strategic roads or emergency maintenance-are carried out usingPAP funds or as projects under the capital budget. 40 Estimated at less than US$40 million per year once failed sections are repaired and road conditions can be held stable through anormal program ofperiodic and routine maintenance. 4' Based on the 2001 surveys by ADB TA 3257-CAM. Reforming Institutions to Improve Service Delivery I.. . 53 3.77 The challenge is to improvepublic expenditure management so that funds are both released and usedfor routineand periodic maintenance, in order of priority. The problem o f inadequate maintenance expenditure will become more important as the size o f the maintainable network increases. For MPWT, the 1,500 km of national roads already rehabilitated require at a minimumUS$ 500 to US$ 1,000 per km per year inroutine maintenance. For MRD,the 11,000 km o f tertiary roads already rehabilitated require at minimumUS$ 300 to US$ 500 per km per year in routine maintenance expenditures. Table 3.17 indicates that the increasing the size o f the maintainable network will have significant impacts on annual maintenance costs. Table 3.17: AnnualRoutine Maintenance Cost (US$) on DifferentNetwork Size Options Size of network(km) US$ ai MPWT 1,500 1,125,000 4,000 3,000,000 7,500 5,625,000 MRD 11,000 5,500,000 28,000 14,000,000 Total 35,500 17,750,000 a/Estimatedunit costs are US$ 75O/kmfor MPWT and US$5OO/kmfor MRD. Source: WorldBank staff estimates. 3.78 Given the likelihood of resource scarcity for the foreseeable future, it is even more importantthat sector expenditures are managedefficientlyand effe~tively.~~ However, sector expenditure management presents serious concerns in this regard. For example, for the primary and secondary network, most rehabilitation works carried out under the national budget, as opposed to donor projects, are not subjected to a transparent process o f planning based on objective surveys and rational criteria. Decisions about the allocation o f funds from MPWT's capital budget (with the exception of counterpart funds for donor projects) are made by MEF and from the FRMRby the OPM, though it i s not clear to what extent decisions are based on the same technical or economic criteria that are applied to other, donor-funded, rehabilitation programs. 3.79 Detailed expenditure data from the FRMR are not available. These problems are compounded by the lack o f a competitive bidding process for large capital projects.43Addressing these shortcomings i s critical for improving the efficiency and effectiveness of expenditures in this important sector. Road transport sector outputs and outcomes 3.80 Roads are fundamental to economic growth and social development. Almost all of Cambodia's border trade with Thailand44,Lao PDR and Vietnam (as well as traffic through the international port o f Sihanoukville) moves by road. Without an arterial network in good condition, the costs o f transporting the country's international and domestic trade and passenger traffic are highrelative to neighboring countries, negatively affecting incomes, prices and welfare throughout the economy. Improved access is a cornerstone o f the Government's NPRS: it helps 42 Internationalexperience suggests four elements are key to puttingthe management and financing of roads on a sound basis: (1) creation of "ownership" by involving road users in the management of roads; (2) stabilization of road financing by securing an adequate and stable flow of funds; (3) clarification of responsibilities so that there is no ambiguity regardingresponsibilities; and (4) strengtheningmanagerial accountability. 43 Government frequently engages in pre-financingarrangementswith contractors, thus building up arrears to suppliers (flows of arrears, though not stocks, are reported inthe TOFE). 44 Recorded trade across the Poipet-Aranyaphathet border amounted to US$ 238 million in 1999, up from US$ 27 million in 1992, partly due to a rerouting of trade following the upgrading ofNR5 and restoration of peace in the area. Some 56,000 tourists crossed the border in2000. 54 Cambodialntegrated FiduciaT Assessment Public Expenditure Review reduce the costs o f farm inputs, increase prices available to farmers, and open up non-farm income-earning~pportunities.~~ 3.8 1 Three decadesof civilwar, politicalinstability,and neglecthad broughtCambodia's road infrastructureto the point in the mid-1990s where sections of even major national routeswere impassableinthe wet season.46With donor assistance, a start ha5 been made onthe reconstruction, rehabilitation, and institutional strengthening needed to bring the network to the condition where it can be maintained on a sustainable basis. Since then, about a third o f the total length o f primary and secondary roads and 11,000 kmo f tertiary roads have been improved. 3.82 In 2002, road conditionsurveys confirmed that 28 percent of the networkis in good and fair condition, 35 percent in poor condition, 28 percent in bad condition and 6 percent under reconstruction or rehabilitation.The main national roads are in better shape than the rest: 40.7 percent was in good or better condition in 2001, compared with 7.7 percent for secondary national roads and 6.5 percent for provincial and urban roads (Figure 3.8).47. Unfortunately, information i s not available with which to reconcile this evidence o f road conditions with information on the lengths o f road recently rehabilitated or to reach conclusions ~- about the effectiveness of maintenance o f road sections once upgraded. Figure3.8: Condition of the MPWT Network, 2001 ' 0Under Conotr'n 10.000 0Poor OFair kaGood E- 8,000 Excellent al ' 6,000 4,000 2,000 0 Main National Secondary National Provincial & Urban All MPWT Roads Source: MPWT/ADB, 2002. 3.83 There has been no survey of M R D ' s tertiary network, but up to 11,000 km of rehabilitation work (mostly regraveling) is reported to have been carried out by MRD in recent years. Mucho fthe rest is acknowledged to be inbadcondition. 3.84 The impact of the poor state of the road network is felt in high transport costs, which impede economicgrowth and service delivery, and in citizensafety. Roadtraffic levels are low, suppressed by poor road conditions, security concerns, and a vehicle fleet that i s 45The results of a survey of the impacts of local road improvements in Siem Reap are given in Jan Sakko, Access, Transport and Local Economic Development, ILO ASIST, 1999. 46Vehicle operating costs (VOCs) on badly-deteriorated roads are some 25-50 percent higher than on well-maintained roads due to higher rates of vehicle wear-and-tear, increasedfuel consumption, and lower vehicle productivity. See, for example, Road Deterioration in Developing Countries, World Bank Policy Study, June 1988. 47 The surveys were carried out under ADB TA 3257-CAM: Strengthening the Maintenance Planning and Management Capabilities at the Ministry of Public Works and Transport. They covered the more important links, amounting to 27 percent of the total length. The condition of remaining national, provincial and urban roads was inferredfrom other data, discussions with MPWT and provincial officials, and informedjudgment. Reforming Institutions to Improve Service Delivery I... 55 expanding fast but from a low base. Nonetheless, traffic has grown more rapidly as some o f these constraints have been addressed. Between 1991 and 2000, the registered four-wheel fleet grew at an average o f 24 percent per year and motorcycles by 15 percent. 3.85 Compared with its Thai and Vietnamese neighbors, Cambodia's road transport industry is embryonic and inefficient. Most goods are carried on old, overloaded trucks with low levels o f utilization. In more isolated regions, goods are transported mainly by motorcycle and animal-drawn vehicles. 3.86 Evenwithout allowingfor under-reporting,Cambodiahas one of the highest traffic fatality rates in Asia: 12.1 deaths per 10,000 vehicles>8The main reported causes are driving under the influence o f alcohol, poor adherence to traffic regulations, and unqualified (i.e,, unlicensed) drivers. Existing record-keeping i s not able to distinguish among other factors that probably also affect road safety: poor road designs, low safety awareness among drivers and pedestrians, unsafe vehicles and speeding. Motorcycles are implicated in a high proportion o f both fatalities and all accidents; they are widely used for public transport and often constitute the largest share o f traffic, particularly in or near settled areas. 3.87 Although rehabilitation efforts have resulted in some immediate improvements in road conditions and institutional performance, there remain doubts about their effectiveness and sustainability over the longer term. Improving the targeting o f sector expenditures to sector priorities, and ensuring that expenditures have higheconomic and financial returns, will be necessary for making roads instruments o f national development. AGRICULTURE49 9 Policy and Planning: Though the NPRS makes progress in sector planning, the agriculture sector still lacks a coherent pro-poor policy framework linked to the budget. 9 Expenditures: The sector is under-funded. In economic terms the shift from wage spending to research (mostly transfers to other institutions) has reduced allocative efficiency. Moreover, sector spending is not targeted to pro-poor programs and regional departments are resource-deprived, > Outputs and Outcomes:Low levels o f funding and the inefficiency o f spending constrain both broad-based economic growth and poverty reduction. Sector outputs in the form o f direct service provision to clients are limited. A lack o f effective sector monitoring makes evaluation o f impact difficult. Sector policy and planning 3.88 Rural issues, particularly improvements in smallholder family farms' productivity and market access, feature prominentlyin the SEDP IIand the NPRS. The NPRS sets out the following sectoral priority objectives: (a) establishing the appropriate legal and institutional framework, (b) improving MAFF's capacity, (c) increasing the knowledge and skills o f farmers, (d) promoting intensification and diversification, and ensuring the security o f agricultural productivity, (e) promoting sustainable natural resource management and conservation, (f) strengthening market access, (g) developing and implementing an irrigation and drainage (ID) information system, (h) strengthening capacity in project formulation and implementation o f 48Comparable rates (2000) are: Vietnam 10.8, Malaysia 5.7, Thailand 5.6, and Singapore 2.9. Developed countries have rates in the range 1.5-1.9.Source: First GRSPASEAN Seminar, March2001. 49This section does not cover land issues. 56 Cambodia Integrated Fiduciary Assessment Public ExDenditure Review sustainable ID projects using a participatory approach, and, among others, (i)facilitation and promotion o f small-scale, private investment inID?O 3.89 Unfortunately, however, the sector still lacks a coherent, pro-poor policy framework. This results in unprioritized objectives (the 2002 MAFF annual report provides a long list o f over 40 objectives, which do not seem to be prioritised, for 2003) and fragmented pend ding.^' Part o f the problem i s the overlapping jurisdiction o f sectoral ministries and the lack o f coordination between them, both o f which are consequences o f the weak administrative capacity that characterizes the rural sector. Without a clear set o f prioritised objectives, it is difficult to analyse sector spending. Moreover, given the weak policy framework, sector expenditures become a function o f historical spending patterns and ad hoc initiatives, which means that there i s no mechanism for focusing expenditures on pro-poor policies. The lack o f an established monitoring and evaluation system means that the results of sector spending cannot be systematically analysed. It is urgent that the Government re-evaluate the role o f the sector by developing a coherent pro-poor policy framework that: (a) prioritizes sector goals and objectives, and develops a strategy to reach them; (b) examines the role for public sector intervention in agriculture vis-his the private sector, and (c) rationalizes the administrative structure o f the sector. Rural sector expenditure^^^ 3.90 The ruralsector is under-funded interms of the size of its nationalbudget allocation relative to its economic importance and its potential for poverty reduction. Though RGC funding for the sector (MAFF and MOWRAM) has increased from 1.35 percent in 1999 to 1.8 percent in 2002 as a share o f the Government's budget (from C R 27.5 billion to 50.0 billion in nominal terms), the size o f the allocation i s low in view o f the relatively large share o f the population, workforce, and the poor who are dependent on agricultural activities for income generation and food security. In per capita terms, current spending was merely US$ 0.93. Unfortunately, however, even given the importance o f the agricultural sector for income generating activities and poverty reduction, it would be unwise to increase sectoral allocations until an agreed, coherent sector policy framework is in place. Given such low overall sector spending, the problem is not so much one o f an imbalance in spending; both operations and maintenance as well as personnel expenditures need to be increased. At the same time, the RGC should look to improvements in efficiency to help fund increased expenditures. A functional review o f the sector (MAFF, MRD, MOWRAM, MLMUPC, etc.) should provide guidance on reducing administrative overlap and fragmentation, thereby releasing resources for more productive use (see Chapter 5).53 3.91 In economic terms there has been a significant shift in composition away from remuneration and toward economic interventions in both MAFF and MOWRAM. Remuneration as a percentage o f recurrent expenditures has declined from 29 percent in 1999 to 17 percent in 2002 according to budget data (for total sectoral spending). Expenditures for operations and maintenance (Chapter 11) have also declined from 57 percent to 48 percent, 50NPRS Action Plan Matrix, pp. 172-185. " "Report on the Situation of Agriculture, Forestry, and Fisheries in 2002," MAFF, January 2003 (unofficial translation). 52 The following covers resources available through MAFF (Ministry o f Agriculture, Forestry and Fisheries) and M O W R A M (Ministry o f Water Resources and Meteorology) to support the development o f agriculture, forestry, and fisheries, includingthe provision o f water to the agricultural sector. 53 MAFF itself cites "overlap" between responsible institutions and departments, "lack of coordination," and "unclear organization structure" as major constraints in the sector. See "Annual Conference on Agriculture, Forestry and Fisheries" (April 10-11, 2002), MAFF, p. 15. Reforming Institutions to Improve Service Delivery I.. . 57 though PAP funding, which totalled 12 percent o f total recurrent spending in 2002, compensates for the decline in Chapter 11resources. 3.92 PAP, which was intended to bring greater alignment to sector policy and spending, does not seem to have had the intended impact. PAP resources do not seem to be directed to high priority services as intended. For example, PAP expenditures on forestry replanting accounted for 28 percent o f total PAP spending in 2001 and 54 percent in 2002. Of these funds, MAFF reports that the majority were spent on remuneration for staff and wages for contract workers (once again indicating the use o f PAP by line ministries to by-pass low civil service pay scales). PAP funds were also spent on animal health, community fisheries, and agricultural research (the Cambodian Agriculture Research and Development Institute, CARDI, and the Royal University o f Agriculture, RUA), as well as the Institute o f Rubber Research (IRR) and animal food for the state zoo. 3.93 I t is also unclear whether the increase in economic interventions (Chapter 30), which more than doubled as a share of total recurrent spending from 1999 to 2002, is pro- poor. In absolute terms economic interventions increased from CR 1,000 million in 1999 to CR 4,427 million in 2002, 87 percent o f which was allocated to CARDI, the RUA, the IRR in 2002 (the remaining 13 percent was allocated to rice seed subsidies). It is not apparent why the sudden four-fold increase in spending on these items became a priority in 2001. Total Chapters 13 and 30 spending on research is thus very high as a proportion o f MAFF's total recurrent budget. Given these high relative and absolute levels o f expenditure, MAFF should undertake a cost- effectiveness analysis to ensure that spending is productive and consistent with sectoral priorities. 3.94 Overall, the level of capital spending in the sector is low, with MEF reporting no capital spending for MAFF between 1999 and 2001. Capital spending, both RGC-financed and donor-financed, is difficult to estimate for this sector and there are discrepancies in data reported from different sources. MOWRAM, however, shows a surge in Government-financed capital spending from CR 236 million in 1999 to CR 3.83 billion in 2000. The rapid rise in spending at MOWRAM raises concerns not only about administrative capacity inthe newly-formed ministry, but also about the proper allocation o f operating and maintenance expenditures to cover the capital base. 3.95 An examination of spending by program in MAFF also raises concerns about the focus on poverty-reducing priorities. Table 3.18 shows spending by service departmedprogram. Forestry accounts for fully 21 percent o f total recurrent spending and nearly half o f total direct service spending. In comparison, agricultural extension accounts for less than one percent o f total recurrent spending.54Animal health and agronomy also account for relatively low shares. Taking account o f the priorities identified in the NPRS would seem to indicate a fairly radical reorientation o f expenditures at the sub-sectoral level. 3.96 Furthermore, an analysis of the regional spending pattern shows that the provinces only receive about 39 percent of total MAFF and MOWRAM spending, while the central level consumes the remaining 61 percent. The share o f O&M expenditures in total recurrent spending made by the central ministries has remained fairly constant at about 40 percent (including Chapters 11 and 13) while the share at the provincial level has increased slightly from 17 percent in 1999 to 20 percent in 2002. Still, given the nature o f the mandates o f these ministries, it would seem that spending i s too concentrated at the central level, with potentially negative impacts on front-line service delivery. ~~~ 54This does not include external assistance, a large share of which is reportedly directed to extension. Due to the lack of data, however, it is not possibleto analyze the total amount of spendingon extensionservices (see Chapter 4). 58 Cambodia Integrated Fiduciaw AssessmentPublic Exuenditure Review Table 3.18: MAFF Budget Allocationsby DepartmentProgram (Share of Total Recurrent), 2002 Departmentmrogram Share Forestry 21.1% Fisheries 11.2% Agronomy 6.3% Animal Health 4.8% Agricultural Extension 0.8% Machinery 2.0% Agro-Industry 0.7% Sub-total 46.9% Rubber 6.6% Educationaland Research 22.2% Sub-total 28.8% Central ministry 23.3% Total 100.0% Note: Does not include external assistance. Source: MAFF. 3.97 Concerns about the efficiency of sectoral resource allocation are not limited to MAFF and MOWRAM alone. Cambodia's rubber companies, which are divested public enterprises under the jurisdiction o f MAFF's General Directorate o f Rubber Plantations, are not profitable, and appear to drain some limitedresources from the sector. In2001, for example, total profits from the seven rubber enterprises amounted to CR 700 million on a capital stock valued at approximately CR 420,000 million, which indicates a rate o f return o f about 0.2 percent. Some companies, such as Chamkar Andaung, have been posting losses for years. Though MAFF does not report regular subsidies to these enterprises, it does provide funding on an "emergency" basis. It is not clear how companies' operating deficits are funded. Rationalization o f management of these companies, and a redirection o f resources away from them, would improve efficiency. Sector outputs and outcomes 3.98 The performance of the agriculture sector is central to the provision of income generating activities, the reduction of poverty, and to increased food security for the majority of the population. More than 70 percent o f Cambodia's population are employed in agricultural production and most are smallholders with less than 2 hectares per l i o u ~ e h o l d . ~ ~ Women make up the majority o f farmers (1.4 million compared to 1.2 million men). The overwhelming majority o f the poor (79 percent) are employed in the agriculture sector, and the highest poverty rates are found among households headed by farmers and households with little or no formal e d ~ c a t i o n . ~ ~ 3.99 In spite of its growth and poverty-relatedpotential,sector performanceis hampered by numerous constraints. For example, a recent study points to a number o f key constraints in the rice sector, in which the majority o f Cambodian farmers operate: (a) low agricultural productivity, (b) high credit costs for farmers and millers, and (c) the need for improvements in seed varieties, access to inputs, and agricultural e~tension.~'The Government's sector strategy would need to address these constraints, as well as those specific to other key agricultural sectors, inorder to providethe underpinnings for more efficient and effective sector expenditure. 55 Cambodia SES, 1999. 56Poverty mapping results from WFP and "Estimation of Poverty Rates at Commune Level in Cambodia" (RGC, October 2002). 57"Rice Value Chain Study: Cambodia," Agrifood ConsultingIntemational (for the World Bank), September 2002. Reforming Institutions to Improve Service Delivery I.. . 59 Box3.7: Womenand Service Provision111:Barriersto Access inAgriculture Although women make upthe majority of farmers, they are aminority amongbeneficiaries of public services, accountingfor only about 20-30 percent ofparticipants inextensionprograms. The direct causes for low service delivery to andutilizationby women are numerous: h Agricultural researchandextensiondonot focusonthe activities offemalefarmers,which is a problem given that menand women specialize indifferent tasks. P Access to extensionservices oftenrequires traveling to the district center which is an average of 9 km from the village. Safety concerns make women hesitantto travel. Moreover, their responsibilities for cooking and caring for childrenmake it difficult for women to be gone for more than a few hours at a time. h The overwhelmingly maleextensionservicestafffinditsocially difficultto engageinface-to-face communication with female farmers. Menare offered training and extension activities even when their wives are the ones carrying out the tasks being studied.Experience inCambodiaas inother countries shows that this knowledge is rarely passedon effectively from husbandto wife. Female extension agents tendto have easier time communicating messages to female farmers, but it i s estimatedthat only about 5 percentof agricultural extension agents are female. 3.100 Moreover, the importance of the sector is not matched by either its resource allocation or administrative capacity. Nor is reliable, consistent data on sector outputs and outcomes available. MAFF does maintain economic statistics on the sector, including data on crop, livestock, and fisheries production, for example, which are produced annually. It has very limited data on sector service delivery. The only information available from MAFF pertains to the number o f vaccines provided and the number o f participants in extension training programs (this data was available in 2002, but not in 2001, inthe annual conference materials). The data that are available reveal the limited scope o f MAFF's provision o f direct services to clients. As would be expected given budget allocations, the number o f participants attending some form o f extension training was very limited (given 2.6 million farmers, and fewer than 20,000 MAFF extension program participants in 2002, it would take 65 years to reach half o f Cambodia's farmers). Establishing a program to monitor MAFF outputs linked to key sectoral outcomes and sector planning instruments i s necessary for improving resource targeting. CONCLUSION 3.101 There is thus considerablescope for improvingthe impactof spendingby improving the key linkages in the policy-expenditure-output-outcome production process. Table 3.19 summarizes the findings. Sector policy development, especially intransport and agriculture, and its link to the NPRS, will necessarily become a priority. The efficiency o f spending in economic, geographic, and programmatic terms, will need to be improved if poverty i s to be reduced further. Indeed, given competing demands for resources over the medium term, it will be essential for the Government to improve the efficiency and effectiveness o f spending in both the social and economic sectors. Improving the accountability o f Government to citizens i s a critical avenue for improving the impact o f public spending. 3.102 The analysis suggests two overarching measures for improving the efficiency and effectiveness of expenditures. First, the priority focus o f spending should be sharpened by improved targeting to the poor. Second, the institutional mechanisms that mediate between resources and services need strengthening. Many o f the problems and their consequences discussed in this chapter are both systemic and institutional in nature-they affect the entire public sector and cannot be resolved at the level o f individual sectors. Thus, these first three chapters on resource generation, allocation, and policy would be incomplete without an analysis o f the institutional underpinnings o f service delivery. In order to improve service delivery, the 60 CambodiaIntegrated Fiduciary Assessment Public ExpenditureReview Government must improve the institutions that translate expenditures into outputs-the public expenditure and financial management system (Chapter 4), the civil service (Chapter 5), and decentralized governance (Chapter 6)-each of which are taken up inturn. Table 3.19: Linkagesbetween Sector Policy,Expenditure,and OutputdOutcomes Policy and Planning Expenditure Output/Outcome Health 9 The healthsectorhas 9 The allocative efficiency of 9 There havebeensome made important progress spending could be increased significant achievements inthe by basing its strategy on (the wage share is too low and sector, especially regarding monitoring outputs and maintenance spending is communicable diseases and outcomes. inadequate). expanded coverage o f the 9 Further work is necessary 9 The shareofspendingatthe population. to cost and operationalize provincial level is too low and 9 However, serious problems the strategy by linking it the high inter-provincial need to be addressed, effectively to the budget. variance results in inequalities including low and uneven inper capitaspendingacross access to services and the the country. inadequate pro-poor incidence 9 The result isinadequate o f spending targeting of resources, which 9 Servicequality continuesto reduces the pro-poor impact o f be a concern. spending. Education 9 The sector planning 9 Therehasbeena significant 9 Steady progresshasbeen process is well-developed increase in Govemment funding made since 1999,in expanding and effectively linked to over the past few years. educational opportunities with the budget andnascent 9 The major share ofspendingis total enrolment growing. MTEF. allocatedto primary education. 9 Enrolmentgrowthhasbeen 9 The wage share ofrecurrent concentrated among the poor. spending has dropped to Repetition rates have also unsustainablylow levels. fallen. 9 The focus onprimary education means that sector spending has become increasingly pro-poor. 9 However, net enrolment ratios and the completion rate at the primary level are the lowest inthe region. 9 Servicequality continues to be a concern. Road 9 Sector policy goals 9 Total expenditures have 9 Although a start hasbeen Transport require clarification. increased significantly, made on reconstruction and 9 Sector goals are not however allocations fall short rehabilitation, the state o f the operationalized in a o f the amounts neededto bring road network remains poor, strategy statement, nor are the network up to a and poor road quality, which they linked to the annual maintainable state. results inhightransport costs, budget. 9 Routineandperiodic is one o f the major constraints maintenance are not funded on on growth. an on-going basis. 9 The FRMRcould, as first priority, finance routine and periodic maintenance o f the currently maintainable network, plus some rehabilitation. Agriculture 9 Though theNPRSmakes 9 The sector is under-funded. > L o w levels of funding and the progress insector 9 The shift fromwage spending inefficiency of spending planning, the sector still to research has reduced constrain both broad-based lacks a coherent pro-poor allocative efficiency. economic growth and poverty policy framework linked b Sector spending is not targeted reduction. to the budget. to pro-poor programs and 9 A lack ofsector monitoring regional departments are makes evaluation o f imoact resource-deprived. difficult. 4. REFORMINGINSTITUTIONSTO IMPROVESERVICE DELIVERYI:THE PUBLICEXPENDITUREAND FINANCIALMANAGEMENTSYSTEM "A critical aspect of NPRS preparation and implementation lies in the link to the budget. Budget allocations need to be consistent with the NPRS and,just as importantly, the execution and management of the budget mustfollow as intended."58 THEREFORMAGENDA 4.1 The Royal Government of Cambodia (RGC) has placed public expenditure and financial management (PEFM) reform squarely on its development and poverty reduction agendas?' Recent initiatives have built on a previous round of reforms, launched after the promulgation of the Organic Budget Law in 1993, which established the institutional architecture for budgeting, planning and budget execution. Priority was given to the development of systems that guaranteed budgetary discipline, through a process of commitment control and centralized payments system. These reforms achieved their objective in delivering-with the exception of a period of fiscal instability following the disturbances of 1998-budgetary restraint, reflected in current balances of over 1.2 percent of GDP in each the last four years. They have also been instrumental in increasing expenditures on the priority sectors of education and health, while reducing spending on defense and security. 4.2 It has become increasingly apparent, however, that weaknesses in the public expenditure and financial management system (PEFM) not only have high costs interms of allocative and operational efficiency but also create unacceptably high levels of fiduciary risk to public funds. The cash-based payments system has emerged as a major constraint. Increasingly, budget execution has suffered from delays and an unpredictable release of funds, due to cash constraints, undermining operational planning, and leading to the build-upof arrears. The system i s plagued by gate-keeping and deficient accounting and reporting systems, thus leading to a weak control environment and increasing opportunities for corruption. Indeed, in comparative perspective, Cambodia's PEFM system ranks below average (as compared to the low income countries assessed by a joint World Bank-IMF diagnostic tool), indicating that Cambodia's systemrequires substantial upgrading(see Box 4.1). 4.3 Attempts have been made to address these problems by implementing pilot initiatives, notably the Accelerated District Development and Priority Action Programs (as recommended by the 1999 PER), together with a "sector wide" approach inaid coordination, first ineducation and health. This has improvedthe alignment of resources with policy objectives and helpedchannel funds to operational units,but their impact has been muted by liquidity constraints and concerns over the adequacy of control arrangements. Within MEF, attention has recently turned to the improvement of treasury operations, cash management, and public accounting so as to address these concerns. At the same time, MEF has made progress in the development of a 58"Cambodia: NationalPovertyReductionStrategy," December2002,p. 127. 59 This chapter i s based inpart on the CambodiaCFAA (see Annex C). 61 62 Cambodia IntePrated Fiducian, Assessment Public ExDenditure Review mediumterm expenditure framework, which, it is hoped, will improve resource allocation and the predictability of resources over the medium term. Box 4.1: Cambodia'sPEFMSysteminComparativePerspective The World Bank and IMF recently developed a diagnostic tool to assess the quality of public expenditure and financial management (PEFM) systems inthe context of the Highly Indebted Poor Country (HIPC) debt relief initiative. One of the objectives of the tool is to "assess the existing capacity for tracking overall public spending, including poverty-reducing spending." The tool assesses PEFM systems according to fifteen criteria that focus on budget formulation, execution, and reporting. The assessment tool also provides an indicative benchmark for each question. The benchmark is not necessarily the ideal, but rather the standard that low income developing countries committed to comprehensive reform should be able to meet. The joint Board paper found that for the twenty-four countries assessed, the mode number of benchmarks met was eight, and the mean was six. Nine of the twenty-four HIPCs were deemed to require "some upgrading" as they met more than half but less than two-thirds of the benchmarks. The remaining fifteen HIPCs were deemed to require "substantial upgrading," as they met less than half the benchmarks. These assessments were agreed to by the country authorities. A Bank assessment of Cambodia's PEFM systemas part of the IFAPER found that Cambodiamet only three of the benchmarks,putting it on par with Niger and Zambia. What the assessment reveals i s that compared with its low income country peers, Cambodia's PEFM system is below average, in spite of reform initiatives undertaken in the context of Bank-Fund programs and with other donors' support. While cross-country tools such as these cannot provide perfect comparability, and while they only provide for comparison with a subset of countries, they are useful for indicating where countries stand relative to a reasonable standard. The comparison also shows that Cambodia's PEFM system i s weak across all three dimensions-budget formulation, execution, and reporting-raising concerns about the capacity of the systemto performthe most basic of tasks. At the same time, as in the case of a number of HIPCs, the low initial assessment provides Cambodia with an opportunity to demonstratefuture progress both in absolute terms and in comparison to other low income countries. Source: "Public ExpenditureManagement: CountryAssessment and Action Plan," draftfor comment, September 2002, IMFNB. 4.4 Given capacity constraints the report proposes a gradualist reform strategy. This will entail the operation of parallel systems for several years to come, while measures are gradually taken to delegate responsibility for financial management to the field level, develop adequate systems for cash management, and strengthen the control environment. The temporal perspective i s necessarily long term, as the requirements in terms of capacity building are substantial. However, efforts to reduce the fiduciary risk to public funds need to be addressed decisively and with urgency. At the same time it will be necessary to strengthen the reform strategy by improving the accountability of service providers-both direct and indirect-to service beneficiaries. Improving incentives through greater accountability i s a promising route to greater efficiency and effectiveness inpublic spending (see Box 4.2). Reforming Institutions to Improve Service Delivery 1.. . 63 ~ ~ ~~ Box 4.2: AccountabilityMechanismsinPractice:International Experiencewith Participatory Public Expenditure Management Broadly speaking participatory public expenditure management (PPEM) involves citizenbeneficiary input at the budget formulation and execution stages as well as at the impact evaluation stage. PPEMhas great potential to improve expenditure processes and outcomes inCambodia. International examples of good practice relevant for Cambodia include: Pro-poor budget analysis in Gujarat, India: DISHA (Development Initiatives for Social and Human Action) pioneered participatory budget analysis in India in 1992 by engaging in review of budget documents, focusing on agency allocations and their intended beneficiaries and discrepancies between policy statements and actual allocations. DISHA's short "budget briefs," which now number about 30 per year covering agencies most involved inpoverty reduction, containprovide budget data and analysis in an accessible form. Observers credit DISHA with improving linkages between planning and budgeting and facilitating the release of funds to priority sectors. Participatory Budget Formulation in Porto Alegre, Brazil: Porto Alegre, with 1.3 million inhabitants, has developed a highly participatory model of citizen involvement in the budget formulation process, which i s usually regarded as the domain of technocrats. Citizens meet repeatedly in formal assemblies during the budget formulation process and are responsible for proposing quantitative rankings of their priorities. An elected Council of Participatory Budgeting proposes revisions to the mayor's budget proposal and lobbies the mayor and the Chamber of Deputies, which ratifies the budget, to adopt a budget that corresponds to the preferences of the citizens. The method i s credited with improved outcomes in access to water and sewerage services, school enrolment, road pavement, and even increased tax collections4ue to greater transparency and motivation to pay. Over eighty Brazilian cities are now following Porto Alegre's lead. Government Watch in the Philippines: G-Watch, an initiative of the Philippine Governance Forum, was established in response to the inadequacies of the Government's performance evaluation systems. G-Watch's objectives are to develop a framework for monitoring public sector performance and to engage civil society, the private sector, and the Government in order to institutionalize a monitoring system. G-Watch selected specific projects to monitor and sent investigators out to the project sites to do field work on project status and implementation. Investigators uncovered both incomplete projects-that had been reported as complete-as well as projects that failed to deliver any services at all. The findings were first presented to the Government before releaseto the public. These cases are just several among many examples of good practice methods available for increasing participation and accountability in public expenditure management. Others include public expenditure tracking surveys (Uganda, e.g.) and citizens report cards (India and the Philippines, e.g.). These methods would be suitable for addressing accountability issues in Cambodia because they could buildon existing civil society networks and initiatives. Source: Songco, TheParticipation Group/SocialDevelopment Department, World Bank (2001). ALIGNING RESOURCESWITH POLICIES AND STRATEGIES 4.5 The purpose of expenditure planning and budgeting systems is to align available resources with Government's development objectives, ensuring that agreed policies and implementation strategies are appropriately financed. This section argues that resource alignment cannot be achieved through Cambodia's conventional budget system, which budgets for institutions and inputs. This problem poses a major challenge for further development and implementation of the RGC's poverty-reduction strategy (see Box 4.3 on the linkages between the budget process and the NPRS). Solutions lie in the development of a programmatic budget structure that matches resources with agency functions and management structures, progressive delegation of decision-making and deconcentration of resources closer to the field level, and the adoption of medium- term perspective to resource programming that can facilitate the reallocation of resources and make resource allocation more predictable. Within this broad framework, specific measures are required to ensure that resources are targeted geographically as well as by 64 Cambodia Intearated Fiducian, Assessment Public Exuenditure Review function, to make adequate provision for capital and maintenance spending, and to integrate external assistance intothe resource allocationprocess. Box 4.3: Linking Policies and Budgets: The Case of the NationalPoverty Reduction Strategy (NPRS) An important aspect of the poverty reduction challenge lies in the linkage of the NPRS to the budget. InCambodiathe lack of coordination betweenMEF and MOP meantthat the preparationprocesses for the NPRS and the 2003 budget were not linked. While there is an NPRS chapter on budget allocations and NPRS priorities are generally being accorded budgetary priority, the process for developing these allocations was not directly linked to the NPRS process. The reasons for the disconnect include: (a) the NPRS timetable, which was not built around the annual budget process; (b) the traditionally bilateral negotiations between MEF and line agencies, which are based on line item budgets; (c) the nature of the NPRS process itself, in that line agency proposals were only partially costed and delinked from annual budgetpreparation; and (d) capacitylimitations at the line agency level and inthe MOP. These shortcomings are reflected in several weaknesses in the NPRS process. One issue that affected the NPRS process, highlighted by arecent OD1study (2002), was the absence of clear guidanceto line agencies as to the criteria that should be used to prepare proposals. Initially agencies were advised to limit their submissionsto limited targetedprograms, but were subsequently advised that their submission should reflect their entire programs, highlighting the poverty relevant aspects. Only about half of NPRS policy and program activities were costed. Overall, the NPRS is not fully resourced and any costing were considered preliminary. This to some extent reflects the varying analytical and policy capacity across line agencies, as well as the underlying nature of the budgetprocess. A disconnect between the PRSP and the budget formulation process is not unusual even in countries where Ministries of Finance do play leading roles in the PRSP process. At the same time international evidence firmly suggests that strong links between policies and programs and the strategic approach set out in the PRSP, and the annual and medium term budget, are critical to the implementation and credibility of national PRSPs. Source: Klugman and Taliercio (2003). Frombudgetingfor institutionsto budgetingfor programs 4.6 For the most part, Cambodia's budget system continues to follow a traditional approach: resources are allocated to institutions through budget ceilings, the institutions prepare bids which comprise a detailed breakdown of the inputs they require, these are then negotiated with the Ministryof Economy and Finance's (MEF) Department of Budget and Financial Affairs (DBFA).The approved budget presents a detailed breakdown of inputsby chapter, distinguishing allocations for salaries and allowances, operating costs, economic and social interventions, and capital. Input allocations are further broken down into sub-chapters, which are used for the purposes o f budgetary control, and line items, with a total o f forty-three line items detailed in agency budgets for operating costs. Agencies have limited authority to alter the composition o f inputs during execution - virement within chapters has to be approved by the Minister o f Economy and Finance, and between chapters by the Office o f the Prime Minister (OPM). 4.7 While this approach has the merit of facilitating control of agency spending, it has hindered allocative and operational efficiency (as indicated in the World Bank's 1999 PER). Since the institution, rather than the function, i s the object of expenditure, the needs o f the institution-administrative overheads-have the first call on limitedresources, rather than service delivery, and little attention i s given to the institutions outputs. Since the institution i s a single entity for budget purposes, it i s often difficult to target resources to specific functions or departments, line managers may not even know how much money they have been allocated, let alone track the allocation o f resources during execution or link resource allocations to the performance of various functions or departments. 4.8 Solutions lie inmoving toward program budgeting, inwhich resourcesare allocated to support functions rather than institutions, with expenditure programs aligned with Reforming Institutions to Improve Sewice Delivery I.. . 65 agencies' policy objectives and internal management structures. The first step in a move toward a program budget structure, i s to clarify the purpose of public spending, the policy objectives and the strategies by which these objectives can be achieved using the available resources.The second step i s to align expenditureprograms with agency management and budget structures. Progress in this regard has already been made in the education and health sectors, where expenditure programs have been aligned with operational management units. The last step i s the definition and monitoring of outputs and outcomes that programs are expected to achieve. MOEYS and MOH have already developed a comprehensive set of performance indicators for PAP, though these are not yet routinely reported through the budget process. 4.9 The key consideration is how fast the budget can be reoriented toward a program approach. Progressinthis area is closely linked to the rolling-out of reforms inbudget execution across government since improvements in planning and programming are unlikely to have a significant impact unless they are accompanied by more predictable and timely delivery of funding. Most agencies have yet to address the first steps: design of a policy framework and implementation strategy. Consequently, while progress can be made over the next two to three years in the priority sectors - agriculture, rural development and transport - implementation of a programmatic approach across government i s likely to take longer, which i s why it i s important for the Government to start sooner rather than later. Programmingfor investment 4.10 During the 1990s, considerable effort was put into the establishment of a rolling, three-year Public Investment Program, managed by the Ministry of Planning (MOP). This sought to match top-down allocations for investment, based on a macro-economic framework for the budget agreedby the Council of Ministers, with programmed allocations for on-going projects and a prioritized listing of new investments presented by the line agencies. MOP i s supposed to appraise line ministries' proposals to determine whether they meet government's development priorities and whether costings are realistic. 4.11 Unfortunately, the PIP has had little impact on the expenditure management process or expenditure outcomes, indicating a lack of progress since the 1999 PER. Project proposals are subject to cursory review. The result is a shopping list of projects. Allocations for capital expenditures, through the budget's Chapter 50, are supposed to reflect the domestic financing requirements identified in the PIP. In practice, no such allocations are made. Instead, the investment budget i s managed by the MEF's investment department, with funds allocated to projects approved by the Council of Ministers or the OPM directly. A listing of investment projects i s presented in the annual budget, though additional projects are approved in-year. Usually expenditures are posted against the nominally responsible agency, though this i s not always the case. Allocations of counterpart funds are made inmuch the same way. As a result, the recurrent budget process i s largely divorced from that of the capitalbudget. 4.12 The PIP has now lost credibility as a planning tool. Agencies recognize that the PIP plays no meaningful role in investment decision-making, submissions are tardy, with scant information on proposed projects and virtually none on actual disbursements, while preparation and Council of Ministers approval of the annual PIPtends to slip outside the budget calendar or- as in2001-is dropped altogether. As proposed inthe previous PER, with the introduction of the MTEFattention should turnto the integrationof the programminginstruments. Programmingfor maintenance 4.13 Planning and budgeting for maintenance presents a specific set of problems, raising concerns about the sustainability of recent and programmed investments, particularly in road 66 Cambodia Integrated Fiduciary Assessment Public Expenditure Review rehabilitation, but also in the social sectors and in agricultural infrastructure. There are few incentives for agencies to allocate scarce resources to maintenance, partly because of the not unreasonable assumption that donors will cover the cost of rehabilitation once assets have deteriorated. 4.14 In the social sectors, centralization of maintenance functions is impractical. Funds are needed at the facility level to effect minor repairs and undertake periodic maintenance of equipment and buildings. In this context, PAP provides a sounder mechanism for allocating resources for maintenance than the conventional budget mechanism. Allocations for maintenance under these programs are no guarantee that funds will not be diverted to address more pressing needs, but facilities can be required to prepare maintenance plans and central agencies can monitor whether adequate measures are being taken to maintain assets. 4.15 In the case of roads, separate provision for maintenance spending is required if maintenance is not to be crowded out by rehabilitation projects. The recently created Fund for the Repair and Maintenance of Roads provides earmarked revenues for the road network, but in the absence of an appropriate allocational and oversight mechanisms, it has not had a significant impact onroad maintenance (see Box 3.6). Managing external assistance 4.16 Reported external assistance flows in 2001, at US$ 472 million, exceed domestic revenues, at US$ 390 million, by a significant margin (see Chapter 2). The bulk of this assistance has been channeled off-budget, both in the sense that resource allocations are not reflected in the Government's budget documents and that funds are not disbursed through Treasury.60Most grant-financed assistance i s disbursedthrough special accounts following donor- specific procedures and so information on actual and programmedexpenditures i s not captured by Government accounting systems. 4.17 Lack of information and the absence of effective instruments to guide the allocation of external financing seriously undermine the integrity and effectiveness of the budgetary system. At present, it is impossible to assess the impact of external project financing on overall resource allocations, let alone its implications for future pattern of on-budget expenditures, as regards provision for operation and maintenance costs for new investments. There i s a considerable risk that the proliferation of donor financed projects has led to policy drift in some instances, by financing investments that are only tangentially related to agencies' development strategies. 4.18 Improving the effectiveness of development assistance will require action by both Government and donors. There is a growing consensus that development assistance i s most effective when it i s channeled to countries with good policies and effective institutions. In fact, donors, both bilateral and multilateral, are moving toward systems that allocate aid based on country performance characteristics. The World Bank's Country Policy and Institutional Assessment (CPIA) i s used to allocate concessional resources to low income countries, while the U S Government has recently proposed a Millennium Challenge Account (MCA) for countries that "govern justly," "invest in people," and "promote economic freedom." Cambodia could increase the effectiveness of development assistance and possibly increase total assistance by making progress on good governance and improving administrative capacity. On the donor side, greater coordination with Government and among donors i s clearly needed. Donors will also be M,Most donors negotiate their assistance programs with the CDC, chaired by the Minister of Economy and Finance, which acts as signatory to their donor country programs and individual project agreements. Most grant financed assistance i s disbursed through special accounts following donor specific procedures. The MEF oversees the disbursement of loan financed extemal assistance. Reforming Institutions to Improve Service Delivery I... 67 challenged to rationalize their salary supplementation practices as the Government moves forward with its civil service reform program. 4.19 At the sectoral level, the solution lies in the further extension of the sector wide approach adopted in health and education, as recommended by the 1999 PER (see Annex C for a complete discussion of progress made on public expenditure management reform since the 1999 PER). In the education sector, considerable progress has been made in the creation of coordination mechanisms involving agency management and external partners. This has helped forge a broad consensus on policy and improve the complementarity of donor inputs. Both the ADB andEUhave indicatedthat they are willing to provide sectoral budgetary support under the rightconditions. 4.20 Improvements in reporting on external assistance flows are critical. At present, the most comprehensive source of information on external assistance i s the Council for the Development of Cambodia's database. This consolidates information on actual and programmed disbursements of external assistancereported by donors directly. Reporting i s patchy and suffers from double counting, but the principal problem i s that the database i s not compatible with the budget structure. The database's usefulness as a monitoring and programming tool could be greatly enhanced simply by recordingthe Government's implementing agency for eachproject. Formulatinggender-responsivebudgets 4.21 There is no functional mechanism to ensure that relevant gender issues are incorporated into the planning process or, if identified, are appropriately funded, even though it is clear that woman encounter there are major obstacles to priority services for women (see Chapter 3). The Ministry of Women's and Veteran's Affairs (MOWVA), by engaging in three types of activities-influencing national development strategies, influencing sectoral strategies, and implementing projects directly-has made some strides in coordinating public programs important to the well-being of Cambodian women. MOWVA has been less successful, however, ininfluencing sectoral policies andprograms. While there is a reasonably highlevel of awareness among line ministries of the need to address gender issues, there is little capacity to put this into practice. At the same time, MOWVA has neither the sector-specific skills nor the institutional authority to influence sectoral planning and budgeting. Though MOWVA has established a system of gender focal points inline ministries, the gender focal points tend to be fairly low level staff without decision making responsibilities. As a result, the gender focal points are not effective in mainstreaming gender issues. In fact, the only other national body to have a formal gender mainstreaming strategy, MOEYS, developed it without significant input from MOWVA. Most recently, MOEYS prepared a Gender Mainstreaming Strategy (2002-2006) in March 2002. Despite these promising actions, however, the strategy was neither costed nor developed as part of the ESSP process. 4.22 M O W V A will have to focus on policy analysis as well as advocacy if it is to effectively mainstream gender issues. The recent surge of policy work around the NPRS has provided a first opportunity. Further work on the development of the MTEF and sector programs will provide another. For this to happen, central and line ministries would need to systematically include MOWVA in monitoring efforts and provide it with the information necessary to monitor activities. Developinga mediumtermbudgetingperspective 4.23 MEF is in the process of establishing a medium-term expenditure framework (MTEF). An MTEF for the period 2003-2005 was presented at the launch of the 2003 budget 68 Cambodia Integrated Fiduciary Assessment Public Expenditure Review process, with further analysis in the NPRSa61 The MTEF provides projections of the aggregate resource envelope for on-budget funds and resource allocations to institutions. Resource allocations are determined by allocating the increment over the aggregatedbaseline ceiling, taken as nominal institutionalallocations for recurrent expenditures and total capital expenditures inthe baseline (2002) budget, with adjustment for projected debt servicing requirements. Forward estimates presented inthe MTEFwere used to guide the preparation of budget limits for the 2003 budgeting exercise, albeit with adjustments during the budget process. Furthermore, MEF has indicated that it intends to present indicative allocations for outer years, based on the MTEF, as part of the final budget documents, which would make the MTEF an integral part of the budgetingprocess. 4.24 So far, allocations have been largely determined "top down" and incrementally, reflecting the limited capacity at agency level to structure their budgets by program and prepare realistically costed forward estimates. However, both the health and education sectors have made a first attempt at "bottom up" costing and expenditure programming. This has entailed the formulation of Sectoral Expenditure Frameworks (SEFs) comprising policy objectives, forward estimates for key programs and, in the case of education, output and outcome targets that the sector expects to achieve. Education's SEF i s consistent with PAPsub-sectoral programestimates as reflected in the ESP. Inhealth, the SEF i s based on broad programmatic categories: provision of basic services, priority public health programs (incorporating national programs), provision of essential specialized services, and sector wide management activities. These categories are split into sub-component areas and the MTEFwill eventually permit the tracking of broad allocations. 4.25 Development of the MTEF will have to be integrated into the MEF's institutional framework and sector budgetlplanning cycle routines. The policy orientation of forward expenditure planning requires a range of new skills-in policy analysis, costing and programming. Experience from Africa suggest that MTEF's may be more effective where existing departments are complemented by new unitsinorder to provide the support necessary for implementation.62Within the MEF,the Economic Analysis and Forecasting Unit could fulfill this function. Across government, the MTEF support structure could be integrated with existing NPRS units and committees. Development of sector expenditure frameworks inthe non-priority sectors is likely to be a lengthy process. SEF's will have to be built on agreed sector policy frameworks and development strategies, and aligned with well-defined and costed programs. Notwithstanding the progress made through the NPRS exercise, many sectors are far from ready. Obviously, priority should be given to the poverty reduction priority sectors identified in the NPRS. However, some consideration should be given to the formulation of expenditure frameworks in those sectors where further savings may be generated, such as defense and security. Without such analysis it will prove difficult to implement a phased reduction in low priority sector without disruption. 4.26 Resource forecasting and in-year predictability infunds release will have to improve if the MTEFis to gain credibility. Ifthere is consistently significant deviation between agreed forward estimates for the priority sectors then agencies will quickly revert to annual budgeting and give scant attention to the preparation of SEFs. This requires highlevelpolitical commitment to the forward estimates, which entails broad discussion of the MTEF throughout the budget preparation cycle, prior to finalization of budget ceilings, and when budget allocations are finalized. It also means that there has to be a sound basis for projectingfuture resource envelopes. h4EF recognizes that its current models are inadequate and is strengthening its forecasting capability. Most importantly, however, the success of MTEF i s dependent on continued improvement inbudget execution. It would be difficult indeed to convince line ministries to take 61 "Medium Term Expenditure Framework,2003-2005" draft document, 12September 2002. LeHouerouandTaliercio (2002). Reforming Institutions to Improve Service Delivery I... 69 three year forward estimates seriously when they are concerned about whether they will receive any funding for the next three months. 4.27 The MTEF should serve as the basis for programming all cross-governmental expenditure policies. Integration of MTEF and the medium term wage framework (MTWF), providing for phased increases incivil service remuneration (see Chapter 5), i s critical. Failure to do so runs the risk of undermining fiscal discipline as well as the credibility of the MTEF. Similarly, the MTEF will have to begin to take into account the huge volume of off-budget external financing. Lack of information on actual and programmed disbursements at an institutional level, hinders the integration of external project financing. While improvements are made in reporting at a cross governmental level, Government can take the lead in guiding the external financing by providing indicative allocations by institution and key programs. Recommendations 9 Develop a strategy, based on recent experience, for moving toward program budgeting over the mediumterm. 9 Integrate the PIP into the MTEF. k Include specific maintenance programs in priority sector under PAP, together with oversight and reporting requirements 9 Introduce a formula based allocation mechanisms for road maintenance funding and develop a strategy for conducting financial and performance audits of road fund expenditures. 9 Register Government implementationagency and recipient of external assistanceinCDC donor financed project database. 9 Integrate the MTEF into MEF's institutional framework and existing NPRS/SEDP organizational structures. 9 Consolidate the mediumterm wage framework with the MTEF. 9 Provide indicative targets for external financing inMTEFfor 2005 budget process. IMPROVING BUDGETEXECUTION 4.28 The critical test of a public expenditure management system is its ability to deliver a predictable and timely flow of funds to the spending agencies, allowing flexibility in the application of resources while ensuring adequate control at a sectoral and aggregate level. Cambodia's system does not currently meet this test. The conventional budget execution system has been effective as a control mechanism, thereby contributing the fiscal and macro-economic stability in recent years, but has failed as a system of delivering resources to spending agencies, resulting in a lack of transparency and a high level of risk (fiduciary risk i s addressed in the following sections). Pilot reforminitiatives, operating in contravention of key controls laid out in the Organic Budget Law of 1993, have a patchy record in terms of improving resource delivery. Nevertheless, they have considerable support at the level of the spending agencies and have now expanded to cover a considerable share of total non-wage spending in the priority sectors. A unified budget execution system i s needed (at the subnational level as well; see Chapter 6). The section concludes by suggesting the design features of a unifiedsystem to be introducedgradually over the medium term. 70 Cambodia Integrated Fiduciary Assessment Public Expenditure Review Budgetary performance 4.29 The effectiveness of expenditure controls at an aggregate level is demonstrated by the Government's success inachieving a current balance of over 1.2 percent of GDPineach year since 1998. This has been critical inrestoringmacro-economic stability. Expenditurereports for 2000, 2001 and preliminary reports for 2002 demonstrate that aggregated expenditures have been kept within budget targets (see Annex A Table A17). Furthermore, execution rates have tended to be fairly close to budgeted levels for both recurrent and capital spending. Nevertheless, excess expenditures do occur, notably in2002 when expenditures salaries and social interventions led to significantly higher levels of expenditures than programmed owing to the implementation of the pay reform. 4.30 There is considerable variance in budgetary performance against initial allocations at the agency level, however, even though aggregate targets have been met (see Table 4.1). Central agencies tend to overspend. In 2000 eighteen of the thirty central agencies registered expenditures in excess initial allocations, eight of these had excess expenditures of over 110 percent of their initial expenditure allocations. Central agencies' outturns were much closer to programmed expenditures in 2001, though total excess expenditures were about 70 percent higher than the overspent institutions' initial allocations inboth 2000 and 2001. Table 4.1: Variance inBudget Execution,2000 and 2001 ExecutionIndicator 2000 2001 Central Provincial Central Provincial No. Budget Agencies, with 30 20 30 20 Execution>110% 8 1 2 Execution> 100% 10 1 3 3 Execution<90% 6 12 8 11 Execution 430% 2 1 1 Excess & Under spending as % Allocation Excess Spending 73.6 30.7 76.5 4.6 Under spending (7.5) (40.7) (5.9) (11.7) Source: Bank estimates based on MEF TOFE. 4.31 For agencies' provincial administrations, the problem is rather one of under- spending. In 2000, twelve of the twenty agencies with provincial administrations reported shortfalls against provincial administrations' budget allocations, amounting to 40 percent of initial allocations. Provincial execution performance improved markedly in 2001and 2002, but provinces' execution rates, the proportion of provincial administrations under-spending, and the overall level of under-spending were still worse than for central agencies. Implementation constraints are one factor contributing to the provinces' poorer execution performance, but this may also be also a consequence of the lower priority accorded to provincial governments in allocating available cash. Execution rates have tended to be lowest for provincial operating expenditures, since salaries and social interventions are generally given higher priority when funds are scarce. In contrast, execution rates for PAP have been significantly higher both in comparison to funding for provincial agencies' operating costs and inrelation to special program activities at the central level. 4.32 The large number of excess votes suggests that expenditure controls are implemented selectively. In-year budget alterations are made to suite changing priorities and to meet pressing obligations. The 2001 commune elections provide a good example-actual expenditures amounted to a reported US$ 10million, as compared with initial allocations of US$ Reforming Institutions to Improve Service Delivery I... 71 0.7 million-and a salutary warning as regards the impact of the national elections scheduled for July 2002. 4.33 It is noteworthy execution rates for the priority sectors have trailed those for the civil administration as a whole, despite the channeling of a substantial share of sector expenditures through streamlined disbursement mechanisms. This i s particularly true of provincial expenditures in the priority sectors, with the exception of education, where provincial expenditures are dominated by payroll costs. In an environment where there i s significant overspending at the agency level, the comparatively poor performance of the priority sectors i s telling. Predictabilityinspending: back-loadingand arrears 4.34 Budget execution tends to get off to a slow start, and proceeds erratically thereafter, with significant back-loading of expenditures in the last month of the year. Table 4.2 demonstrates the scale of this problem: in 2000 and 2001 around 30 percent of total recurrent expenditures, and over 40 percent of operating costs, were not posted until December, the last month in the financial year. Again, the level of back loading i s higher for health and education than for the civil administration as a whole, and i s significantly higher than for defense and security. In 2001, for example, although 92 percent of the budget provision for health was disbursed for the year as a whole, only 2 percent was disbursed in the first quarter of the year, 7 percent inthe second, 11percent in the third, and 73 percent in the last quarter. The situation in 2002 was largely unchanged, with just 8 percent of the allocation disbursed in the first quarter. The pattern ineducation-indeed across Government-is much the same. Table 4.2: Back-loading:Share of ExpenditurePostedinDecember,2000-2001 EconomidSector 2000 2001 Central Provincial Central Provincial Civil Recurrent 43.4 22.8 31.6 29.9 Ch. 10Salaries 11.6 19.4 14.1 23.7 Ch. 11: Operatingcosts 38.2 25.0 42.2 34.9 Ch. 13: Specific Program Activities 47.7 100.00 47.3 70.0 Capital Domestic Financing 10.8 18.2 Defense and Security 17.2 18.5 Education 63.2 21.1 34.7 30.2 Health 59.7 31.7 71.4 43.6 Agriculture 41.2 18.6 30.7 28.0 RuralDevelopment 31.5 22.1 13.2 27.4 Source: Bank estimates based on MEF TOFE. 4.35 Back-loadingis also a problem for salaries and wages. The coefficient of ~ariability~~ for salaries and allowances of central agencies has increased from 38 percent in 1999 to 52 percent in 2001, while variability i s higher still at provincial level, rising from 41 percent to 65 percent. Moreover, variability tends to be higher in for the priority sectors than it i s for the civil administration as a whole, particularly at the central level (at 122 percent for MOHin 2001, 297 percent for MOEYS and 151percent for MAFF).Figure4.1 shows the highvariation inmonthly salary payments inM O Hin2001. 63 Coefficientof variability i s definedas the ratio of the standarddeviationof monthlyreportedexpenditureson salaries twelfth of the totalper month-the coefficientof variability wouldbe zero. and allowances to the average. If the monthly reported expenditures were distributed evenly across the year - one 72 Cambodia Intearated Fiduciaw Assessment Public Emenditure Review Figure4.1: MonthlySalary PaymentsinMOH,2001 I 1 250,000 200,000 -.g150,000 -+- Centra average 100,000 --e Rovinca average Centralaamn stratson 50,000 11 Rovincial admnistraton 0 -50,000 I Month I 4.36 Back-loading at the commitment stage is accompanied by delays in the release of funds into the following budget year. In part, the end of year surge is due to clearance of payments of expenditures against the previous years budget during an extended complementary period-so the back loading i s actually worse than the data would suggest. However, the extended complementary period, together with delays in payments to suppliers, lead to a flatter aggregateexpenditure profile over the year. TOFEreports record expenditures on a modified cash basis (as mandates are authorized) rather than on a cash basis (when payments are made to suppliers). Figure 1compares the expenditure profile for 1999 to 2002 on a modified-cash and cash basis. Whereas modified-cash expenditures show a stop-start trend with a huge end of year surge, cash expenditures show a much flatter profile. There i s still an end of year peak and subsequent drop in expenditures in the early months, but cash expenditures tend to remain significantly higher than those recorded on a modified-cash basis throughout the first quarter as the previous year's mandates are processed and liquidated. As the back loading has increased in recent years, this complimentary period for the closure of previous years accounts has tended to extent-already beyond the one month foreseen inthe budget law-into the second quarter. This has had knock-on effects, since mandates from the previous year's budget take the available cash, delaying the liquidation of payments arising from the current exercise, so that in2002, the budget was operating in net arrears-cash payments exceeded commitments-until September as compared with March in2001 and February in2000. Figure 4.2: Expenditureson Modified-Cashand CashBasis, by month1999 to 2002 (CR million) I --Cash Bask 400.000 -- 300.000 2W.m 100.OOO 4.37 The delayed release of mandates has been accompanied by delays in payments to suppliers, leading to an accumulation of arrears. InSeptember 2002 Treasury reported arrears Reforming Institutions to Improve Service Delivery I... 73 to the private sector and autonomous public agencies of CR 147 billion (about US$ 37 million), equivalent to 7.8 percent of total spending in that year, of which CR 29 billion predated that budget year. However, the Treasury data are likely to underestimate the extent of the problem, since this covers only payment orders that have been charged against Treasury off-setting accounts. This does not capture arrears to many of the smaller companies inthe form of payment orders that they have not been able to cash. Nor does it capture arrears arisingfrom unauthorized commitments made by spending agencies. Discussions with line agencies suggest that agencies do take delivery of goods and services on credit, on the understandingthat payments will be made once authorized, inorder to bypass the prolonged authorization process. 4.38 Where there funds release is unpredictable, the budget ceases to be a useful planning tool, seriously undermining agency performance and expenditure control. Delays inthe approval of commitments and the releaseof fundsprevent timely implementationof agency work plans, undermining the effectiveness of spending in education, where school materials should be available at the beginning of the school year, or road maintenance, where funds should be available after the rainy season. Recourse to unauthorized commitments to by-pass these problems creates another series of perverse incentives, not least of which i s the tying of agencies to suppliers with whom they have credit arrangements, thereby hindering attempts to introduce a transparent procurement process.64Agencies have every incentive to withhold and underreport revenues, since this generatescashthat can easily be usedto finance expenditures. Constraints inbudget execution 4.39 Problems arise at three distinct points in the payments process: the initiation of commitment requestsby agencies; the approval of commitments; and the execution of payments. 4.40 At an agency level, the centralization of responsibility for budget preparation and monitoring and poor understandingof budget management procedures are the heart of the problem. Agencies encounter difficulties in initiating and programming payments due to: (a) late communication of budget allocations, since agencies only begin to prepare and clear mandates after the budget law i s passed by parliament and they have been notified of detailed allocations by MEF, usually in January; (b) poor expenditure planning, since agencies do not prepare systematic expenditure plans for budget execution, mandates for expenditures tend to be drawn up shortly before funds are required65;(c) batching expenditure mandates for submission to MEF particularly where, as in the case of payroll, the preparation and clearance of mandates require information from provinces; and (d) end of year expensing of budget balances, as agencies rushto submit mandates for any remaining budget balances, reflecting both inadequate expenditure planning and weak execution capacity at agency level. Moving to a program approach, in which agencies line departments are responsible for preparing budgets and monitoring their implementation, would help integrate budget and operational management. This i s a prerequisite for effective budget and procurement planning. 4.41 Excessive centralization of commitment authority at the MEF compounds these problems. Under Cambodia's centralized budget system, all commitments are subject to authorization by MEF financial controllers. The financial controllers fulfill a pre-audit function for the agency that they oversee, this entails verification that the pro forma invoice or requisition presented by the agency has the requisite agency authorizations, that it is reasonable in terms of cost, and that it includes adequatebudget coverage and the appropriate supporting documentation. See the CountryProcurementAssessment Report, forthcoming. @ 65 Moreover, agencies andMEF fail to take into account andprepare for seasonalpayments such as the bulk purchase of medical supplies, amounting to about CR 40 billion, half of health sector operating costs, and payments for school examinationsupervision (aroundCR 5 billion), which have a significantimpacton cashrequirements. 74 Cambodia Integrated Fiduciary Assessment Public Expenditure Review Although financial controllers can, in principle, clear commitments up to CR 20 million (about US$ 5,000), various endorsements are required for most payments, including that of the head of the DBFA, and sometimes including the minister. Inevitably this takes time: agencies complain of frequent delays at this stage, sometimes of weeks or months. It also creates some uncertainty as to whether commitments will be authorized, or releasedto the agency even if authorized. MEFmay decide to retain mandates when it believes that there will be insufficient cash to cover the commitment. 4.42 For the most part, however, delays in budget execution occur at the payment stage, leading to growth in arrears. After agencies confirm the receipt of goods and services, a payment order and supporting mandate are returned to the financial controller for liquidation, confirming the amount of the liability to the supplier. The controller verifies the payment order and authorizing payment, requiring several endorsements. Once authorized, the payment order i s submitted to Treasury for payment, or to MEF's Foreign Currency Unit if the payment i s in foreign exchange. At this point, payment orders may be retained by either MEF prior to authorization or by Treasury, where there are insufficient funds to cover the expenditure. Again, there are delays, and uncertainty as to whether payments will be effected and the agency i s left dependent on gatekeepers inMEFand Treasury. REFORMINITIATIVES PRIORITYACTION PROGRAMS - 4.43 Frustration with poor budget performance led MEF, MOH, and MOEYS to find ways of channeling budget funds to priority sectors that bypass the rigidities, delays, and uncertainties created by this over-centralized budget execution process. These pilot measures, the Accelerated District Development program (ADD) and the Priority Action Program (PAP), have forced some change in spending priorities through a change resistant budget execution system and have thereby ameliorated some of the associated problems. At the same time, they have also created new concerns, particularly as regards the adequacy of controls. 4.44 The ADD program was introduced in 1996 to overcome problems in getting budget funds to the district level. Participating districts physically collected cash from the Ministry of Health in Phnom Penh, by-passing the provincial governors and the provincial treasury system altogether. Operating funds (excluding wages and drugs), which were provided in a single block, could be spent according to district needs. Despite ADD'Sfocus on small disbursements, it introduced a package of modem budget execution principles to Cambodia for the first time. Advance payment via an imprest arrangement, and post-audit of spending rather than pre-audit by MEF,greatly reduced line itemspecificity inbudget allocations. Creationof budget management centers able to decide themselves how to allocate budget funds between the numerous categories of runningcosts marked a significant deconcentration of expenditure management responsibility closer to the field level. 4.45 A parallel reform, PAP, intended to channel funds directly to health and education spending units, followed in 2000. PAP sought to ensure that high priority activities gained access to their full budget allocation by: (a) removing PAP funding from the discretionary monthly cash allotment system, by providing 25 percent of appropriations for PAP automatically on a quarterly basis; (b) reducing line itemdetail in the budget for PAP allocations, thus giving greater management responsibility for the allocation of operating costs to spending units; (c) introducing budget management centers (BMCs)-including national programs, hospitals, provincial healthdepartments and operational districts for health, and school districts ineducation -in spendingagencies to manage their own budgets in accordancewith pre-approved plans; (d) replacingpre-audit of spending actions by post-audit, a function supposedto be carried out by the Reforming Institutions to Improve Service Delivery 1. .. 75 PDAFs for sub-national priority actionprograms on a quarterly b a d 6 ;and (e) creating dedicated accounts inprovincial treasuries (F'Ts) that were not subject to discretion over cashrelease. 4.46 Initially started as a pilot program, PAPhas expanded rapidly. PAPwas launched in MOH(seven provinces and somenational hospitals) andMOEYS (all schools inall provinces). It has since been extended to MAFF and MRD. Furthermore, the amount of funds channeled through Chapter 13 has increasedrapidly inrecent years. Chapter 13 accounted for just 1percent of M O H non-wage recurrent spending in 1997, rising to 9.6 percent in 2000 and then to 36 percent in 2001. Even larger increases are reported in education, where Chapter 13 accounted for 37 percent of non-wage recurrent spending in2001, rising to 67.5 percent inthe 2002 budget. 4.47 A recent addition to this multiplicity of budget execution procedures is the Commune/Sangkat Fund (CSF). The CSF was created to finance operating and investment costs of the recently elected commune councils, though it has entirely different expenditure management procedures compared with regular or PAP procedures. The provincial treasuries act as cashiers but authorization of expenditure i s solely with the commune chiefs. There i s no requirement of "pre-audit" and it i s not clear what other audit arrangements will be inplace (see Chapter 6). Assessingthe impact of reforms 4.48 At a systemic level, the impact of the reforms is profound, for the simple reason that they do not do not comply with the 1993 Organic Budget Law. While provision i s made for imprest accounts for small volumes of expenditure, there i s none for the continuous replenishment of advances, nor does the law provide for the delegation of prior review authority to line agency staff. 4.49 There i s some evidence that the special Chapter 13 programs have improved budgetary performance against annual allocations. However, this evidence is only consistent for the health sector, where the impact tends to be greater at the provincial level (see Table 4.3). For MOEYS, for which execution data can only be computed for central level allocations, the improvement i s less clear cut. What the annual execution data hides i s the erratic and back- loadedPAPdisbursements. For example, in 2001, delays inthe approval of the MOEYS program structure meant that 95 percent of PAPfunds were releasedinonly the last quarter. As aresult of delays in disbursement, 29 percent of M O H PAP allocations for 2001 were carried over for execution in the following year, together with 44 percent of the allocations for MOEYS at the central level. The disbursement pattern for 2002 shows similar backloading. Moreover, funds have been disbursed sporadically, often monthly, rather than follow a quarterly disbursement pattern. 2002 data also suggests that execution rates vary greatly between provinces - from 56 percent to 85 percent inMOEYS and from 61 percent to over 100percent inthe case of MOH. 4.50 There is no conclusive evidence that PAP and ADD have had a significant impact on agency performance and outcomes. A preliminary review of the performance of PAP versus non-PAP provinces in the health sector found that PAPprovinces were more successful on some counts (antenatal coverage), while on others (contacts per capita and rate of attended deliveries), non-PAP provinces fared better. A study of school progression in 2000 and 2001 in MOEYS indicated much lower repeat rates for students inPAPprovinces than non-PAP provinces in 2000 due to the additional funds available for remedial teaching, which freed up places for new students. For example, the reductioninfirst grade repeatersinthe 14non-PAPprovinces was less 66 Since PAP i s not an imprest account arrangement, there i s no provision for verification o f accounts before replenishment. Instead, mandates are released on the basis o f the spending agencies' quarterly cash plan, together with procurement documentation for the prior quarter. The post-audit i s supposed to entail a review o f these procurement documents. 76 Cambodia Integrated Fiduciary Assessment Public Expenditure Review than 8,000 students compared to about 50,000 inthe ten PAPprovinces. Inthe PAPprovinces the new first grade intake increased by about 34 percent (60,000 students) as compared to an increase of 7 percent (20,000) in the non-PAP provinces. MOEYS attributes the improved performance to the additional fundingprovidedto PAPprovinces for remedialteaching.67 Table 4.3: Chapter 13and 11BudgetExecution2000,2001andNovember2002 ~ ~ EconomicStructure 2000 2001 2002(November) Chap 13 Chap. 11 Chap. 13 Chap. 11 Chap 13 Chap. 11 Central Administration Education 62.0 48.1 77.7 50.0 28.5 Health 93.0 86.1 98.0 93.0 84.4 14.0 Agriculture 103.5 70.2 94.0 79.5 41.4 Provincial Administration Education 121.5 108.6 53.8 Health 66.5 97.8 82.6 86.6 54.6 Agriculture 86.4 89.3 59.6 4.5 1 IfPAPhasfailed to significantly improve budget execution, this islargely due to the fact that problems of cash management erratic, delayed and back-loaded disbursement - - persisted. However, it i s also clear that spending agencies have been slow to spend the funds, resulting in the diversion of some funds to faster disbursing agencies. In many instances local managers, who were used to a pre-audit system, in which responsibility for authorizing spending lay with others, were reluctant to assume responsibility for spending that would now be post- audited68.The rapid scaling up of PAPhas brought to light significant capacity building needs in the BMCs, in areas such as budget preparation, costing, accounting records and reports, financial control procedures, and control systems. For example, in MOEYS PAP has placed increased responsibility on the Finance Department to supervise, monitor and support the BMCs at the district and national levels. While the Finance Department has increased to 13 staff to monitor and provide support, they are learning by doing and operating with a manual system. Continued strengthening of finance departments' capacity, introducing accounting standards and computerization, developing controller capacity, and improving the capacity of the provinces and districts to maintain records of expenditure are necessaryfor improving PAPperformance. 4.52 Control and accountability for PAP expenditures can only assured where there are effective internaland external audits, yet post-audit capacity i s extremely limitedinthe priority ministries. In MOEYS there have been no systematic post-audits of PAP expenditure since the start of the program, although the MEF controllers at the provincial level reportedly undertake control exercises from time to time. InMOH only a summary statement of spending i s required by MEFbefore release of subsequent cash allotments (supporting documents for each purchase are retained by Provincial Health Departments). Moreover, checking of procurement documentation by financial controllers at the national level (and PDAF at the sub-national level) does not appear to be effectively implemented. 4.53 Although the impact of Chapter 13 reforms has not met expectations, this is due more to deficiencies in the cash management system rather than inherent problems of the PAP and ADD. Certainly, the consensus in the priority sectors is that PAP, while far from perfect, i s preferable to the conventional budget execution system. It has given agencies greater flexibility in the use of funds and it has managed to get funds down to the field level. Yet it is clearly unacceptable to operate multiple budget execution systems simultaneously, particularly 67MOEYS,PAP Basic Education2000and2001Update. 68CambodianHealthPromotionandPrimaryHealthCare Project,A Training Needs Assessment on PAP, April 2001. Reforming Institutions to Improve Service Delivery I. .. 77 where these operate at the margins of the law. As elimination of PAP i s not an option, the issue is the transformationof PAPinto a mainstreammechanismof budget execution. BEYONDPAP TOWARDUNIFIEDBUDGETEXECUTION - A SYSTEM 4.54 Over the medium to long-term, the objective is to have just one budget execution system in place for non-wage expenditures across the whole of Government. The system should facilitate the release and application of funds by spending agencies, while ensuring adequate control at the agency and aggregate level. Such a system will have many of the features of PAP, including delegated responsibility for expenditure control, a programmatic budget structure, post-audit of expenditures by oversight institutions, predictable release of funds, and flexibility over the composition of spending. The experience of PAP to date indicates that the development of adequate technical capacity at the agency level will be the a constraint in the rolling out the pilot initiatives, since capacity constraints continue to hinder the effective functioning of the new systems inthe priority ministries. The strategy must also take into account the risks of a undermining expenditure discipline by proceeding too fast with the delegation to line agencies, without adequate training or the development of supporting control systems at agency level. 4.55 In line with this gradualist approach, this section outlines a strategy that seeks to address some of the shortcomings of the conventional budget execution system, strengthen expenditure management within the existing PAP agencies, and provide a framework for gradually rolling-out the piloted systems across Government. Improving the ConventionalBudget Execution System 4.56 Provisions for the outposting of financial controllers in the organic law have only recently begun to be implemented on a pilot basis in MOH and MOEYS, following the recommendation in the 1999 World Bank PER. The out posting of financial controllers would make them more responsive to the needs of the spending agencies, provide mechanism for capacity buildingin financial management at the agency level, and allow them to develop a role in support of expenditure programming and planning as well as financial control (though care mustbe taken to ensure that procedures for out posted controllers do not contradict existing PAP procedures). Outposting i s most likely to work where the financial controller becomes accountable to the spending minister, while working under the technical supervision of the MEF, as a cadre of the MEF, though in Cambodia the controllers are still accountable to MEF. 4.57 Significant improvementsinperformancecould be made under conventional budget execution arrangements by simplifying the commitment and payment authorization process. Outposting will only have an impact if the financial controllers commitment authority for individual transactions i s increased significantly and the controllers can submit payment orders direct to Treasury and the Foreign Currency Unit. While MEF clearance for larger payments is sound practice, not least to ensure that appropriate procurement procedures are followed, continued centralization of all transactions processing would defeat the purpose of out posting. Communication of monthly or quarterly allocations, which would constitute the aggregate limit for commitments should be sufficient for the purposes of expenditure control, as long as MEFimmediately sanctions officials that fail to comply with these limits. 4.58 Further improvementswould depend on the development of information systems to support processing and commitment control. Delegation of authority to agencies is only practical where there i s timely information on commitments. Where financial controllers are out posted summary reports of mandates submitted, outstanding commitments, and mandates to be submitted in the next three months would suffice. Not only does this allow DBFA verify whether 78 Cambodia Integrated Fiduciary Assessment Public Expenditure Review commitments are within limits, it also provides information on future cash requirements and i s an essential part of any cashplanning system. 4.59 Budget classificationswill have to be improved. The present classifications, focusing on detailed line items, serve little purpose, reduce flexibility, and complicate accounting while contributing nothing to improved control. Since the introduction of new budget classifications requires considerable investments in training across the administration, and will take several years to implement, it i s wise to undertake a comprehensive restructuring of the budget classification and chart of accounts. Key features of the revised structure should include an economicAine item classification consistent with GFS standards and provision for prograddepartmental classification. Mainstreamingdelegatedbudget executionmechanisms 4.60 The measures proposed above would have a significant impact on the effectiveness of expenditure management systems under PAP as well, but additional measures are requiredto strengthen and deepenthe PAP reforms inthe priority ministries. 4.61 As a first step, financial management procedures should be standardized. In the absence of adequate guidance on internal financial operations for PAPfrom MEF,bothM O H and MOEYS have developed their own procedural manuals to guide the budget preparation and budget execution processes. While agencies should be encouraged to develop systems that meet their specific needs, basic procedures should be consistent across government if the integrity of the system i s to be maintained. Preparation of appropriate regulations by MEF i s a matter of urgency, as PAPi s rolled out to MAFFand MRD. 4.62 Computerized accounting systems are also required. Some progress was made by MOH in introducing Peachtree accounting software in the provincial health departments. Computerization of BMC's in MOEYS was rejected on account of the number of units and consequent cost, but reliance on manual accounting systems has hindered timely reporting and tracking on expenditures. Some intermediate level of computerization-between BMCs and provinces-may be viable. Standardization across government, while allowing some flexibility to meet agency needs, i s critical. MEF should take the lead in the identification of appropriate software and oversee implementationinpriority sectors. 4.63 Streamlining funds release would improve the efficiency of the execution system. Due to liquidity constraints, the quarterly release of cash advances didnot function properly. The current design (sub-decree 102) uses tranches to channel resources through PAP, but in practice the number of tranches per year has been quite small. While just-in-time cash releases based on monthly cash plans would be ideal, this would likely be difficult given limited capacity for planning and present liquidity constraints. Pragmatic solutions lie in formalizing the monthly release of funds, while seeking to ensure predictability through a quarterly cash allotment mechanism. It should be recognized that a cash "tranche" regime i s an extremely inefficient mechanism for budget execution for a government facing liquidity constraints, draining the Treasury of resources that could be used elsewhere. These constraints can be partially addressed by making properuse of the banking system for funds releaseand agency payments. 4.64 PAP should also be rolled out further at the agency level, but only where the responsible units can meet clearly defined readiness criteria. The first step in mainstreaming PAPis to roll-out these programs within the priority sectors. At presentjust seven of the twenty three provinces and municipalities operate PAPSin the MOH. In education, PAP operate across all provinces, but only 144 of the 185 education districts have a BMC. Extension of PAP inboth sectors should be phased, both geographically, using readiness criteria to guide the identification of new provinces and districts, and, as the basic architecture of the PAP system i s strengthened, Reforming Institutions to Improve Service Delivery I... 79 by transferring spending under Chapters 11 to Chapter 13. Institutional constraints can be addressed by applying readiness criteria, which define the minimumrequirements before PAP can be launched and for threshold levels of expenditure. These would include: agreement on an appropriate programstructure; placement of the financial controller and an adequately staffed and trained financial management unit; establishment and training of BMC staff; and computerization of accounting procedures. Rigorous application of the readiness criteria can be used as an incentive for agencies and-within agencies, provinces and districts-wishing to access PAP funds.69 . 4.65 Before rolling-out PAP beyond the priority sectors it would be wise to undertake a thorough evaluation of its impact so far. This would entail analysis of operational efficiency (especially in terms of timeliness of funds release), compliance (the adequacy of the control arrangements, through an independent audit), and effectiveness both as a tool for delivering resources to service delivery units (through expenditure tracking) and in improving agency performance (through assessment of performance indicators). Ideally, this assessment would be carried out before further budget support significantly increases the volume of funds channeled through this mechanism. Recommendations 9 Outpost financial controllers to MOH and MOEYS and prepare schedule for out posting to other priority ministries. 9 Define job-descriptions, institutional arrangements, and authorization procedures for out posted financial controllers. > Evaluate operational efficiency, compliance and performance of PAP before further expansion of programs. 9 Prepare standardized financial management procedures and supporting manuals for PAP, taking account of sectoral variations. 9 Review of options for progressive computerization of PAPoperations. 9 Require preparation of quarterly cash plans for Priority Action Plans and introduce monthly funds releasebasedon a quarterly allocation. 9 Define and apply readiness criteria for rolling out of PAP at agency and cross- governmental level. TREASURY OPERATIONSAND CASHMANAGEMENT 4.66 Cambodia's Organic Budget Law of 1993 provides for a consolidated fund and central payments system, managed by the National Treasury (NT) and its network of provincial treasuries. All revenues should be deposited with Treasury, unless otherwise provided for by law. Treasury should make all payments on behalf of government ministries and agencies, save for petty cash payments through imprest accounts. Payments should only be made on the basis of payment orders authorized by MEF. All payments above CR 2.5 million (about US$600) are supposedto be paid through the banking system. For this purpose, the NT maintains accounts with the National Bank of Cambodia (NBC), the Government's banking agent, together with sub-accounts inthe provinces where the NBC has delegations. 4.67 While the Treasury management system provided for inthe Organic Budget Law is extremely centralized, subject to problems of delay and gate keeping that characterize 69 Strengtheningof intemal andextemal control environment is also a priority (see last section). 80 Cambodia Integrated Fiduciary Assessment Public Expenditure Review earlier stages of the payments process, it constitutes a sound framework for cash management. The liquidity constraints that lie at the heart of the problems of poor budget execution, both through conventional and pilot mechanisms, can be traced back to the failure of Government institutions to comply with the law's requirements inkey respects. 4.68 However, some of the basic systems that should underpin any financial management system are completely absent from Cambodia's National Treasury. There are no standardsto govern treasury operations or public sector accounting, except some decrees and ad hoc circulars issued from time to time. Nor are there any requirements to audit the accounts. Given the lack of automation, voluminous accounting transactions are carried out by usinghand-held calculators or even manually. As in the case of most procedures governing financial operations in Cambodia, the focus of the NT remains on compliance and control through multiple regulations and procedures. The problem i s seriously compounded by the absence of regular internal or external audit. Some oversight functions are, however, performed by a number of Government unit through physical inspections of the line agencies. According to anecdotal evidence, this has also resulted in rent-seeking behavior at different levels of financial control, thus encouraging the agencies to flout established rules. Excessive controls seem only to have resulted in strengthening these systematic aberrations. There i s also a serious capacity problem of lack of basic skills, which i s reflected inthe constant revision of accounting data duringthe year. The present section focuses on these serious shortcomings of the NT and cash management methods and suggests measuresthat can be taken to bring these practices back inline with sound practice. Causesof liquidity constraints 4.69 Liquidity constraints are not caused by an uneven flow of revenues over the course of the year (Figure 4.3). At an aggregate level, domestic revenue flows appear to be fairly stable. Government has not had recourse to domestic bank financing to cover an aggregate financing deficit. Inrecent years, reported expenditure on a cash basis has tracked domestic revenues fairly closely duringthe early months of the year. This pattern breaks down at the end of the year, when expenditures are significantly higher than reported revenues, and, in2002, following a significant increase inrevenues during the first quarter, which allowed the Government to buildup stronger balances. However, over the year as a whole, the Government has sought to constrain expenditures inorder to deliver a positive current balance. Figure 4.3: Domestic Revenues and Treasury-Executed Expenditure (Cash Basis),2000-2002 300.000 -Domestic Revenues I -BExpcndiivrc C a i h Basis A A 100.000 4.70 The Government's cash resourcesare not managed from a single fund as provided for under the law, but are fragmented so that Treasury only has access to a small part of Government's total deposits. At end June 2002 about 10 percent of the RGC's financial assets were available to Treasury for the purposes of budget execution (liquid assets then amounted to Reforming Institutions to Improve Service Delivery I.. . 81 about 6.6 percent of the total Treasury executed budget, riel liquid assetsjust 2.8 percent). By the end of December, liquidassets had dropped to 5 percent of total assets (see Table 4.4). Table 4.4: Structure of Government Financial Deposits Structure of Deposits1 No. of 613Ql2QQ2 1213112002 Cash Holdings Accounts M.Riel % MRiel % MEFLiquidAssets 124,116 26.1 79,551 17.5 CentralTreasury Cash 3,301 0.7 3,686 0.8 Provincial TreasuryCash 21,281 4.5 6,634 1.5 ConsolidatedFundAccount 1 29,154 6.1 9,255 2.0 Other Riel Accounts 2 316 0.1 316 0.1 Dollar Accounts 6 70,063 14.7 59,660 13.2 EarmarkedBudgetary Support 259,521 54.5 264,740 58.4 Riel Accounts 7 104,978 22.1 153,899 33.9 Dollar Accounts 8 154,543 32.5 110,840 24.4 ProjectSpecial Accounts 38,312 8.1 70,7 17 15.6 RielAccounts Dollar Accounts 93 38,312 8.1 70,717 15.6 Ministry Special Accounts 53,831 11.3 38,631 8.5 Riel Accounts 10 7,501 1.6 9,034 2.0 Dollar Accounts 49 46,330 9.7 29,597 6.5 Total GovernmentDeposits 475,780 100.00 453,640 100.0 InRiel 166,532 35.0 182,826 40.3 InDollars 309,248 65.0 270,814 59.7 Source: Bank estimates based on MEF TOFE reportfor December 2002. 4.71 Agency revenue retention is a major cause of fragmentation. Untilrecently, tax and customs revenues were retained in separate accounts and released to Treasury periodically. Government has now established a Treasury Single Account (TSA), requiring large tax payers and customs revenuespayable inRiel to be paid directly into the Treasury's revenue accounts and for these to be swept into the TSA daily. Eightrevenue accounts have been closed. Nevertheless, line agencies-notably the Ministry of Commerce, Ministry of Post and Telecommunications, and MAFF-continue to receive and retain revenues, which are not necessarily fully reported to Treasury, and some of which are held in agency accounts at the NBC. Not only does this hinder the consolidation of available cash, it also undermines efforts to plan and monitor revenue yields for the purposes of forward cash planning. Further efforts are neededto consolidate revenues and establish an adequatereporting framework where there i s a legalbasis for agency retention. 4.72 Segregation of foreign currency and riel accounts is another source of fragmentation. Foreign currency revenues are surrendered to MEF, whose Foreign Currency Unit is responsible for managing foreign currency accounts and transactions. Deposits in these accounts constituted around half of the MEF's liquid assets at end of June 2002 and over three quarters of liquid assets in December (see Table 4.3). In principle, foreign currency receipts should be surrendered to the National Bank of Cambodia and the TSA credited with the local currency counter value. In practice, these accounts are managed separately, primarily for the purpose of making dollar currency transactions within Cambodia. Mandates and payment orders for foreign exchange transactions-mostly for capital expenditures-are processed through the financial controllers and submitted to the Foreign Currency Unit for payment. There are now transparent criteria for transfers from foreign currency to riel accounts, though such transfers are occasionally authorized when cashconstraints become acute. 4.73 By far the largest part of the Government's financial assets at any one point in time is constituted by external budgetary support-amounting to over half of total assets in June 82 Cambodia Intearated Fiduciarv Assessment Public Exvenditure Review and December 2002. These funds are deposited in special account and earmarked for specific purposes, rather than surrendered to the TSA for the general budget execution. Deposits are occasionally used to replenish Treasury, but they are not readily available, and certainly not managedas part of a comprehensive cashmanagementstrategy. On occasion Treasury is required to reimburse advances from the earmarked accounts when funds become available. 4.74 Problems of fragmentation are aggravated by the use of cash for nearly all payments, which results in a systemic weaknessthat open up avenues for corrupt practices. The exception are salary payments to the Senate and National Assembly and offsetting payments to large suppliers, whereby suppliers' accounts at Treasury are credited against future tax liabilities. The TSA i s not usedfor payments to suppliers by check or bank transfer. Instead, cash i s withdrawn by Treasury for payments to be made directly to creditors. Transfers to provinces are similarly made in cash, including those for the payment of salaries. This results inlarge cash balances being held by central Treasury and provincial treasuries: at the end of June and December 2002, cash holdings were of the same order of magnitude as deposits in the TSA. Additional deposits are held by spending agencies, especially in the priority sectors where the advances for PAP are drawn in cash and held by the BMCs, who then make payments to suppliers. Not only does the widespread use of cash greatly increase the risk of fraud, it makes it difficult to exercise basic controls, such as reconciliation - at the centralTreasury, for instance, in and out movements of cash are verified daily but not the stock - and leaves the cash management system hostage to the logistics of shipping large volumes of cash between NBC, treasuries, and BMCs. Inefficiencies will inevitably occur with idle cash holdings balances in some areas that are neededelsewhere. Implementation of PAPhas aggravated these problems, by requiring Treasury to mobilize large sums for quarterly advances. 4.75 Cash management is further hindered by the lack of short-term financing instruments for revenue and expenditure smoothing. The Organic Budget Law allows Government to borrow up to 10percent of revenues for in-year financing. However, Government has made a commitment not to recourse to domestic lending-though this commitment may waver as expenditure pressures mount in the run up to the 2003 elections. Government may also issue short-term borrowing by issuing Treasury Bills. So far, however, Treasury bills have only been issued for the purposes of bank capitalization but the market for such operations i s likely to be very limited. Instead, Government tends to rely on non-bank financing-ither informally through arrears-or, inthe case of investment projects, throughpre-financing arrangements. 4.76 Lastly, liquidity constraints have also been aggravated by poor communication between Treasury and the DBFA and lack of effective instruments for in-year expenditure programming. Inthe past, Treasury has allocated the available cash without close reference to the budget's priorities, often with direct intervention by the OPM and provincial governors determining how limited cash will be applied. MEF created a Cash Management Committee in early 2002 to address these concerns. Cash projections were prepared for central and provincial administrations by month for the whole of the year, and the committee has monitored implementation inrelation to the plan. However, the Committee has yet to assume an active role in scheduling the authorizationof mandates, thereby ensuring that commitments reflect projected cashbalancesand reducing the risk of arrears, and inprioritizing the payment of suppliers. This i s partly because, inthe absence of a systematic registrationof mandates in a database, it i s difficult to identify which mandates and the volume of commitments that are outstanding and are being processedthrough the system at any one time. Nor, given the limitations of revenue projections, does the Committee have access to reliable forward estimates of receipts. Reforming Institutions to Improve Service Delivery I... 83 Towarda strategyfor improvedcashmanagement 4.77 While increased use of the banking system must constitute the core of any program to strengthen cash management, this will have to be phased in gradually. Scant use i s made of the banking systemfor private sector transactions, particularly inthe provinces, partly because distrust of the banking system is deep-rooted, and has been aggravated by instability in the banking sector inrecent years, partly becausethe use of banks provides opportunities for another group of gatekeepers, and partly because of the sparse banking network outside the main urban centers. Similarly, while use of a single, national currency for Government transactions is highly desirable, this too is likely to encounter resistance from the private sector. Cambodia is a highly dollarized economy, and though riel may be used more widely in rural areas, the dollar still remains important as a store of wealth. 4.78 Within Government it is impractical to design reforms simply in terms of implementing the provisions of the 1993 Organic Budget Law. Introduction of PAPhas led to a delegation of the payment function from Treasury to the spending agency that it would prove difficult to reverse in the medium term. In this context, a program of reform for measures to strengthen cash management will have to address a range of institutional constraints over the medium-term. Such a program should focus on: (a) increasing the use of the banking sector for budget execution; (b) further integrating the Treasury system, especially with regard to foreign currency operations and revenues; (c) improving planning and prioritization incash management; (d)revampingearmarking practices; and (e) improving management information systems. 4.79 Making increased use of the banking system for budget execution transactions, including PAP, is a priority measure. MEF recently indicated that NBC will begin to make transfers to some provinces through its provincial branches from early 2003. This i s an important first step. It i s also important to begin using the banking systems for payments. This could be achieved by requiring that all payments above a certain level (the law states CR 2.5 million) be made by check or bank transfer. The extensive use of the banking systemby the World Bank and other donors projects demonstrates that this i s feasible, though NBC's capacity needs to be strengthened as well. PAP agencies should also begin to use the banking system for payments. This would bepossible by establishing sub-accountswith the NBC at central and provincial level, and by selective use of the private banking network (as the NBC does not have delegations in most provinces). Ideally, PAP accounts would function on the basis of a zero balance, cleared regularly, however, if this i s not feasible, monthly replenishment i s advised. 4.80 Further integration of the Treasury system is necessary to reduce the fragmentation that impedes effective cash management. Integrating foreign currency operations should start with the integration of information on foreign currency balances into the MEF's cash management plan. Over the medium-term, the Foreign Currency Unit and its functions should be transferred to Treasury. At the same time a progressively larger share of revenues should be paid through the banking system to TSA sub-accounts, and procedures for the timely surrender and reporting of agency revenuesneed to be clarified. 4.81 Improving forward planning and prioritization in cash management is another key reform measure. At the agency level, attention needs to turn to in-year programming of expenditures and procurement planning, so as to provide an informed basis for cash planning at the central level. Agency's should be required to report monthly on the commitment mandates submitted, mandates authorized, payments falling due, and mandates to be submitted in the next three months. For MEF, and the Cash Management Committee, forward cash plans should be regularly updated. These should provide broad guidance on Treasury' s prioritization of payments and identify specific payment orders to be expensed as a priority. PAP release procedures also need revision. As the share of expenditures passing through PAP increases, it will become 84 Cambodia Integrated Fiduciary Assessment Public Expenditure Review impractical to provide quarterly cash advances. Monthly advances are feasible and could actually improve predictability if accompanied by quarterly cash allotments that serve as a guide to the volume of monthly releases. 4.82 Given the legal requirement for consolidation of funds, earmarking should be strictly limited to those earmarked revenues provided for in legislation and by financing agreement. All other earmarking arrangements should be terminated. Additionally, the revenues and expenditures associatedwith authorized earmarking arrangements should be clearly identified inbudgetdocuments andthe TOFEreport. 4.83 Lastly, implementation of a computerized information system for Treasury operations and MEF commitment controls is needed. A simple accounting system would allow Treasury to compile basic accounting registers, such as cash books and ledgers, by automating existing manualprocedures. For MEF, while financial controllers are still basedinthe DBFA,information systemsfor tracking mandatesand commitments are critical, since this is the only way to provide information on outstanding obligations for the purposes of cashplanning. Recommendations > Define and apply threshold for national and provincial Treasury expenditure operations to be executed by check or bank transfer, and use banking system for all intergovernmental transfers. > Undertake an inventory of all revenue accounts and draw up a schedule for the elimination of revenue retention accounts that have nobasis inlaw. > Review of requirements for computerization of national and provincial Treasury operations. > Integrate information on foreign exchange balances in MEFlTreasury cash management plan. > Consolidate revenues into Treasury Single Account, other than those with legalbasis. > Require agencies to present commitments, outstanding commitments, and accounts payable intheir monthly financial reports. STRENGTHENINGTRANSPARENCYACCOUNTABILITYAND 4.84 Cambodia's budget law, which is complemented by a series of decrees and sub- decrees, establishes a sound legal framework for oversight of the public expenditure management system. Key features include requirements for: review and approval of the budget by the National Assembly; all revenues and expenditures of the government to be included and reported separately in a single budget document; submission of all revenues to Treasury; the separation of responsibilities of authorizing officers and payment officers; presentation and audit of annual financial statements. Budget documentation includes a detailed presentation of projectedrevenues and the proposed application of funds by central agencies and provinces. 4.85 However, in spite of a comprehensive legal framework, the multiplicity of controls, and the centralized nature of the expenditure management system, weak financial management practices still undermine transparency and accountability, posing serious fiduciary risks. A sound internal control environment requires not only an appropriate legal framework but also an organizational culture of compliance with the law, ethical and professional standards, a motivated and competent civil service familiar with rules and regulations that they are expected to apply, management systems that provide timely and accurate information for decision making, and appropriate oversight arrangements. While the legal framework i s sound, Reforming Institutions to Improve Service Delivery 1.. . 85 the Cambodian internal control environment lacks many of the above requirements. This section examines the underlying causes of these risks and the measures that may be taken, over the mediumterm, to strengthen expenditure and financial management. The internal control environment 4.86 In recent years, agencies have adopted a number of practices that facilitate their operations but do not comply with the principles or requirements of the Organic Budget Law. Some of these practices are officially endorsed, while others merely tolerated. Not only do these practices create fiduciary risks-notably regarding the large amount of cash circulating in the expenditure system, which makes it difficult to reconcile accounts, and the inadequate reporting of agency revenues-they undermine the rationale of rule-based expenditure management. Indeed, officials routinely and innovatively circumvent rules to facilitate budget execution. Use of informal credit arrangements with suppliers, for instance, allows agencies to procure goods and services in anticipation of future payment, at higher cost, but also in contravention of procurement rules. Agency officials exchange cash advances into dollars to make payments to suppliers, sometimes changing dollars back into riel when payments have to be made inlocal currency. 4.87 Despite the efforts of the MEF, notably through its training institute, staff are poorly trained, particularly in basic financial management skills, and unfamiliar with the regulations they are supposed to apply. There are no minimumqualification requirements for technical staff, nor are there recognized professional qualifications inthe field of accounting and financial management. A generation of skilled personnel that would occupy senior management posts today was lost duringthe PolPot regime. 4.88 Within this weak control environment, multiple controls have created a series of gatekeepers throughout the expenditure management system: in agencies, in MEF, in Treasury, and in the banks. The result i s increased probability of corruption. Gatekeepers have little incentive to facilitate operations: they may be risk averse, referring the smallest matters to their superiors for decision, causing excessive delay. In the worst cases they may be corrupt. There i s a general acceptance, based on widespread anecdotal evidence, that "facilitation" of transactions i s a common necessity (see Box 4.4). Box 4.4: A Traditionof Rent-seeking? A Cambodian newspaper recently reported that provinces and municipalities are required to make payments to officials inMEF, NT, the Ministry of Interior (MOI), as well as local banks, inorder to secure release of cash for municipal and provincial expenditures, and claimed to have to documentation to prove its allegations. "`It i s true that we have to pay people for their services to the relevant ministries,' said one official in the Kep municipality. `If we do not pay them, it is hardto get the funds or we receive our funds too late."' About 30 percent of appropriated funds are "`going into everyone's pocket inthe ministries even though the funds are for development,' the official said. `It's a tradition."' According to the press account, "Other provincial leaders confirmed that they too were ordered to pay a certain percentage of their development funds to the ministries." Inone case it was reported that payments were made throughout the system, to MEF (about 5 percent quarterly), M O I (about 5 percent), NT (about 2 percent), and PDAF (about 1.3 percent). Asked for clarification, MEFasserted that this account was neither truthful or accurate. Source: The CambodiaDaily, December 5, 2002,page I. 4.89 Government is aware of these problems. Chapter 11 expenditures are identified as being a particularly high risk, since it i s here that agencies have adopted informal practices to facilitate budget execution. Concerns have also been raised regardingPAPexpenditures, owing to the absence of effective pre- or post-audit reviews. However, rarely have civil servants been dismissed-one recent case was identified related to the dismissal of a personnel officer for payroll fraud-let alone subject to criminal proceedings. 86 Cambodia Integrated Fiduciary Assessment Public Expenditure Review 4.90 There are no immediate solutions to these systemic problems, though greater attention to the gravity of the fiduciary risk problem would likely galvanize efforts to improve transparency and accountability. MEF's strategy i s to gradually strengthen the internal control mechanisms at agency and central level. Significant progress can also be made by making the public expenditure process more transparent and improving downward accountability to clients. These technical improvements are an essential part of any solution. However, in the end, highlevel political commitment will be neededto identify and sanction cases of corruption, reward good performance, and gradually instill a culture of compliance and performance throughout the administration. Civil service reform i s a prerequisite for successful financial management reform(see Chapter 5). Public accounting 4.91 The Cambodian financial accounting system, as currently practiced, does not meet with any acceptable accounting standards and neither provides accurate nor timely information necessary for the financial management of a modern state. The system has evolved over the years and i s primarily based on French accounting principles. However, over time, other accounting conventions have been adopted either due ideological reasons (Soviet conventions during Khmer Rouge regime) or in response to changing fiscal management requirements. Over time the system has deteriorated owing to the successive crises and incompatible systems imposed on this basic model. The lack of attention given to the system over the years has led to an inability to respond to the increasing demands of budget management in particular, because basic systems, such as a standardized accounting methodology, do not exist. Moreover, the system i s manual; there i s an extremely low level of automation in the NT and PTs. The result is the systemthat provides late, incomplete, and inaccurate information. 4.92 Responsibility for accounting restswith the NT and its provincial branches. Treasury prepares trial balances daily, based on statements from the NBC and records of movements against accounts and cash transactions, together with monthly financial reports, for internal purposes. Both line agencies' Departments of Administration and Finance and MEF financial controllers maintain records of mandates issued and cleared, but neither the agencies nor h4EF keep a database which would allow them to track these transactions or consolidate commitment records. Agencies are also supposed keep record of expenditures and prepare monthly reports, though this practice does not appear to be systematic, indicating a lack of progress since the 1999 PER. 4.93 The chart of accounts, established in 1994 and subsequently adapted to accommodate changing expenditure management practices, does not provide a satisfactory framework for consolidating information on revenue and expenditure operations. These are used as the basis for a final balance, presented to the National Assembly about six to nine months after the end of the financial year. Weaknesses in the current system include: (a) poor coverage, as the Treasury system fails to capture large volumes of transactions (operations in foreign currency, those using retained revenues and earmarked funds, those arisingfrom PAPadvances); (b) lack of reconciliation of flows and stocks (moreover, reconciliation of banking information does not appear to be systematic), including between Treasury and line agency reports; (c) use of offsetting arrangements, whereby suppliers accounts in Treasury are credited with payments and debited for tax, which undermines transparency; (d) lack of detail (control accounts have few subsidiary ledgers for analysis); (e) poor articulation with the expenditure system (the accounting system records transactions at the payment stage, with little information on commitments); (f) poor integration of provincial treasuries, leading to delays and difficulties in consolidating accounts; (g) use of a largely manual system, generating considerable risk of computationalerror; (h) lack of a comprehensive accounting manual detailing policies and procedures supplementing Reforming Institutions to Improve Service Delivery I... 87 the diverse decrees and sub-decrees for expenditure management (as a result practices may vary among provincial treasuries); and (i) the focus on pre-audit, which leads to superficial verification of supporting evidence for the completed transactions. 4.94 Government's main financial report, the Table of State Financial Operations (TOFE), prepared monthly by MEF, is made available to an external audience, including the National Assembly and external partners, and will soon be posted on the MEF website. While the timeliness and detail of reporting i s far superior to that found in many low-income countries, the TOFE suffers from the weaknesses of the underlying Treasury accounting system and difficulties inreconciling the various sources. Moreover, as a tool for expenditure monitoring the TOFE i s deficient in three respects. First, presentation on a modified cash basis effectively captures commitments but can give a false impression of progress in budget execution. This is particularly true at the end of the year when mandates continue to be authorized and expensed duringan extended complimentary period. While adjustment is made at an aggregate level, it is not possible to determine actual expenditures against individual agencies. Second, there i s no explicit link between earmarked revenues and associated expenditures. Given the size and strategic importance of some of these earmarked funds, notably the FRMR, this is a significant omission. Third, the treatment of capital expenditures i s unclear, for the most part these are presented only at an aggregate level and, where pre-finance operations are used, the actual expenditures are significantly lower than statedon a modified cashbasis. 4.95 Given these shortcomings, the accounting system does not meet acceptable accounting standards and cannot provide assurance that public funds are used for the purposes intended. Under TCAF', the RGC has set up an Accounting Reform Group. Three priorities have been identified: reformulation of the chart of accounts, consolidation and reconciliation of accounting information, and the computerization of Treasury operations. So far, progress has been slow, as the group has become bogged down trying to resolve too many theoretical and practical issues simultaneously. A more pragmatic approach would involve focusing on basic procedures and information accuracy before tackling more complex issue of accounting principles and standards. 4.96 The main objective of the first phase of the reform program should be to produce accurate accounting information based on a unified chart of accounts. This will be easier to implement if there are no changes to the current budget classification, budget execution rules, or the payment system over the short-term. The following implementation program i s recommended: (a) definition of an acceptable set of accounting policies and procedures (based on international standards) to be used inpreparation of government financial statements (the priority i s to clarify the procedures to be applied following existing financial management practices); (b) preparation of a revised chart of accounts combining the current budget classification coding of revenue and expenditure and treasury accounts, and expanding the asset and liability accounts to include all government transactions, together with simple reporting formats"; (c) development of a simple accounting manual fully explaining the linkage of new chart of accounts to budget classifications and providing clear rules for coding of financial transactions, followed by adequate training. 4.97 Computerization is critical to the timely, accurate preparation of accounts based on transaction information and the ability to reconcile between accounting records. This is feasible using a fairly inexpensive, standalone PC accounting software package in both national and provincial treasury offices. Training needs should not be underestimated: staff in the national and provincial treasuries would need training in the use of the computerized accounting system. 'O Over the medium-term, attention shouldbe given to the reformulationof the chart of accounts, including a revised budgetclassificationconsistentwith GFS standards. 88 Cambodia Integrated Fiduciary Assessment Public Expenditure Review Fairly quickly it should be possible to generate all government financial statements from a computerized system. Although an integrated financial management system should be the ultimate goal, the technical competency of staff and current development of accounting rules and regulations do not warrant embarking on a complex system for the time being.71 Development of oversight arrangements 4.98 Though there has been progressinimproving the audit function, public oversight of government financial management remains extremely weak. Legislation establishing the National Audit Authority (NAA) was passed in March 2000, providing for an institutional framework, responsibilities, and competencies that are consistent with international standards. The NAA i s independent of the executive, reports to the National Assembly, enjoys financial and administrative autonomy, and i s authorized to determine the scope and methods of audits. An Auditor-General and two Deputy Auditor-Generals are appointed for five year terms by royal decree at the recommendation of the Government and approved by a two-third majority vote of the National Assembly. The current Auditor-General i s from the CPP, and the deputies from minority parties. The NAA carried out an audit of the year 2001 budget implementation (MEF's Budget Settlement accounts) as its first audit a~signment.~'The audit was completed in September 2002 and was submitted to the National assembly. However, the report has yet to be made public. 4.99 As a young organization with a clearly defined mission, NAA appears to be well motivated. However, NAA acknowledges that a considerable amount of on-the-job training and further investment in equipment, including computers, i s required for it to be fully effective.73 The ADB i s supporting with technical assistance in training. Under the original program put together by ADB, training in theory and practice spanning 12 and 16 months respectively were planned. However, due lack of technical assistance resources, this has not taken place. This should be considered a priority, since the audit function i s critical inensuring adequate oversight. 4.100 Inparallel with the creation of the external auditor, the RGChas sought to buildup its internal audit capacity. MEF's General InspectionDepartment undertakes compliance audits at the agency level, though it does not appear to have an audit program. Line ministries have also been encouraged to establish internal audit units. 4.101 Legislative oversight of public expenditure i s provided by the National Assembly's Finance and Banking Committee (FBC), and the Assembly as a whole, though the link between legislative oversight and greater accountability is weak. The FBC reviews the RGC's budget proposal and financial statements before submittingthem to the plenary session. The FBC has also received advice from an independent think-tank, the Cambodia Development Research Institute, on the Government's budget proposals. While there i s active debate on budgetary issues, the FBC's deliberations are neither public or published. The draft budget law are voted inplenary sessions after Government presentation and cursory debate. Though parliamentarians and senators have the authority to propose reallocations to the budget law, they rarely do so. These debates are reported in the local press. For the most part, however, press coverage has tended to 71 The World Bank has successfully assisted in installing stand alone PC-basedcomputerizedaccounting systems for project accountingin Cambodia. Cambodian technicians have used simple software (such as Peachtree)unaidedafter basic training. A more complexprojectuses Solomons software andi s operatedby Cambodian staff with the assistance of one expatriate accountant. Similar systems are now being piloted in seven RILGP provinces to manage CSF '*The accounts(thoughnot for the entirePT). NAA has adopted Intemational Standards of Auditing, as Cambodia does not yet have national standards, and hasjoined INTOSAI(IntemationalOrganizationof SupremeAudit Institutions). 73 The World Bank's peer review of the NAA recommendsthat the NAA not be used for the audit of Bank projects at this time. Reforming Institutions to Improve Service Delivery I... 89 focus on investigations of corruption in the public sector rather than broader expenditure policy issues. Moreover, inpast years the Assembly has been given too little time to review the budget indetail. Typically the draft budget law has beenpresented inmidNovember for adoptionby the end of December. The Assembly and the Government agreedthat in2003 the budget law will be presented in mid October. Providing adequate lead time i s necessary, though not sufficient, to enable the National Assembly to fulfill its oversight role. Strengthening the FBC's capacity- through technical assistance as well as hiring of staff with the requisite skills-to review the budget law would help it to holdthe Government to a higher standardof accountability. 4.102 While oversight arrangements are fairly sound, their effectivenessis severelylimited by capacity constraints. These will have to be addressed over the medium to long-term, by increasing both the numbers and in-service training of financial management personnel. In the meantime, the effectiveness of existing oversight arrangements could be greatly improved by ensuring that the NAA's audit reports are made available publicly. Moreover, attention should be given to the development of "bottom-up" oversight arrangements to complement formal oversight by elected representativesthrough NAA and the National Assembly. Recommendations 9 Define an acceptable set of accounting policies and procedures for existing expenditure execution procedures, including PAP. 9 Prepare a simple accounting manual, providing clear rule for coding of financial transactions and underlyingprocedures. 9 Prepare a modified chart of accounts covering all government financial operations (alongside PAPmanual). 9 Review requirements for computerization of National and Provincial Treasury accounts and financial reporting basedon a modified chart of accounts. 9 Develop and arrange financing for a medium-term capacity building program for the National Assembly's FBC. 9 Develop and arrange financing for amedium-term capacity buildingprogramfor NAA. 9 Ensure that the Government's financial statements are audited annually by the NAA. > Publishthe audited financial statements and make themaccessibleto the public. CONCLUSION 4.103 Itis clear that weaknesses inthe PEFMsystem have highcosts interms of allocative and operational efficiency while also creating unacceptably high levels of fiduciary risk to public funds. Indeed, incomparative perspective, Cambodia's PEFMsystem i s considered weak. The paradox i s that in spite of the multiplicity of controls and the centralized nature of the PEFM system, weak financial management practices pose serious fiduciary risks. The fundamental problems result from a weak control environment, severe weaknesses in Treasury operations, especially in the area of cash management, inadequacies in the public accounting system, and weak internal and external auditing and oversight. These problems are so serious that the Government's systemcannot be relied upon to expend resourcesinan accountable manner. These systemic weaknesses raise concerns about fiduciary malfeasance and increase the likelihood of corruption. Comprehensive PEFM reform, strengthened by effective accountability mechanisms, offer the way forward. 5. REFORMINGINSTITUTIONSTO IMPROVE SERVICE DELIVERY11: THE CAMBODIANCIVIL SERVICE "Yet, recognizing that adequate remuneration is central to the future of reforms and to the Civil Service becoming an effective provider of public services and a reliable development partner, the Government is systematically investigating ways and means to accelerate pay reform within the confines of macro- economic stability, sound fiscal management and sustainability. Every possible avenue will be explored and there is no panacea (emphasis added)." l4 INTRODUCTIONAND REFORMAGENDA 5.1 The Royal Government of Cambodia's (RGC) commitment to poverty reduction requires a public sector able to deliver a high quantity of high quality services across the national territory, including rural areas where many of the poor reside. The civil service is the institution that is charged with transforming Government and donor resources into poverty- reducing services. Civil servants are the link between resource delivery and service delivery. Yet given the serious challenges, widely acknowledged by the RGC, donors, NGOs, and Cambodian citizens, afflicting the civil service-low pay, low skills, and thus low capacity-comprehensive civil service reform will have to be accelerated in the short term andcarried out over the medium term if the Government's vision of poverty reduction is to become reality. Indeed, one of the principal risks to the NPRS i s the capacity of the civil service to deliver. Moreover, it i s clear that low public sector wages provide a breeding ground for corrupt practices. At the same time it i s apparent that low pay i s a leading cause of Cambodia's relatively poor standing on public sector performance among its peers. The World Bank's Country Policy and Institutional Assessment (CPIA) ranks Cambodia in the fourth lowest quintile among fellow low income countries on issues pertaining to public sector management and institutions, indicating significant room for improvement. 5.2 Though the problem of civil service reform is widely acknowledged and there is growing consensus on the path reform will have to take, the pace of reform will have to quicken and the scope of reformwill have to be expanded if service delivery is to improve in the medium term. The RGC has outlined its goals for civil service reform through the National Program for Administrative Reform (WAR),and its implementation strategy through the Strategy to Rationalize the Civil Service (SRCS), approved by the Council of Ministers in October 2001 (See Table 5.1). However, the RGC also recognizes the need for a more far- reaching approach. The NPRS rightly asserts that "every possible avenue" will have to be explored. Simply put, without comprehensive civil service reform, it will be difficult for the Government to implement its poverty reduction agenda. 74"Cambodia: NationalPoverty Reduction Strategy," December2002, p. 102-103. 91 92 Cambodia: Integrated Fiduciary AssessmentPublic Expenditure Review Table 5.1: The National Program for Administrative Reform and Strategy to Rationalize the Civil Service, 2002-2006 Goals Desired Outcomes 1. Accelerate pay andemploymentreform Improved Quality of PublicAdministration 2. Reorganize and redeploy staff to priority composition and distribution of the civilControl > Ability to document and the service 3. Improveservice delivery areas > Essential instruments to manage and motivate 4. Strengthencapacity to plan and managethe personnel administrative reform Empowermentof Civil Servantsto be More Effective and Productive > Increase civil service salaries in line with available resources > > Gradually decompress the wage structure Increase competitiveness of civil service remunerationand recruitment Promotion of Transparency,Responsibility and Ethical Behavior Strategies Improved Quality of Public Administration 1. Develop and operationalize a HumanResourceManagement InformationSystem. 2. Improve management of payroll, including automation. 3. Introduce new employee classification system in order to promote staff based on performance and seniority. 4. Meet needs of ministries relating to corporate (back office) services. Empowermentof Civil Servantsto be More Effective and Productive 5. Maintain workforce stability by increasing size of civil service by only 0.3% between 2002-2006. 6. Increase averagecivil service pay. 7. Introduce a system of allowances targeted to priority tasks and functions. 8. Implement Priority Mission Groups. 9. Undertake further analytical work to investigate ways and means to accelerate pay and employment reform, including an operational review/functional analysis and options for establishment control. 5.3 The nature of the constraints on civil service capacity requires a comprehensive approach. Such an approach would: (a) improve civil service pay, to attract and retain skilled staff who would work the required hours in the required locations, especially for high level management and priority sector staff; (b) rationalize civil service employment, to ensure that human resources are wisely deployed in highpriority sectors and administrative functions; and (c) strengthen civil service management, to guarantee that human resource expenditures are subject to controls (viz., on hiring and promotion) and linked more closely to the budget formulation process. Insome of these areas, such as workforce control, key measures such as the introduction of an establishment register, would allow the Government to build productively on its previous successes, while in other areas, notably pay and employment policy, the Govemment will be challenged to pushitself farther and faster than currently envisioned. 5.4 Given the gravity of the problem, the RGC committed in 2002 to an accelerated program of civil service reform that adequately addresses the problems of low pay, weak administrativecapacity, and the need for employment rationalization.The reformprogram- scheduled to be completed in early 2004 in time for the new Govemment to take concrete steps forward-will buildon the SRCS. The RGC, basedon its Governance ActionPlan (GAP)and the National Programof Administrative Reform (WAR),has madenotableprogress inthe following key areas: (a) civil service remuneration, in which average monthly pay increased to approximately US$ 33 as a result of the introduction of a new classification system in 2002 Reforming Institutions to Improve Service Delivery II:The Cambodia Civil Service 93 (according to 2002 budget data); and (b) documentation and control of the workforce by carrying out a census, completed in2000, that identified a reported 9,000 "ghost" workers and issuing civil service identification cards, by developing an automated payroll system (with automatic production of payroll lists), as part of the Human Resource Management Information System (HRMIS), and by solidifying the legal framework to cover all civil servants under the Common Statute. The RGC will also undertake additional initiatives in 2003, such as the Priority Mission Group (PMG) program.75Thus, while the RGC has undertaken important reforms, much more remains to be done. 5.5 The chapter begins by providing a brief overview of the Cambodian civil service in comparative context before addressing, in turn, (a) pay reform, (b) employment reform, (c) mediumtermpay and employment reform options, and (d) humanresource management. CAMBODIAIN INTERNATIONAL SIZE OF THE CIVIL PERSPECTIVE: THECOSTAND SERVICE76 5.6 Compared to its low income country peers, the Cambodian civil service is about average in terms of cost (the wage bill as a percentage of GDP, current expenditures, or revenues) and size (civil servants as a percentage of population). Inrecent years Cambodia's total wage bill, which includes defense and security expenditures as well as civil administration costs (the latter include the salaries of elected officials in the executive and legislative branches, as well as the judiciary) averaged 3.90 percent of GDP (1999-2002), including salaries, allowances, and supplements, declining from 4.12 percent of GDP in 1999 to 3.64 in 2002.'' The comparison with other low and lower middle income countries for which recent data is available shows that Cambodia falls toward mid-range andjust slightlybelow the set average of 4.5 percent (the low income country worldwide average i s 5.4 percent). As a percentage of current spending, Cambodia falls just slightly below the set average of 39 percent (Figure 5.1). Both relative cost indicators show that Cambodia's total spending on wages i s not inadequate by international standards, 5.7 In terms of the wage bill as a share of revenues, Cambodia, at 33 percent, falls slightly below the low and lower middle income country average of 37 percent (Figure 5.2). This would suggest that Cambodia is not an outlier either interms of over- or under-spending on the wage bill relative to resources; Cambodia falls in the mid-range where most of the countries in this sample cluster. However, it does suggest that for Cambodia to increase the wage bill significantly, it would have to increase revenues, both tax and non-tax, significantly. To move 7s Starting in 2003 the implementation of the Priority Mission Group (PMG) program will begin to address major service delivery bottlenecks, with the first PMGs targeting change initiatives underway. PMGs are groups of civil servants who will be tasked with focusing on identified "priority missions." To motivate these groups, they will be providedwith special monthly allowances, dependingon the staff`s category,rangingfrom CR 190,000to CR 520,000 (approximately US$48 to $130) for a pre-determined time period (approximately twelve to eighteen months). The Govemment plans to create PMGscomprisinga total of 1,000civil servants in 2003. PMGs are intendedas an interim measureto allow the Govemment to make progress in strategic areas of service delivery in the short term while more long term solutions are developed. It i s clear that PMGs will not, nor are they intended, to resolvethe problemof low remuneration. Rather, they are intendedto facilitate implementationof a select number of key reform initiatives, and this i s the standardagainst which they shouldbe evaluated. The intention of the PMG approach i s to substitute for the ad hoc arrangements employed by donors to supplement salaries (often in violation of donors' own policies and procedures). The basic concem that arises, given both scarce financial and administrative resources, is whether the PMG programwill yield positivenet benefits.The answer depends on the total cost to Govemment anddonors as well as the administrativeimprovementsyieldedby the program. Moreover, given scarce capacity, it may be a challengeto implement the program. The Govemment should evaluate the program in late 2004 to determine whether further resourcesshouldbe directedto PMGs or to other highpriority reforms. 76 To be clear, comparisonsof internationaldata such as these shouldbe used only to probe findings developedin the context of country-specificanalyticalwork, not to suggest specific policy reforms. 77 Actual expendituresfor years 1999-2001,budgetedfor 2002. 94 Cambodia: Integrated Fiduciarv Assessment Public Exvenditure Review from the low revenuesAow wage bill equilibrium to a highrevenueshigher wage bill equilibrium will require substantial improvements intax and customs administration (see Chapter 1). Figure 5.1: Total Central GovernmentWage Billas a Percentageof Current Expenditures:Low and MiddleIncomeCountries Upandadl Madagascar g/ Kenyaf/ C d o d i a d / Rwanda& Sudanc/ Costa Ricad/ Thailand d/ Y m , Rsp. c/ Haiti bl BurkinaFaso ai I 0% 10% 20% 30% 40% 50% 60% Source: IMF Government Finance Statistics,SIMA Database, Cambodian budget data. Years: a/ 1993, b/ 2000, c/ 1999,LU2001, e/ I992,j7 1996,g/ 1995. Figure5.2: Wage Billas aShareof Total Revenues, Low andLower MiddleIncomeCountriese' Uganda dl Kenya W Cambodia d/ Sudan cl Thrulandd/ Costa Rica d/ MadagascarW Burkina Faso a/ Yemen cl Haiti cl Rwanda a/ 070 10% 20% 30% 40% 50% 60% Source: IMF Government Finance Statistics, SIMA Database, Cambodia (Budget Law, TOFE, and IMF staff estimates). Years: a/1992-1993, b/ 19951996, d1999-2000, LU2001. e/Includes tax and non-tax revenues, as well as grants. 5.8 With respect to the share of civil service employment to the total population, Cambodia, at 1.4 percent in 2001, falls just below the low income country average of 1.6 Reforming Institutions to Improve Service Delivery II:The Cambodia Civil Service 95 percent (see Figure 5.3). This simple comparison suggests that Cambodia's i s not a bloated civil service. Butneither is it one of the leaner inthe developing Figure5.3: CivilServiceEmployment as a Share of Population: Low Income Countries 4% 0.0% 0.5% 1.O% 1.5% 2.0% 2.5% 3.0% 3.5% Source: Cross-NationalData on Employment and Wages,WorldBank, 2002; CambodiaBudget Law 2001. Years: a/ 1993,b/ 1994,c/ 1995,d/ 1996,e/ 1997,f/ 1998. Note: Thefigure for Cambodia is based on the latest available populationfigures. An estimate of 1.2% can be derivedby using UnitedNationspopulation projectionsfor 2003. 5.9 A closer look at staff allocation to the priority sectors-health and education-also shows that Cambodia falls below the low income country averages. Cambodia's education and health sector public employment account for 0.7 and 0.2 percent of the population, respectively, while the low income group averages (1996-2000) are 0.9 and 0.6 percent, respectively. The comparison suggests that Cambodia has room to increase employment inthese sectors, though, given the average size and cost of the civil service, also suggests that net reallocations would have to be financed by reductions inthe non-priority sectors 5.10 This comparative analysis suggests that: (a) Cambodia's civil service is average in terms of cost and size; (b) Cambodia, which collects a relatively low amount of revenues, spends a relatively low amount on wages; and (c) the total size of the workforce i s average, but the priority sectors (education and health) are under-staffed. The question is: Given poverty levels, severe budgetary constraints, and extremely low wages, can Cambodia afford to '*Interestingly,as a percentage of the total civil administration wage bill, Cambodia spends a significant amount on contractualworkers (over 3 percent in 2002). The amount paidto contractual workers (including "floating" personnel) in 2002 would finance 6,053 civil servants (approximately 3.7 percent of total civil servants) at the prevailing average wage rate. 96 Cambodia: Integrated Fiduciary Assessment Public Expenditure Review be average? The answer is "no." The chapter takes up this challenging question by analyzing current civil service pay levels, civil service size and distribution, and the human resource management system. Finally, the chapter suggests a number of options for moving forward with pay and employment reform, but notes that these must be taken up as parts of a whole inorder to make far-reaching progress a reality. CAR now has support to examine these reform options in greater detail as part of a comprehensive programof analytic work. IMPROVINGCIVIL SERVICEPAY: CURRENTPAY LEVELS EVIDENCE AND OF INADEQUACY "Government, donors and NGOs all acknowledge that low wages and low skill levels are contributing to the low morale and poor performance of the public sector. The NGO community reiterates that inadequate government salaries remain one of the major obstacles to the delivery of quality public services that could dramatically reduce poverty. NGOs working in Cambodia are particularly disillusioned by the lack of progress in this area, and especially the ones working in the health, education, legal and judicial sectors.'J' 5.11 The most pressing issue facing the Cambodian civil service is undoubtedly the low levelof pay for most civil servants, inrelationnot only to wage levelsoutside the service, but also to the cost of living. Indeed, it is widely acknowledged that civil service pay, which averaged approximately US$ 33 per month in 2002 after a considerable increase, i s inadequate." Comparing the average wage to per capita GDP finds that a Cambodian civil servant only makes slightly more than the annual per capita GDP. Cambodia's ratio of the average civil service wage to per capita GDP i s one of the lowest inthe region. The NGOForum, for example, has cited low salary levels as one of the "major obstacles'' to the delivery of poverty-reducing services. Moreover, public officials themselves view low salaries as the most important cause of corruption (World Bank, 2000, p. 27). For many reasons low levels of public sector remuneration require urgent attention. 5.12 With the exception of a very limited number of high level staff, most civil servants earn very little, in either absolute or relative terms (Table 5.1).8l Moreover, the compression ratios (the ratio of average pay of each category to the lowest category) are very low by international standards. The problem i s compounded by the fact that functional allowances have been set at their lowest level as the "experience and efficiency" criteria have not yet been developed by CAR. Implementing regulations should be developed as soon as possible to allow for use of the higher level allowances. A compression ratio of under two will not be sufficient to attract the necessary quantity of highly qualified candidates to civil service positions. If the Government cannot attract qualified, experienced high and mid-level managers and professionals to spearhead programdesign and implementation, the RGC's ability to deliver evenbasic services will be impaired. No doubt the Government's ability to attract these professionals is constrained by distortionary salary supplementation practices among donors and NGOs (see paragraph 5.40). But the fact remains that decompression of salaries must become a basic element of the ''NGO Forumon Cambodia, "A Rapid Comparison of the NGO Statement to the 2002 CG Meeting and the Second Draftofthe PRSPinCambodia," November2002, p.8. *'Thesalariesofjudges andprosecutorswere also increasedin2002. 81Approximately 700 officials received the special "AA" allowance, which i s based on rank. The special allowance, which pertains to five functions (directors of department, vice governors of provincelcity, departmental inspectors, deans, andgovemors of districts), substantiallyincreasestotalremunerationby the equivalentof betweenUS$ 130and US$205 per month. The Government's aim in offering the AA allowance has been to begin to decompress the remunerationstructure by increasingrewards to high-level administrativeofficials. However,one problemencountered has beensomewhatunclear criteria for decidingto whom to grant the "AA" allowance. Inpracticeit appears that some "acting" officials, who do not normally holdthe particularpost full time, have beengiventhe allowance. This problem stems inpart from the lack of an effectiveestablishmentcontrolmechanismthat mapsindividuals to approvedposts. ReforminP Institutions to Imurove Service Deliverv II:The Cambodia Civil Service 97 Government's reform program. The precise content of the decompression plan should be determined after further study, but it i s clear that significant decompression needsto take place. Table 5.1: Median Monthly Remuneration, 2002 (US$) Category Level of Education US$/month Compression Ratio A Secondary school years 40 1.9 B Secondary school ++ 42 years 32 1.5 C Secondary school 26 1.2 D Other 21 1.o Source: CAR Note: Compression ratiosfor each catego8y reflect the mean salary in that category as compared with the mean salary of Category D civil servants. 5.13 The Government's Strategy to Rationalize the Civil Service (SRCS) is based on a medium term wage bill framework (MTWF) for the period 2002-2006.83The MTWF would The end result - though not the ultimate objective - of the existing MTWF i s to increase the allow the average civil service wage to be increased from US$ 28 in 2002 to US$ 51.5 in2006.84 average civil service wage to that of the current minimum wage in the garment industry. However, the average wage in the garment industry, as in other parts of the private sector, i s much higher than the minimumwage. Table 5.2 shows that average wages data across the private sector are much higher than the average civil service wage. For example, a hotel manager would earn about 60 percent more than a senior public sector manager. Comparisons inthe same sector also show the low absolute level of wages. Anecdotal evidence suggests that Englishas a second language teachers earn about US$ 12-15 per hour, as compared to US$0.25 per hour for a public school teacher (assuming an average wage of US$40per month and a 40 hour workweek). Table 5.2: Private Sector Monthly Wage Data, 1999 CR (`000) US$ Cigarettes 318.2 81.6 Textiles 332.7 85.3 Hotels 248.6 63.7 Construction 267.7 68.6 Memoitem: Exchange Rate 3,900 Source: Cited in Abrahart (2000). 5.14 Evidence from the line ministries also confirms that civil service remuneration is much too low (see Box 5.1 for evidence on approximations of market rates in the health sector). At the ministry level several different mechanisms are used to increase de facto remuneration. The problem i s that thoughthese mechanisms respond to an urgent problem, they may do so in an ad hoc way that creates distortions in the system. For example, M O H allows health establishments to allocate 49 percent of user charges collected to staff remuneration, which, however, results in wide regional differentials in civil service pay. Moreover, given that user '*Th i s follows the OECD approachof comparingmeans rather than extremes, which ensures that ahandful of salaries will not dramatically skew the ratio. All compression ratio approaches can be misleading if there are significant monetary allowancesnot capturedinthe calculations. s3 CARS, "Civil Service Rationalization Strategy: Medium Term Wage Projection, 2002-2006," 27 June 2002. It shouldbenotedthat the macroeconomicframeworkhas sincebeenrevised. 84 The MTWF estimates the average monthly wage at US$ 28 in 2002, though actual expenditures put the average closer to US$33 (TOFEreport, April 2003). 98 Cambodia: Integrated Fiduciary Assessment Public Expenditure Review charge revenues are much higher in referral hospitals as compared to district health centers, the MOH scheme further biases incentives in favor of working in urban as opposed to rural areas (and biases service delivery toward curative as opposed to preventive services). M O H also uses per diems allowed through the PAP system to increase pay. In 2002 MAFF used PAP to finance staff remuneration for its forest replanting program. MOEYS has created a special program in PAP ("PAP 1: Education service efficiency," budgeted at CR 11,200 million, or approximately US$ 2.8 million in 2002) to finance redeployment of staff and increased remuneration, including financial incentives to teach in remote and difficult areas and for managementAeadership functions. Another PAPprogram (2) finances additional pay for remedial instruction (allowances for double-shift and multi-grade teaching continue to be paid out of Chapter 10).These examples make clear the desperation at the line ministry Box 5.1: Triplingor Quadrupling?A HealthSector "Experiment" Shows that PublicSector Wages Needto beIncreasedby Multiplesof Current Levels The Ministryof Health has engaged in several innovative experiences with contracting-in and - Jut, as well as other pilot reforms, of health service delivery. In a number of different provinces the managementof district health centers has been turned over to third parties, mostly NGOs, operating under a number of managerial arrangements. Under the contracting-in arrangement the contractor receives funds to provide technical support to the operational district (OD), has management responsibility for the OD, but has limited administrative authority over M O H staff. While personnel and O&M costs for the OD remain the responsibility of the ministry, the contractor receives an annual supplement of US$0.25 per capita, over which he has complete control. Under the contracting-out arrangement the contractors were given full responsibility, including over the budget and staff, for the OD (staff became directly employed by the contractor). Included in this set o f pilot projects were a number o f control districts; ODs were randomly assigned to the contracting in, contracting out, and control groups. A subsequent study found that the ODs operating under the contracting-out arrangements registered rapid increases in the quantity of services provided to clients as compared with the contracting-in and control ODs. Not surprisingly, one of the measures that most o f the contractors instituted was salary reform. The data show the range of salaries paid by contractors (CO and CI) to their public health staffs (in US$, monthly average): Pereang (70-180), Ang Roka (77-160), Memot (80-200), Cheung Prey (40-loo), and Kirivong(40-80). Taking the average monthly wage as US$ 28 (May 2002), it i s apparent that contractors felt they needed to pay multiples, ranging from about 1.5 in the lowest cases to 7 in the highest, of that wage in order to attain results. The average of the midpoints of these ranges is US$ 102, indicating a multiple of over 3.5 with respect to the average public sector wage. According to a Medicam report, "The provision of incentives by the NGO has been an approach applied by all NGOs. The NGOs consider this as the most important necessarycondition for effectively running district health services (p. 20)." Source: "Contracting and Similar Experiences in Health in Cambodia, Pieter FeenstrdMedicam, December 2001. 5.15 To resolve the demand issue the Government has committed to undertaking a labor market survey to determine the remuneration needed to be offered by the Government to provide the incentives needed to recruit and retain well-qualified staff. The labor market analysis should include a comparator pay survey, attitudinal surveys among staff and potential recruits, and studies of private sector as well as NGO markets. The results of the analytical work, and the Government's subsequent policy, will likely focus on targeting pay increases to specific cohorts of civil servants inhighpriority areas, as well as to higher level management. Thus, labor market information is urgently required as an inputto a revised MTWF. Average monthlysalaries at the NAA are also reportedto be muchhigher (aroundUS$170). Reforming Institutions to Improve Service Delivery II:The Cambodia Civil Service 99 Recommendations P Decompress civil service salaries starting with high level managers(Category A). The precise content of the decompression plan should be determined inconjunction with other variables and after further study. P Increase civil service remuneration significantly inthe short to mediumterm. The pay strategy should be basedon the forthcoming survey of labor market conditions. RATIONALIZING CIVIL SERVICE EMPLOYMENT: CURRENT THE SIZE AND DISTRIBUTIONOFTHE CIVILSERVICE 5.16 Both the size and allocation of the civil service act as constraints on efficient service delivery. On the one hand, the size of the civil service has been increasing, though unevenly, since the mid-1990s. On the other hand there i s some evidence that the deployment of civil servants i s problematic from the perspective of delivery of high priority poverty-reducing services. Large shares of civil servants are deployed in low priority sectors and in some sectors inadequate numbers of civil servants are deployed in the provinces. Moreover, only a small number of civil servants are women. Rationalizing civil service employment could increase resources for the priority sectors and promote greater efficiency inpersonnel management. 5.17 The size of the civil service has been increasing since the mid-l990s, though data sources are inconsistent. CAR'S census of the civil service recorded a total of 163,000 employees in 2001. According to the 2003 budget law, the size of the civil service now stands at 166,872 (this figure does not include ministers and other political appointees, which number 906 in 2003). According to the budget laws, the size of the civil service appears to have increased significantly since 1994, from 148,353 to 167,778, representing an increase of 13 percent. Rather than an actual increase in hiring, however, the increase of 6.6 percent in 2000 seems to correspond to a readjustment of the figures after an extension of the retirement age from 55 to 60. The explanation for the increase of 2.2 percent in2003 i s unclear, though it i s possible that some of the increase could be due to the requirement to post clerks to each of the newly elected commune councils. 5.18 The fact that the RGC does not use one set of official numbers when counting its civil servants creates confusion and undermines robust analysis. With the completion of the census, the budget law data should have been based on CAR'Sdata. It i s unclear why this has not happened, though it seems that CAR'Sdatabase has not been utilized outside of CAR. Granting wider governmental access to the database would improve the quality of policy analysis and i s necessary to ensure transparency. The accomplishments of the civil service census have been somewhat obscured by the lack of available data. For example, though it has been reported that the recent census and the automation of the payroll led to the removal of over 9,000 ghost workers, this reduction i s not apparent either in the budget law or the CAR figures, nor i s it visible inthe difference between the two sets of figures (see Table 5.3).86Greater transparency i s necessary both for showcasing the Government's accomplishments and for improving the management of the workforce. It i s imperative for the Government to use a single, unified database, and for MEF to have complete access to CAR'S database for budget preparation purposes, and for the line ministries to have ("read only") access to the database for personnel and budget managementpurposes. 5.19 As indicated by CAR's projections for 2003-2006 the size of the civil service is programmed to decrease slightly. The SRCS calls for the recruitment of approximately 22,875 new civil servants (20,075 in the education sector), attrition of 9,025, and departure (contracting 86CARS has committed to providingthis data, however it has not yet beenmade available. 100 Cambodia: Integrated Fiducian,Assessment Public Exvenditure Review out, privatization, and voluntary departure) of 15,750 (12,550 in the non-education sectors) over the period 2003-2006. However, it is unclear how the increase in staffing in 2003 will affect CAR'Smedium term scenario. The SRCS scenario will have to be updated to reflect the actual staffing situation. The apparent lack of coordinationbetween CAR and MEF, as well as other line ministries, undermines attempts to plan over the medium term (see section on MTEF below), though efforts to improve coordination are underway. Table 5.3: Historical Size of the Civil Service by Source b' ` Budget Law C A R 1994 148,353 n/a 1995 143,855 n/a 1996 147,086 n/a 1997 153,372 n/a 1998 156,731 161,466 1999 155,233 162,772 2000 165,539 162,991 2001 164,397 162,969 2002 164,219 165,369 2003 167,778 166,381 2004 n/a 165,953 2005 n/a 164,595 2006 n/a 163,469 2004-2006figures based on CARprojections. b/Budget law data includepolitical appointees and contract and jloating stafi while CAR data do not. Distributionof civil servants by category, function, region, gender, and age 5.20 There is not a single, updated published source of information on the Cambodian civil service. Information must be pieced together from the budget laws, CAR database, and old State Secretary of the Civil Service (SSCS) reports. Information on grade, functiodministry, and center/province i s available in the budget law. Additional information i s sometimes available from CAR on an ad hoc basis. The quality of the SSCS 2000 report is impressive, and the RGC should take steps to re-introduce it on a priority basis with updated information from the CAR database. This would require greater coordination between CAR and the SSCS, as well as increased capacity within CAR. 5.21 The majority of civil servants (40 percent) pertain to category C, which is reserved for those with a secondary education (see Table 5.4). Nearly two-thirds of the civil service have a secondary educational level or lower. Only 11percent of civil servants have completed college, though some 24 percent have some university training. The low educational level of the civil service, a legacy of recent Cambodian history, places a serious constraint on the quality of government performance broadly and on the delivery of essential services in particular. This capacity issue can only be resolvedover the longterm. Reforming Institutions to Improve Service Delivery II:The Cambodia Civil Service 101 Table 5.4: Civil Service Numbers, 2003 Category Education Non-Education Total % A 6,747 12,061 18,808 11% B 21,713 18,203 39,916 24% C 51,601 15,961 67,562 40% D 4,395 27,324 31,719 19% Other aJ 606 8,261 8,867 5% Total 85,062 81,810 166,872 100% Source: 2003 Budget Law.87 "Includes personnel on term contracts, "jloating," and on leave withoutpay. 5.22 Just over half of total state employeesserve incore (central) government functions, and of these, the vast majority work indefenseand public order:* Social development, at 39 percent in 2003, accounts for the second highest share of employees and six ministries, while economic services account for only 9 percent of government staff, though it encompasses ten ministries. At first glance, given the need to increase the share of civil servants working in social services (see below), and given the already low share of civil servants ineconomic sector services (viz., road transport and agriculture), any reduction and/or reallocation would appear to have to come from the core government function. This hypothesis will be probed further inthe functional analysis study to be undertakenby the RGC. 5.23 The education sector commands by far the largest share of civil servants (that is, excluding defense and security). In2003 education is programmed to account for just over half of all civil servants. By CAR'Sestimates the share to education will have to increaseto 61 percent by 2006 to meet the sector's needs. Given a very slight decrease in primary school enrolment from 2003-2005 and a large estimated increase in secondary school enrolment, from about 740,000 in 2003 to over 1,100,000 in 2005, and given pupil-teacher ratio objectives of 45 at the primary and 30 at the secondary levels in 2005, further increases in the sector are required. MOEYS estimates the need for a total of 109,000 teachers and staff by 2005. Holding the total number of civil servants constant at CAR'Sprogrammed figure of 163,469, the sector would account for nearly 67 percent of the total. The reallocation toward education will thus require further reductioninthe size of the non-education workforce. 5.24 The vast majority of civil servants, approximately 79 percent, are based in the provinces. MOEYS and the MOH account for 62 percent and 11percent, respectively, of these provincial staff. The Ministry of Interior accounts for the next largest share, at 7 percent, while MAFF accounts for 4 percent. Of the central ministry-based staff in PhnomPenh, MEF, at 12 percent, accounts for the largest share, followed by MOH at just under 12 percent and MAFF at 11percent. The Council of Ministersaccountsfor 6 percent while the Ministryof Culture and Art accounts for 5 percent of the total. See Annex A Table 5.4 for a breakdown by ministry and location. 5.25 These aggregate data, however, mask some of the problems surrounding the geographic distribution of civil servants. Anecdotal evidence indicates that low pay and the absence of "moonlighting" opportunities inthe provinces mean that civil servants prefer to work inPhnomPenh and the provincial capitals. The result is an oversupply of civil servants in urban areas and inadequate service delivery in the provinces, where most Cambodians live. The problem i s more acute for the higher professional civil servants, which results in a skewed distribution of skills and has a negative impact on the quality of service delivery. MOH, for *'Datainclude "ItisunclearwhetherthesedatamatchthoseintheHRMIS. defenseforces, but not police, for which data were unavailable. 102 Cambodia: Integrated Fiduciary Assessment Public Expenditure Review example, reports difficulties in maintaining its presence of doctors evenly across the provinces, which has resulted in differentials in the doctor-patient ratio by province, and MOEYS has serious problems attracting teachers to rural areas. The surplus staff in the urban areas tend to work in administrative rather than operational positions, further reducing their usefulness. In the education sector, for example, though about 20 percent of the total staff are in non-teaching positions, which is not high, MOEYS suffers from a high proportion of administrative staff at the level of secondary education. In primary education, the teaching: non-teaching staff ratio averages 7:1, compared with a very low 2.5:l in secondary education, where clearly efficiency gains are possible. This i s even more unfortunate inthe context of an overall shortage of teachers. At the same time, a number of ministries have large percentages of their workforces basedinthe capital, when, given the nature of their work, one would expect the preponderant share of staff to be located in the provinces: MOWRAM (55 percent), Ministry of Environment (56 percent), Ministry of Post and Telecommunication (65 percent), and MAFF (41 percent, where the situation i s even worse at district than at provincial level, for example in terms of agricultural engineers and agronomists). It i s impossible to say, a priori, whether these staff allocations make sense, but these figures c o n f m the need for a functional analysis to improve the efficiency and effectiveness of these scarce resources. 5.26 Gender distribution of the civil service is also uneven, with a nearly 70 percent male workforce and the majority of female staff aged between 18 and 35 (Figure 5.4). The Government has committed, as part of the Millennium Development Goals (MDG), to increase the percentage of women in decision-making positions within the government. Targets for 2008 have been set at 12.5% for women with the rank of Minister, 17% for Secretaries of State and 16.7% for Under Secretaries of State (the corresponding figures for 2003 are 8.7%, 8%, and 8%, respe~tively).~~address this concern, MOWVA has suggestedthat an affirmative action policy To for women may be necessary. It i s recognized that, at least in part, the problem stems from the fact that many fewer girls and young women complete secondary school and university, as compared to their male counterparts, making capacity buildingan attractive option for redressing inequality. However, it has also been suggested that deep-seated prejudices about the nature of appropriate work for women and concern about women's roles in society also have an important influence. A study of obstaclesto greater gender equity inthe civil service should be conducted to provide the necessary analytical underpinnings for moving forward with policies that promote greater gender equity. Figure 5.4: Age and Gender Distributionof the Civil Service in2000 Source: State Secretary of the Civil Service 89 Targets agreedJune2003 by MOWVA. Reforming Institutions to Improve Service Delivery 11: The Cambodia Civil Service 103 5.27 This section on employment and deployment suggests that: (a) reallocation of human resourcesto the priority sectors is necessary, given the Government's commitment to poverty reduction; (b) some compensated departure is in order both in the non-priority sectors and of lesser skilled workers; and (c) redeployment from center to province is necessary in some sectors. Additional options such as contracting in, contracting out and autonomization should be considered through a study of alternative service delivery (ASD) mechanisms begun inmid-2003. A functional analysis/operational review would be necessary to ascertain what the overall level of employment ought to be, where exactly the sectoral inefficiencies are, and what composition of public sector employment makes the most sense for Cambodia. The Government has agreed to undertake a functional review," which should be sure to cover the appropriate institutional arrangements, organization, processes, and staffing of government functions inorder to improve efficiency and effectiveness inservice delivery. Recommendations > Develop fully an employment strategy to (1) redeploy civil servants to the priority sectors, from central to provincial offices, and from staff to line functions; and (2) provide for compensated departure in the non-priority sectors and of lesser skilled workers. This strategy should be based on a functional review that covers the appropriate institutional arrangements, organization, processes, and staffing of government functions (based on analytical work to be carried out inmid-2003). P Develop a unified, consistent human resource database, linked to the budget formulation > process. Re-introduce the SSCS civil service annual report on a priority basis with updated information from the CAR. P Conduct a study of factors, such as education levels, social mores, and actual civil service practices in selecting and promoting women to provide the necessary analytical underpinnings for moving forward with policies that promote greater gender equity inthe civil service. P A Y LEVELS THEMEDIUM TERMW A G E FRAMEWORK AND 5.28 The previous two sections make the case that civil service remuneration needs to be increased and decompressed, and that employment needs to be rationalized by reallocation and compensated retrenchment. The purpose of this section is to elucidate possible routes to comprehensive pay and employment reform. The section suggests that strategies will have to include increased resources for civil service wages inboth absolute terms (that is, by increasing the total amount of resources available for recurrent expenditures) and in relative terms (that is, by reallocatingresources from other uses to fund a higher civil service wage bill and by reducing the size of the civil service), given resource constraints. The section also argues that Government will also have to marshal additional resources from donors to fund an increase in the wage bill. The simulations conducted inthis section are meant to be illustrative and not to suggest specific policy targets. 5.29 Inthe recent pastthe Government hasattemptedto containthe total wage billto be less than 40 percent of current spending, as recommended by the IMF. The total wage bill (Chapter 10) has declined steadily since 1998 as a percentage of recurrent expenditures, both in terms of the budget laws and actual execution, even as actual expenditures in riel have risen (in nominal terms) from 1998-2001 (see Table 5.5). This i s due in part to the combination of a ceiling on the wage bill with an increase in O&M spending in the priority sectors. The defense CARS, "Operational Review of Service Delivery," October2002 104 Cambodia: Integrated Fiduciary AssessmentPublic Expenditure Review security (DS) wage bill, after increasing by nearly 12 percent from 1998-1999, has declined steadily from 1999 to 2002, with the largest decrease coming in 2001-2002 (budget law). Thus by 2002 the DS wage bill i s programmed to be lower than in 1998. On the other hand the civil administration (CA) wage bill has steadily increased over the period 1998-2002 both in programmed and actual expenditures. Inrelative terms the share of the DS wage bill has declined from 66 percent in 1998 to 48 percent in 2002, while the CA share has increased from 34 to 52 over the same period, which i s indeed a very positive development on the functional resource allocation front. It i s also worth noting that interms of executionrates the CA wage billhas either matched or exceeded the DS wage bill, though overall wage bill execution rates have been deterioratingover the past few years. Table 5.5: The Wage Bill,includingCivilAdministrationandDefense/Security,as a PercentageofRecurrentExpenditures,BudgetedandActual, 1998-2002 1998 1999 2000 2001 2002 (CR millions) Civil administrationwage bill (budget law) 144,730 171,014 218,115 235,285 258,625 Defenselsecuritywage bill (budget law) 282,140 315,000 309,250 307,000 265,350 Total wage bill (budgetlaw) 426,870 486,014 527,365 542,285 523,975 Total wage bill (actualexpenditures) 447,541 518,034 517,208 509,086 544,389 Total wage bill (budget law) as pct. Ofrecurrent expenditures 48% 44% 40% 38% 33% Total wage bill (actual) as pct.of recurrent expenditures 48% 47% 43% 36% 41% Civil administrationwage billas pct. of total wage bill (actual) 34% 37% 41% 44% 52% Defensehecuritywage bill as pct.of total wagebill (actual) 66% 63% 59% 56% 48% Civil administrationwage bill: executionrate 106% 112% 98% 95% 109% Defensehecuritywage bill:executionrate 104% 104% 98% 93% 99% Overallwage bill:executionrate 105% 107% 98% 94% 104% Source: Budget laws and TOFE. 5.30 While the 40 percent constraint was met in2000 inthe budget law, it was not met in practice, as actual expenditures reached 43 percent of total recurrent spending. Nonetheless, in2001 and 2002 the programmed wage billwas set to decline to 38 and 33 percent, respectively (the actual wage bill was 36 percent of current spending in 2001). Given the wage increase effected in 2002 due to the introduction of a new classification system, however, the wage bill i s expected to be higher than that programmed in the 2002 budget (it appears that the fiscal impact of the new classification system was not fully factored into the 2002 budget law). The latest available data for 2002 (April 2003 TOFE) show the actual CA wage bill as 118 percent of the budget, and 37 percent of current expenditures. 5.31 Taking into account the existing macroeconomic framework, Table 5.6 provides an estimate of the MTWF for the period 2003-2007.91Based on most recent (December 2002) projections available (for GDP, the ratio of tax collections to GDP, and the ratio of current spending to GDP), the total civil administration wage bill i s estimated as a function of three key variables: (a) change in tax collections, (b) change in the wage bill ceiling as a percentage of current spending, and (c) the functional composition of wage bill spending (CA and DS).Inthe base case, assuming that there are no additional increases in tax collections (above those already projected), assuming Government allocates a full 40 percent of current spendingto the total wage 91 This MTWF, referred to throughout the rest of the chapter, i s based on the most recent Ih4F estimates and presents the total CA wage bill envelope for the period. For these reasons there are differences with CAR'SMTWFreferred to earlier. Reforming Institutions to Improve Service Deliverv II:The Cambodia Civil Service 105 bill, and assuming that the DS share of the total wage bill remains at 51 percent over the period, the resource envelope for the CA wage bill i s expected to grow by 65 percent from CR 349,706 in 2003 to CR 575,309 in2007 (inmillions of nominalriel). The base case thus permits a significant increase in the CA wage bill based on sustained GDP growth and substantial increases in tax collections. Table 5.6: Cambodia: MediumTermWage Bill Framework,2002-2007a/ 2002 2003 2004 2005 2006 2007 Projections Undercurrent assumptions Wage bill ceiling: 40% of current expenditure (CR million) 638,339 713,686 813,165 910,408 1,029,480 1,174,101 Wage billceiling: 40% of current expenditure (US$ million) 162,637 177,414 195,676 212,646 233,454 258,495 CurrentExpenditures (CR millions) 1,595,847 1,784,214 2,032,913 2,276,0202,573,701 2,935,252 Revised assumptions (variables in bold) Revenues as apercentageof GDP 12.1% 12.7% 13.4% 13.9% 14.3% 14.5% Additional growthinrevenuecollectionsas percentageof GDP (percentagepoints) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Revenues as apercentageof GDP (revised) b' 12.1% 12.7% 13.4% 13.9% 14.3% 14.5% Currentexpendituresas apercentageof GDP 11.1% 11.4% 11.9% 12.2% 12.6% 13.1% Currentexpendituresas apercentageof GDP (revised) 11.4% 11.9% 12.2% 12.6% 13.1% A. Revisedcurrent expenditures (CR millions) - 1,784,214 2,032,913 2,276,0202,573,701 2,935,252 Wage bill ceiling as percentageof recurrent expenditures 40.0% 40.0% 40.0% 40.0% 40.0% 40.0% B.Revisedtotal wage bill ceiling (CR millions) 713,686 813,165 910,408 1,029,480 1,174,101 Defensehecurityas percentageof wage bill 51% 51% 51% 51% 51% 51% Annualdecrease (cumulative, percentagepoints) 0% 0% 0% 0% 0% 0% C. Revisedcivil administrationwage billceiling (CR millions) 349,706 398,451 446,100 504,445 575,309 Source: IMF projections and World Bank staff estimates. a/Neither Priority Mission Group costs nor any voluntary departure costs have been included in theframework owing to the lack of detailed cost estimates, including thepossible contribution of donorfunding. b'Assumes that all additional revenues are spent on recurrent expenditures. 5.32 A reallocation of resource to the civil administration wage bill is one of the central issues for the Government's medium term expenditure framework, yet the proposed MTEF has yet to incorporate any strategic objectives on the economic composition of expenditures. The MTEFdoes propose a strategic reallocation to the priority sectors: "The strategic reallocation of resources toward the priority sectors-specially education and health-and away from defense and internal security i s one of the objectives of the proposed MTEF."92Yet the MTEF neither analyzes nor programs the economic classification of spending in terms of wage and non- wage for the period. The problem i s that with large programmed compositional shifts to the priority sectors, especially health and education (from 8.4 percent of current spending in 2000 to 12.9 in 2005, and 14 percent in 2000 to 20 percent in 2005, respectively), the total envelope for recurrent spending has been increasing, while the wage billhas not nearly increased as much. The result i s a declining share of the wage bill as a percentage of total recurrent spending. Moreover, 92 "Medium TermExpenditureFramework,2003-2005" draft document, 12September 2002,p.1. 106 Cambodia: Integrated Fiduciary AssessmentPublic ExpenditureReview the MTEF must be explicit about the source of financing for the wage reforms, including the Strategy to Rationalizethe Civil Service (SRCS). Figure 5.5: Cross-Country Comparisonsof PersonnelExpendituresas a Percentageof Recurrent SpendinginEducationand Health I 1 Years: d2001, b/1997, 1998-99,c/1997, d/1997, 1998. Source: Budget data (Cambodia), UNESCO,SIMA Database, WorldBank PERs. 5.33 The share of personnel expenditures in total recurrent spending on education and health is significantly lower in Cambodia than in other East Asian countries. Table 5.7 shows that the MOEYS wage bill as a percentage of recurrent spending i s slated to decline from 75 percent in 1999 to 57 percent in 2005 (the trend i s thus accentuated by the MTEF), while the MOH wage bill has actually fallen by more than half from 1998 to 2001 as a percentage of recurrent spending. It i s possible that the economic composition of spending in the priority sectors has become inefficient, as it i s clear that the composition i s driven from the top down rather than from the bottom up. The preliminary expenditure data for 2002 show a welcome reversal of this trend, but the data are likely to change further as non-wage expenditures are posted over the next couple of months. Further steps need to be taken to improve the efficiency of non-wage recurrent spending and correct the imbalance inthe economic composition of spending, especially in the priority sectors. The Government plans to revise the MTEF, along with CAR'S MTWF, and indoing so it should allow for substantial increases inthe civil administration wage bill. Increasing civil service wages in these sectors is the obvious solution, and one short-term option i s to use targeted functional allowances to boost priority sector wages. This remedy would need to be undertaken in the context of a systematic pay and grading restructuring to provide sustainablepay relief, however. Table 5.7: PersonnelExpendituresas a Percentageof Recurrent Spending, 1998-2005,inEducation, Health,andAgriculture 1998 1999 2000 2001 2002 2003 2004 2005 MOEYS a/ nla 75% 73% 61% 67% 54% 55% 57% MOHb' 24% 16% 13% 10% 20% d a nla d a MAFF n/a 30% 27% 13% 28% nla n/a d a "1999-2001 are actuals (TOFE),2002from TOFE (Feb. 2003),and 2003-2005from MTEF, Remunerationfrom PAP is recategorized as personnel expenditure. b' 1998-2001are actuals(TOFE)and 2002from TOFE(Feb.2003). No data availablefrom MTEF. 1999-2001are actuals (TOFE)and 2002from TOFE (Feb.2003).No data availablefrom MTEF. 5.34 The challenge for the Government is to assess to what extent additional shifts in resource allocation could open up possibilities for further civil service wage increases. A preliminary analysis of the MTWF suggests that there are four key variables, or levers, that the Government could manipulate to increase public sector remuneration: (a) the tax effort, (b) the Reforming Institutions to Improve Sewice Delivery II:The Cambodia Civil Service 107 ceiling on the total wage bill, (c) the share of the wage bill allocated to the civil administration, and (d) the overall size of the civil service. The details of the simulations are presented in Appendix 5A. 5.35 The simulations show that a combination of policies-namely, increasing the wage bill ceiling as a percentage of expenditures, reallocating from defensehsecurity to civil administration, increasing tax revenues, and retrenching staff-can have a significant increase on average remuneration.For example, the projectedaverage monthly wage in2007 i s nearly one-third higher in the alternativehmproved scenario (US$80 as compared to US$ 61) and nearly two-thirds higher inthe radicalloptimistic scenario as compared to the base case (US$ 100 as compared to US$ 61). Moreover, if these gains were focused on the priority sectors through targeted allowances the impact would be even greater. These simple projections, which are intended as illustrative and not to prefigure any specific outcomes, show that, given the right policies, civil service remuneration can be increased from the Government' s budget before resorting to internationalaid to fund pay reform. 5.36 There is no doubt that difficult decisions will have to be made by the Government over the medium term. But there is also no doubt that one of the most important binding constraints on improving the quality of poverty-reducing services i s the low level of public sector remuneration. The benefits and costs of reallocations from non-wage recurrent spending to the wage bill, of a shift in the functional composition of the wage bill from defensehecurity to civil administration, and of retrenchment will have to be carefully weighed. 5.37 Making progress on pay reform, however, would likely attract support from a broad coalition of stakeholders, including donors as well as civil servants. For example, a recent survey found that public officials support retrenchment of civil servants if such a measure would permit "an increase in salary and benefits for remaining government workers" (World Bank, 2000:33). This finding runs counter to conventional wisdom by suggesting that the gravity of the low pay problem might make mobilization of support easier if it is clear that retrenchment would be linked to a significant pay increase. Many donors would also support comprehensive pay reform as a way to improve governance and as a pre-condition for the reduction of corruption, bothof which would increasethe effectiveness of external aid. 5.38 Even with a substantial reallocation of resources, however, it is unlikely that adequate wage levels are attainable. It is clear that additional donor resources will be required, though the precise amount would depend on the details of the remuneration reform. A very preliminary analysis carried out in 2001estimated that the cost of a more competitive pay policyg3 would run approximately CR 507 billion per year (note that this estimate is based on a particular scenario and i s presented for illustrative purposes, not to suggest specific policies). The key assumptions of the scenario were that average monthly salaries would be increased to US$ 228 for senior management (estimated as about 75 percent of equivalent pay inthe NGO sector), US$ 70 in the education sector, and US$ 50 in the non-education sectors. Taking into account the resource envelope from the MTWF, the external financing requirement for 2003-2005 would range from approximately US$ 80 million inthe base case to US$ 24 million under the optimistic scenario. Table 5.8 provides details on the potential financing gap. As the table indicates, the gap will vary widely depending on the amount of domestic resources shifted to the CA wage bill. This analysis i s meant to serve as an indication of the orders of magnitude of the cost of reform, rather than as the final word. Much more work i s needed to work out the details both of the remuneration reformand the need for external financing. 93 Bannock Consulting, "Developing strategies for Priority Mission Groups towards sustainable civil service reform," November 2001. 108 Cambodia: Integrated Fiduciary Assessment Public Expenditure Review 5.39 One potentially promising option is to approach external financing for civil service pay reform through budget support. It has been suggested that by pooling donor resources currently used for project management unit (PMU) salary supplements and channeling them through the budget to support the wage bill, progress could be made toward meeting the reform- induced financing gap. The basic premise of this approach i s that rationalization of salary supplementation, which i s estimated to range from US$ 6 to 10million per annum, could form an important pillar of a major program of civil service wage reformthat could be put inplace on a "sunset" basis (ie., declining as revenues and civil service management capacity build). If, for example, current salary supplementation totals US$ 8 million per year, then the total additional financing requirement over a three year period (presented inTable 5.8) would range from US $56 million inthe base case to zero inthe optimistic case. Pooling donor salary supplements as budget support for civil service reform would also eliminate the distortionary impact of such supplements inthe Cambodian labor market and strengthen the public sector itself. Table 5.8: EstimatedExternalFinancingRe uirementsunderWage BillScenarios,2003- 2006 a / # (CR billions unless otherwise noted) ~~ ~ ~ ~ ~ ~ ~ ~~~~ 2003 2004 2005 2006 Estimated cost of reform 507 507 507 507 Resources available c/ Base case 350 398 446 504 Improved scenario 381 443 505 582 Optimistic scenario 420 497 577 676 External Financing Requirement Base case 157 109 61 3 Improved scenario 126 64 2 d a Optimistic scenario 87 10 nla d a Total financing requirement (US$ millions) Base case 80.4 Improved scenario 46.8 Optimistic scenario 23.6 Source: World Bank staff estimates. a/ Remuneration reform scenario based on Bannock Consulting (2001). b' Retrenchment scenariosnot included in these calculations. 5.40 Much work would be needed to design and implement a pooled donor approach to remuneration reform through budget support. A fully costed reform proposal would have to be developed, donors and NGOs would have to agree to rationalize their current salary supplementation practices, and the Government would have to hold extensive dialogue with donors on the budget support mechanism. Inaddition, internationalexperience with pooled donor support for salary reform would have to be in~estigated.'~The Government has agreed to carry out the analytical work necessaryfor fleshing out options for pay and employment reformusinga pooled donor support mechanism and rationalizing salary supplementation practices. This work i s a necessarypillar of any further reformand should be carriedout as soon as possible. 94 A recent example of such an approach is Bolivia, where multilateral and bilateral donors are financing a five-year salary enhancement scheme. Donor financing will support, on a decliningbasis, a portion of the incrementalrecurrent costs necessary to increase salaries for managerial and professional posts in priority ministries. See "Towards a MediumTermPayReformStrategyfor the MozambiquePublic Service," TheodoreValentine, October 2001. Reforming Institutions to Improve Service Delivery II:The Cambodia Civil Service 109 Recommendations 9 Develop a strategy to increase and decompress civil service remuneration by increasing the overall wage bill through greater tax collections, reallocating from lower- to higher- priority sectors, such as from the defensehecurity to the civil administration sector, redressing the imbalance inwage to non-wage spending, and retrenchingcivil servants in the lower priority sectors while providing compensation. 9 Revise the MTEFto make sure it i s consistent with CAR'SMTWF. 9 Use greater functional allowances to boost priority sector wages as a short term measure. 9 Carry out the analytical work necessary to develop options for pay and employment reformusinga pooled donor support mechanism and rationalizing salary supplementation practices (to be carried out inlate 2003). STRENGTHENINGCIVIL SERVICEMANAGEMENT: SAFEGUARDINGMANAGERIALAND FIDUCIARY CONTROLS 5.41 The Government has made progress on several fronts in the area of personnel management. The most important reforms to date include the civil service census and the development of the Human Resource Management Information System (HRMIS), which has allowed for simplification and automation of the payroll, as well as facilitated the review of human resource management policies. Building on this progress, the RGC i s well-placed to address short term priorities, such as strengthening establishment control, and medium term priorities, such as improving merit-based recruitment and promotionpractices. Payroll Management 5.42 It is clear that poor treasury management exacerbates the already serious problem of low wages by under-paying civil servants, and paying sporadically and unevenly during the year. The fact that the process of preparing and certifying the monthly civil service payroll l i s t has been improved by reducing the steps from fourteen to six (at both the provincial and central levels). This new process should improve and hasten the process up to the point at which it is taken over by the Treasury. 5.43 The Government has also made substantial progress by introducing an automated payroll through its HRMIS. Each month, the HRMIS system produces a hard-copy pay bill for ministries and provinces. The pay bill lists each employee within the ministry or province, their unique staff number, their job designation, their sectiodunit, and certain other information relating to family dependents, including the amount of functional allowances (the systemdoes not provide individual pay slips). The pay bill is then verified as accurate by the personnel section of the relevant ministry/province and it forms the basis for payment of salaries to staff. The relevant personnel official obtains Treasury approval for release of the appropriate funds, based on the pay bill, and the funds are then disbursed to staff, who sign to indicate that they have received the relevant amounts. There has thus been substantial progress toward the policy objective of "automating" pay, and, by extension, reducingcorrupt and/or opaque payroll practices. 5.44 The calculation of individual monthly remuneration is somewhat complex, due to the nature of the classification system. Civil service wages are broadly divided intwo types of payments: base salaries, which account for over 58 percent of total pay, and allowances, which include functional, pedagogical, family, and risk allowances. The first three types of allowance are the most significant in terms of proportion of the wage bill, representing respectively, 27 percent, 3 percent, and 3 percent in 2002. Once these pay and allowance factors have been determined it i s possible to calculate pay by adding together the basic pay and all the allowances. The concern with the complexity of the system is that it acts as a tax on administrative capacity 110 Cambodia: Integrated Fiduciary Assessment Public Expenditure Review both in the central oversight and line agencies. Even given the automated payroll, clerical staff must provide important support services to the overall system. A simpler system would consume fewer administrative resources and result in fewer potential errors and thus greater control. Though such a system might make more sense in the future when salaries are higher, at present there would be benefits to simplifying the system. A further drawback i s that individual pay slips, which would make the system more transparent, are not produced by the automated payroll system (though a limitednumber of highrankingofficials do receive pay slips). The Government should introduce individual pay slips, as resourcespermit, over the mediumterm. 5.45 Another factor that obfuscates actual remuneration is the use of overtime pay, which, at more than US$6.9 million in 2002, is not an insignificant amount. Though data on use of overtime pay are not available by ministry or civil service category, some simple calculations demonstrate that it i s a significant amount inrelative terms (there i s some indication that a significant share of overtime pay i s allocated to teachers). Assuming that overtime pay was distributedequally across civil servants during the year would mean that the average monthly salary would have been US$ 3.52 higher in 2002. At an average salary of US$ 28 (programmed for 2002) before overtime, the additional compensation would represent an increase of over 12 percent. The regulations and procedures for authorizing overtime pay should be reviewed inlight of the significant share of the annual wage bill allocated to overtime, especially inthe absence of an establishment register and discrepancies inactual numbers of civil servantsby official source. Staff recruitment and promotion 5.46 Ideally, recruitment is undertaken by way of competitive examinations, which are supervised by the SSCS. However, anecdotal evidence indicates that actual practices differ?' Permanent employment i s dependent on the results of an assessment after an obligatory probationary period of 12months. Merit-basedrecruitment needs to be strengthened significantly in practice. To the extent that salaries are increased and decompressed, it will become more important to address weaknesses inherent inrecruitment and promotion practices. 5.47 Greater attention to career development through merit-based promotions would enhance performance and provide civil servants with greater incentives to undertake professional development. Staff are appointed to a particular category according to educational qualifications. The only way a civil servant can change category i s by acquiring the necessary level of education through attendance at college or university. Progressing from one grade to another can be done inthree ways. The first consists of an assessment of the employee's abilities and performance by the ministryhe or she works for, after a formal request from the civil servant. Each year in January, civil servants are supposed to fill in a professional evaluation form, which i s in turn annotated by the department's head, though this i s not commonly done in practice. Ideally, the form i s then passed on to the SSCS, which forms an opinion of the quality of the employee and checks that this promotion corresponds to a vacancy in the relevant grade and that it is within the limits of the ministry's budget. The second way is for the civil servant to undertake long-term training at the Royal School of Administration. Finally, promotions can be achieved through an internal examination, which i s also supervised by the SSCS. All grade promotions have to be ratified by the SSCS. Progression from one class to another i s automatic with every two years spent in the civil service. In conjunction with its efforts to decompress salaries, CAR should review options for enhancing career development and strengthening the performance evaluation system for highlevel civil servants (category A). 95 Each ministry or institution wishing to recruit civil servants sends a proposal to the State Secretariat of the Civil Service (SSCS), and then to CAR for approval. MEF addresses the financial aspects and, if approved, includes the proposal in the draft of the annual budget submitted for adoption by the Council of Ministers and the National Assembly. Only when the proposalis adoptedby MEF andCAR is the agency ableto undertakerecruitment. Reforming Institutions to Improve Service Delivery II:The Cambodia Civil Service 111 5.48 Based on the economic necessity for staff to undertake private paid work, controls over absenteeism, and other disciplinary functions, appear to be ineffective. Each ministry has inspectors that are supposed to monitor staff attendance, among other things, but they do not do so inpractice. Punishment for moonlighting and absenteeismi s rare. As salaries are increased, it will be important to reestablishthe inspectionfunction. Information ManagementSystem and Establishment Control 5.49 In many respects CAR'S Human Resource Management Information System (HRMIS) is an admirable system. It combines both Khmer and English languages and now contains a wealth of information about the civil-service workforce that was previously either unobtainable or difficult to obtain. Stringent validation procedures have been applied to ensure the progressive elimination of errors, and to ensure that data are comprehensive and accurate. There are, however, anumber of issuesthat require attention. 5.50 While there are procedures for controlling the entry of names into the HRMIS database, the system itself cannot determine whether a particular appointment, promotion, or transfer is valid or even within budgetary limits. Though CAR has attempted to address some of these system weaknesses manually, no simple, comprehensive system exists at present for identifying individual posts inthe civil service. This, inturn, means that there i s no means for defining a link between a civil servant and a particular post. There is, therefore, no possibility of producing an establishment register, and thus there i s not an adequate instrument for exercising control over the established w ~ r k f o r c e . ~ ~ 5.51 It appears, therefore, that establishment control is the weakest aspect of the system.97The key problem is the absence of any separate table within the database listing every post and assigning it a unique number. This, in tum, means that there i s no uniquely definitive linkage between posts and employees and no possibility of producing a comprehensively accurate establishment register. 5.52 While the total number of posts in a particular establishment could probably be inferred, such a reconciliation would almost certainly have to be performed manually, as the system could not do it, and there would be no guarantee of its accuracy.98In such circumstances, the whole point of an establishment control system would be defeated. Moreover, over time, it would become progressively harder to ascertain what a baseline establishment figure 96 Ministries, departments,units and post designations have all been appropriatelycoded and these details are heldin separate tables within the database.There appear to be approximately 100post designationsin additionto the requisite number of ministerial, provincial unit and department codes (in database programmingterms, devising and entering these codes would not havebeena significantly difficult task and it is reasonably safe therefore to assume that this has been achieved to a high level of accuracy). All these codes are unique. They do not and cannot, however, uniquely define each post, because the posts themselves are not includedin the system, and there appears to be no method of doing so, Le., no consistently reliable and individual post identification and numbering system in the ministries/provinces. 97 There does not appear to be any provision for holding the following, typically necessary, information about each post: date of financial approvalfor the post's creation, type and amount of allowances attachedto the post, and status (Le. permanent or temporary, full- or part-time). This also means that additional information on salary scale and appropriate grade/classification, which would be useful to collect, i s not included(as i s commonin traditional French systems). 98 This could be done by noting the total of sequentially-numbered employeeson each monthly payroll, (together with any vacanciesindicatedas such), and cross-checkingthe figure with some other list or total obtainedelsewhere giving the numberof approvedposts inthat establishment (this latter figure would presumablyremainfairly static from month to month, though it probably would not have been subject to the same validation tests as the staff data). By such a means, it might also bepossibleto ascertain whether the particularministry had exceededits authorizedestablishment, Le., employedmorepeoplethan it was entitledto do, or possiblyfallen short of its requirements. 112 Cambodia: Integrated Fiduciary Assessment Public Expenditure Review should be, or should have been at some pre-determined point, Le., the system's inception or the beginning of a financial year. 5.53 The importance of the lack of an establishment register cannot be underestimated. Without the ability to produce an establishment register, the system cannot properly control posts and staff within ministries and provinces, because, as outlined above, it cannot assign a unique relationship to each. The current process of validating appointments, promotions and transfers depends heavily upon the instituting ministry/province. There i s no way of independently validating or verifying the information provided. This clearly indicates that, although the HRMIS i s at present a completely centralized function, it i s limited to a purely reactive role. It only records, rather than independently authorizes and verifies, the data provided to it. Moreover, the highpercentage of resources spent on contractual workers and overtime further argues the case for tighter establishment control.99The possibility of artificialhnapproved or purposely-inflated establishment totals being used by particular establishments would be considerable and very difficult to correct. It i s therefore necessary that the Government create an establishment register to improve civil service management and budgetary control. The Government i s well-placed to strengthen establishment control based on its previous reform efforts and building on its promising HRMIS. Recommendations 9 Develop options for improving management and control of civil service employment, including formulation of an establishment register proposal, and piloting of the proposal inselect ministries. 9 Integrate the HRMIS database into the budget preparation process by providing access to > MEFand (read only) access to line ministries. Improve establishment control across the civil service based on fine tuning of the pilot experience. 9 Develop effective monitoring mechanisms to assess civil servants' performance and introducedisciplinary measures to promote better performance. CONCLUSION 5.54 This chapter argues that given the serious constraints on civil service capacity, a comprehensive pay and employment reform program is needed. Such a program would significantly improve civil service pay, particularly for essential skills, rationalize civil service employment and deployment, and strengthen civil service management. From the donor community to NGOs to public officials themselves there i s clear agreement that low capacity in the civil service i s perhaps the leading constraints on Cambodia's development. The urgency of the situation and potential impact of far-reaching reform means that the Government will be challenged to pushitself considerably farther and faster than currently envisioned. 99 The procedures for authorizing and calculating overtime compensation should be reviewed, especially as the new automatedpayroll system (andhardcopypay bill) does not seemto include a column for overtime compensation. Reforming Institutions to Improve Service Delivery II:The Cambodia Civil Service 113 APPENDIX 5.A 1 The purpose of this appendix is to present different scenarios as a function of changesinthree key variables. Based on the variables definedinTable 5.6 (A, B, and C, above, highlighted), a sensitivity analysis was conducted. Three scenarios were analyzed: a. Pessimistic - 0.3 percentage point reduction in revenues/GDP (0.3 percentage points lower each year), no revisions in the wage bill ceiling, and no shift in composition from DS to CA; b. Zmproved - 0.2 percentage point increase in revenues/GDP (0.2 percentage points higher each year), wage bill ceiling set at 42 percent, and 1 percent annual change in composition from DS to CA; C. Optimistic - 0.3 percentage point change inrevenues/GDP (0.3 percentage points higher each year), wage bill ceiling set at 45 percent, and 2 percent annual change in composition from DS to CA. 2 The scenario analysis shows that, based on existing macroeconomic and fiscal assumptions,significant gainscan be made within the Government's mediumterm resource envelope, assuming reallocations from non-wage operating and maintenance costs (including efficiency gains) to the wage bill and assuming reallocations from the DS to the CA wage bill (see Table 5A.1).Vis-his the base case, the total civil administrationwage bill in 2007 can be 17 percent higher under Scenario 111, or as much as 39 percent higher under the more optimistic Scenario IV. In US dollar terms, under Scenario I11it i s feasible to expect average monthly wages in the civil servicelooto increase to approximately US$ 72, while average wages could reach US$ 85 under the optimistic scenario. While the use of averages necessarily entails a gross calculation, the analysis indicates that there i s room for maneuver under existing macroeconomic and fiscal projections. Table 5A.1: Civil Service Average Monthly Wage Projections, 2003-2007 a/ 2003 2004 2005 2006 2007 I.Basecase 35 41 47 54 61 11. Pessimistic Scenario 34 40 46 52 60 111. Improved Scenario 38 46 53 62 72 IV. Optimistic Scenario 42 51 60 72 85 Source: WB staff estimates. a`Based on CAR'sprojected worqorce size (2007 set equal to 2006). 3 Given tight fiscal constraints and the low level of wages, an analysis of the impact of a reductionin size of the civil service should also be undertaken. A very preliminary scenario analysis i s presented here for the purpose of stimulating discussion rather than proposing specific policies or targets. Retrenchment would have to be accompanied by adequate departure packages, as envisioned by CAR, and possibly other measures (e.g., retraining). Three scenarios are considered"': looCivil service wage bill projections subtract salary costs for the royal, legislative, and judicial branches, and also subtracts an estimate for overtimepayments, and contractualand "floating" pay costs. Basedon CAR estimates of the historical difference between the total civil administrationwage bill and the civil service wage bill, the ratio of civil service to civil administration salary costs averages to 84 percent based on projectionsfor the period 2003-2007. In 2002 CAR estimatesthat the cost of overtime and contractualand floating staff was nearly 14percent of the total civil administrationwagebill (overtimeaccountedfor 10.7percent). lo'Naturalattrition is basedon CAR's estimates(about 1.2percent on averageper annum). 114 Cambodia: Integrated Fiduciary Assessment Public Expenditure Review a. SRCS baselinescenario-in which a total of 15,750 civil servants are retrenched over the period 2003-2006, with most of the reductionoccurring in2005-2006; b. Alternative scenario-in which 31,500 civil servants (double the baseline scenario) are retrenched over the same period, with most of the reductionoccurring in2005-2006; and C. Radical scenario-in which the size of the civil service i s set to be equal to one percent period).IO! of the PO ulation by 2006 (assuming modest population growth of two percent over the Table 5A.2: Projected Civil Service Size (Year End) by Scenario, 2003-2006 2003 2004 2005 2006 GTR NTR I.SRCSScenario(Basecase) 166,381 165,953 164,595 163,469 15,750 1,900 11. Alternative Scenario 163,731 159,263 153,445 147,719 31,500 17,650 111. Radical Scenario 158,803 152,188 145,062 138,308 40,911 27,061 Source: CAR MTWF and World Bank staff estimates. Note: Gross Total Reduction (GTR) is the sum of the number of retrenched employees only, while Net Total Reduction is the difference between the size of the civil service at the start of 2003 as compared with the end of 2006.All calculations in the table include attrition and new hiring asprogrammed by CAR. 5 The retrenchment analysis shows that in the first couple of years of the retrenchment program the gains are quite small and there is not much difference between the scenarios, taking into account the differential costs of severance packages (see Table 5A.3).lo3 By 2007, however, the comprehensive and radical options yield more benefits as compared to the base case. The average monthly wage in the radical scenario, for example, i s 18 percent higher. Yet, inabsolute terms, salaries are still low. Table 5A.3: Projected Monthly Civil Service Wage (US$), 2003-2007 2003 2004 2005 2006 2007 I.SRCSScenario(basecase) 34 40 45 52 61 11. Alternative Scenario 34 40 47 55 68 111. Radical Scenario 34 41 49 58 72 Source: CAR MTWF and World Bank staff estimates. Note: Size of workforce in 2007 set equal to 2006. 6 A final simulation (see Table 5A.4), in which the retrenchment scenarios were combined with the wage bill scenarios, shows that a combination of policies-namely, increasing the wage bill ceiling as a percentage of expenditures, reallocating from defensehecurity to civil administration, increasing tax revenues, and retrenching staff-can have a significant increase on average rem~neration.'~~ For example, the projected average monthly wage in 2007 i s nearly one-third higher in the alternativehmproved scenario and nearly lo'Settingthe size of the civil serviceequal to one percent of the populationwould placeCambodia on the low endof the low income countries spectrumof share ofpublic sector employment to population.See Figure5.3. lo3Table 5A.2 presents the projectedsize of the civil service according to the three scenarios. Note that ineach of the scenarios it is assumedthat all reductionswill come fromthe non-educationsector. At the same time the projecteddata take into account normal attrition and programmed new hiring over the period, as estimated by CAR'SMTWF. The next step was to estimate the cost of the retrenchment program, which was based on CAR data. It was assumed that each retrenchedcivil servant would receive a departure package, the value of which was equal to the average cost of eighteen months of wages (the annual average wage as estimatedby CAR over the period was usedto calculate the severance package values). The next step was to calculatethe average monthly civil service wage after retrenchment (Table 5A.3). lo4 The altemative retrenchment scenario i s combined with the improved wage bill scenario and the radical retrenchment scenario is combinedwith the optimistic wage bill scenario (the base line scenario remains the same in bothcases).The only difference betweenTables 5.7 and5.9 is that Table 5.7 does not factor inretrenchmentcosts. ReforminP Institutions to ImDrove Service Deliverv II:The Cambodia Civil Service 115 two-thirds higher in the radicaVoptimistic scenario as compared to the base case. These simple projections, which are intended as illustrative and not to prefigure any specific outcomes, show that, given the right policies, civil service remuneration can be increased from the Government's budget beforeresortingto intemational aid to fundpay reform. Table 5A.4: CombinedWage BillandRetrenchmentScenarios,AverageMonthlyWage (US$) 2003-2007 2003 2004 2005 2006 2007 Base case AlternativeRetrenchmentScenario + 34 40 45 52 61 RadicalRetrenchment Scenario+ ImprovedWage Bill Scenario 39 47 57 68 80 OptimisticWage Bill Scenario 44 56 69 85 100 Source: SRCS, CAR MTWF, and World Bank staff estimates. Note: Size of workforce in 2007 set equal to 2006. 6. REFORMINGINSTITUTIONSTO IMPROVESERVICE DELIVERY111:DECENTRALIZATION "The devolution of power carried out through decentralization and deconcentration will shift thefocus of development towards the people with the view to implementing structural adjustment, strengthening grassroots governance, reducing poverty and inequality, and promoting gender equity at all levels."`05 REFORMAGENDAAND INSTITUTIONALFRAMEWORK 6.1 The RGC is undertaking a decentralization program to devolve power to newly created semi-autonomous elected governments at the commune and sangkat (urban commune) level and move toward enhanced roles for provinces and municipalities as deconcentrated entities of the center.lo6It i s important to emphasize, however, that commune level reforms are inthe very first stages of implementation, while provincial level reforms are still under development, though provinces and communes have existed as administrative levels of the central government since the French colonial era (the numbers and distribution of the various administrative structures are summarized in Table 6.l)."' Decentralization, however, i s considered a critical element of the public sector reformagenda (Governance Action Plan, 2001). 6.2 Improving service delivery in the face of severe administrative capacity constraints will likely mean pursuing both devolution to the communes and further deconcentration to the provincial administrations. Devolution in principle i s "the transfer of authority for decision making, finance and management to quasi-autonomous units of local government (World Bank, 2000)." As yet, however, Cambodia's decentralized communes have no mandated responsibilities and only very limited financing. While the communes may eventually become important service providers, this i s not likely to happen for many years. Strengthening devolved provincial administration i s a more promising near term measure, though deconcentration implies, by definition, less decentralization than devolution. The aims of this chapter are thus modest. First, it addresses commune level decentralization by offering some guidelines on further developing the devolved system. Second, it offers some suggestions for improving expenditure management and service delivery at the deconcentrated provincial level. 6.3 As the Government proceeds to deepen and extend the commune decentralization process, it will need to further develop the legal and institutional structure as it pertains to: (a) expenditure and revenue assignment and management; (b) monitoring and evaluation systems; and (c) the inter-governmental transfer system. On expenditure and revenue assignment, the priority i s to undertake the necessary analytical work to inform the upcoming commune boundary review. On expenditure and financial management, the advantage i s that the system i s generally well-developed, though there are some procedural complexities that might "Cambodia: NationalPovertyReductionStrategy," December2002, p. 108. lo` "Provinces" is used throughout the rest of the chapter to denote both provinces and municipalities, while "communes" is usedto signify bothcommunes andsangkats, unless otherwise noted. lo'The 1993 Constitution establishes provinces (khet) and municipalities (krung) as the principal temtorial sub- divisions of the Cambodian state. Provinces are further divided into districts (srok) and communes (khum), while municipalities(which have the same status as provincesin selectedurbanareas) are further divided into sectors (khan) andurbancommunes (sangkat).Villages exist inruralareas andurbanareas. 117 118 Cambodia: Intearated Fiduciarv Assessment and Public Exvenditure Review prove difficult for the less experienced communes. These will need to be addressed. The challenge i s to support implementationof the system by closely monitoring progress and to adjust andor simplify the systems and procedures, if necessary. This entails the establishment of an adequate commune monitoring and evaluation system as a key strategic measure to safeguardthe accountability across the decentralized system. The most immediate priority for further developing the inter-governmental transfer system i s setting up the Commune/Sangkat Fund Board to manage fund resources transparently. Over the medium term, and tightly linked to devolution of specific functions to the communes, i s the determination of the fixed resource share allocated to the communes and sangkats, which will be necessary for providing the communes with some measureof resource stability. Table 6.1: Number of Districts/Sectors,Communes/SangkatsandVillagesby ProvinceMunicipality inCambodia No. ProvinciaVMunicipalities Districts/Sectors CommunedSangkats Villages 1 Bantey Meanche 8 64 623 2 Battambang 13 96 733 3 KampongCham 16 173 1,758 4 Kampong Chhnang 8 69 553 5 KampongSpeu 8 87 1,308 6 KampongThom 8 81 732 7 Kampot 8 92 478 8 Kandal 11 147 1,087 9 KokKong 8 33 130 10 Kraches 5 46 250 11 Mondulkiri 5 21 90 12a/ PhnomPenh 7 76 637 13 PreahVihear 7 49 208 14 Prey Veng 12 116 1,137 15 Pursat 6 49 501 16 Ratanakiri 9 49 240 17 SiemReap 12 100 875 18 SihaknoukVille 3 22 94 19 StengTreng 5 34 128 20 Svay Rieng 7 80 690 21 Takeo 10 100 1,116 22 Otdor Meanchey 5 24 231 23 a/ Key 2 5 16 24 Pailin 2 8 79 Total 185 1.621 13.694 a/Denotesmunicipaliry, sectors, sangkats,and villages. Source: Prakas No 493 PRK, on Number, Name and Boundaries of the Communes and Sangkats of the Kingdom of Cambodia,30 April 2001. 6.4 As the RGC moves forward with its development of a plan to deconcentrate responsibilities to the provinces, work on a number of preparatory measures would smooth the transition. First, there i s a need for an overarching deconcentration framework and analysis of options for Cambodia, with a focus on the objectives to be achieved through deconcentration. CAR, which i s charged with developing the framework, should provide a clear set of guidelines to the sectoral ministries for developing approaches, and needs to coordinate this work closely with them. This framework should also address the relationshipbetween provincial governors and line ministries provincial administrations. Second, there i s a need to improve coordination between ministries and provinces on expenditure management, given the finding that there is no positive correlation between poverty incidence, or the number of poor, and provincial budget allocations for health, education or other sectors, which results from high expenditures on Reforming Institutions to Improve Service Delivery III:Decentralization and Deconcentration 119 administrative overhead and a lack of alignment between provincial spending and sectoral priorities. Inthe short term, sectoral earmarking at the provincial level would allow line agencies to improve pro-poor targeting of spending. 6.5 Weak institutional capacity at all levels remains a great challenge to designing and implementing decentralization and deconcentration reforms. Strengthening capacity, both in terms of training commune and provincial level civil servants and clarifying the institutional support structures for decentralization and deconcentration reforms, i s critical for success. This also would entail clarifying the role of villages in the commune structure. The most important ingredient for success, however, i s likely to be meaningful accountability relationships between local elected officials and citizens. Bolstering and safeguarding the democratic integrity of the commune systemi s an important challenge for the Government. DECENTRALIZATIONTHECOMMUNESANDSANGKATS TO 6.6 The first commune elections were held in February 2002. The commune councils are elected with a five-year mandate. Councils consist of five to eleven members depending on the population and location of the commune. The councils are elected on a proportional basis, such that more than one politicalparty can be represented. The presidentof the council (commune chief) i s the individual who receivesthe most votes on the majority-party candidatelist. 6.7 The legal foundation of decentralized government is straightforward, and is composedof two laws passedin2001on commune administrativeand electoral systems. The 2001Law on Commune/Sangkat Administrative Management does not assign mandatory sectoral responsibilities to the communes and sangkats. At this early stage, it simply provides a framework in which the communes are empowered to: (a) maintain public order and security; (b) organizeandmanagepublic services; (c) enhance the welfare of citizens; (d)promote socioeconomic development and upgrade living standards; (e) protect and preserve the environment, natural resources, and culture; (f) coordinate citizens views in order to promote tolerance and mutual understanding;and(g) performgeneralaffairs to respondto the needs of citizens. 6.8 Mandatory functions have not been assignedbecausemany communes have little or no service delivery capacity. In many countries, decentralization (in the sense of devolution) refers to the transfer of existing functions and resources from a higher to a lower level of government. Accordingly, the strategy of initially empowering communes to cover only limited administrative functions and to provide a few services that are local in nature (and have not previously been provided to any significant extent by the central government) i s a pragmatic way to develop the commune system, as long as it i s properly embedded ina larger program of public sector reform. 6.9 However, the lack of assigned functions means that communes have no mandated expenditure responsibilities. Assigning functional responsibilities is typically regarded as the first step in the process of devolution. The design of the inter-governmental transfer system and the granting of own-source revenue rights-allowing for sufficient funding to meet expenditure obligations-would then follow. Cambodia's strategy i s to allow communes to identify initial priority services through participatory mechanisms and to develop them locally, which should stimulate some basic local capacity and build connections to residents. While this strategy i s reasonable as a first phase, given capacity constraints at the commune level, the Government should develop a plan and methodology for phasing inmandated responsibilities over the medium term for two reasons. First,the question of the adequacy of inter-governmental transfers and own- source revenuescannot be addressedin a vacuum. Second, there i s a riskthat lack of coordination with national ministries might result induplication of functions, thereby decreasing the efficiency of resource use (and possibly resulting in more wasteful spending on general administration). 120 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review Though transfers to the communes presently account for a small share of total government spending, resource scarcity and poverty reduction demand that they are spent efficiently. Allocations to the communes have an opportunity cost, and while it may be that the marginal benefit of commune expenditure i s higher than the opportunity cost of line ministry expenditure, such an outcome shouldnot simply be assumed. 6.10 The RGC should develop a strategy for devolving service delivery responsibilitiesto communes-and adapting the role of line agencies-as the system matures. If it i s found, for example, that the majority of communes are spending on water supply and irrigation services, the Government should consider devolving these functions to the communes and then revamp line agencies currently responsible for these functions so that their role changes from direct service provision to regulation and oversight. The general principles of expenditure assignment are well- known from the decentralization literature (see, e.g., McLure and Martinez-VBzquez, 2002). The principle of subsidiarity, which states that services should be delivered by the level of government to which the benefits accrue, and thus depends foremost on the size of local government entities, i s of particular relevance for Cambodia at the moment in light of the upcoming commune boundary review (mandated to occur before the July 2003 elections). 6.11 The commune boundary review will have a major impact on the type of services that communes can provide and on the types of own-source revenues that communes can collect. The commune boundary review should thus be guided by explicit analysis of assumptions about expenditure and revenue assignments. For this reason the review will need to be supported by analysis that addresses both of these assignment issues. Moreover, as i s the case in many countries, it i s clear that one system will not fit boththe urban and rural sectors. The Government could take advantage of the boundary review to clarify differences in expenditure and revenue assignment between communes and sangkats, and to further elucidate the relationship between sangkats and municipalities. Inter-governmental transfers 6.12 Communes are entitled to receive revenues from the national budget as well as to collect their own revenues (they are not allowed to borrow). Communes are required to use national guidelines to establish a system for the management, monitoring and control of their finances. The Commune/Sangkat Fund (CSF), whose sub-decree was approved by the National Committee to Support the Communes (NCSC) inFebruary 2002, i s capitalized from domestic and external sources. The contribution from the RGC budget in 2002 was US$ 5 million (1.2 percent of recurrent domestic revenue). According to the MTEF, the contribution i s expected to grow to 2.5 percent, or US$ 12 million in FY 2004 (see Table 6.2). .Thoughthe law requires that a share of the RGC budget be allocated to the CSF, the obligation i s not statutorily defined. Fixing an allocation in due course-once there i s greater clarity about expenditure assignment-would be necessary for providing the communes with some measure of resource stability. 6.13 CSF resourcesare heldinthe NT and transferred to provincial treasuries, where each commune has an account. The CSF is to be managed by a CSF Board, which i s responsible for setting CSF policies. International experience suggests that it i s good practice to have a separate and more independent management structure for this type of fund rather than housing it under an existing ministry or committee. MEFchairs the Board, and other RGC members are drawn from MOI, MOP and the CDC. Inaddition, three representatives of commune councils will be selected by MOI. The Board i s not yet functioning because no CS representatives have been chosen. A Technical Secretariat to support the Board, however, has already been named by the MEF.While Reforming Institutions to lmprove Service Delivery 111: Decentralization and Deconcentration 121 the CSF Board i s being set up, control over CSF resources i s vested in the NCSC. Establishing the Board as planned i s a critical next step inthe decentralization process.'08 Table 6.2: IntergovernmentalTransfersto Communes:BaseCaseResourceEnvelope (CR billions) 2002 2003 2004 2005 ~ ~ Budget MTEF MTEF MTEF Total domestic recurrent revenue 1,791 2,035 2,286 2,547 Commune Fundas % of recurrent 1.1 2.0 2.5 2.5 revenue Commune Fund(CR millions) 20 40 57 63 Distributed as: Recurrent expenditures 6,660 13,551 19,030 21,207 Capital expenditures 13,340 27,142 38,117 42,478 Source: MTEF. 6.14 Governors are required to prepare quarterly reports on CSF operations for MOI. Based on this information, MOI/DOLA i s required to prepare quarterly reports for the CSF Board. The CSF Board Secretariat i s required to prepare reports for the RGC and all donors and international financial institutions that support the CSFFund. 6.15 CSF transfers are divided into general administration (GA) and local development (LD) comp~nents.'~~ CSF Board sets component weights, but the GA cannot exceed 30 The percent of the total and the LD cannot be less than 70 percent. The GA share i s allocated to communes in proportion to the number of councilors. The LD share i s allocated according to a formula with three components: equal share, a share proportional to population, and a share proportional to relative poverty. The CSF Board determines component weights subject to the constraint that at least 40 percent must be based on the poverty component. The FY2002 allocation formula, however, used weights for the equal share, population, and poverty components, respectively, of 50, 30 and 20 percent. 6.16 Commune councils were divided into two CSF allocation groups in 2002. The more advanced ones (category 1-506 of 1,621 in 2002) that participated in Seila received full allocations (GA and LD), while new communes (category 2) only received GA funds. In2002, a total of US$ 6.42 million dollars was allocated to the communes. Of this amount, US$ 2.21 million was allocated for the GA component and $US4.21million for the LD component. The highest LD allocation to a single CS is US$15,530 and the lowest is US$5,400 in2002. 6.17 It would be preferable to adopt a more graduated approach to phasing in access to development resources based on capacity and performance. A recent proposal has suggested doing away with category 2 status and creating two category 1 groups, 1-A and 1-B, with the former receiving the full and the latter a much reduced LD allocation. This does deal with the problem of some councils having no resources to deliver services, but it uses inappropriate criteria for ensuring that higher status communes can spend resources wisely. If weaker councils initially received small development grants, those performing adequately could be successively lo*Verify status. '09CSF resources are held in a deposit account in the NT. RGC budget contributions to this account are to be transferred in three installments:50 percent on or beforeMarch 1, 30 percent on or beforeJune 1,and20 percent on or before September 1. Extemal contributions to the CSF account are supposed to be transferred from an MEF foreign deposit account in the NBC maintainedto receive grants and loans from donor agencies and international financial institutions,however in practicethis has not yet happened (e.g.,the CSF deposit account in the NBC has not yet been established).The use of commercial banks will bepilotedinsome provincesin2004. 122 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review awarded larger amounts in subsequent years until they reached the full allocation. The transfer system could also be generally used to provide hard incentives for all communes (by using ex post sanctions for poorly performing communes, e.g.). The plan for 2003 i s for all communes to receive the LDallocation. 6.18 There i s some concern that councilor allowances approved in May 2002 seem high relative to the part-time effort required of elected councilors. In the aggregate, these allowances have also come up against CSF principles, which state that a maximum of 30 percent of CSF allocations should be devoted to administrative expenses."' The broader issue, however, i s that the overall level of commune administrative expenses i s high, and that in some communes administrationfunds are the only resources being provided. Own-source revenues 6.19 Commune councils have the right to raise own-source revenues (OSRs), namely tax and non-tax revenues and service charges. The decentralization law assigns land, real estate, and rental taxes to the communes, but it does not provide details. The law requires further elaboration of the commune revenue sources, including types, rates and collection processes. It also requires RGC compensation for agency functions performed by communes on behalf of a state agency. 6.20 Previous experience with raising revenue at the commune/village level is limited to three types. The first are fees unrelated to government that are paid to community structures (e.g., pagoda committees) and local NGOs to build or maintain facilities. The second includes contributions required by international organizations for development projects, such as L O rural roads or UNICEF water and sanitation projects. The third i s the system of capital and recurrent contributionsused under Seila. 6.21 Developing commune OSRs will be essential, over the medium term, for ensuring financial sustainability, and is regarded by the RGC as a key element of its "second phase" of decentralization. MEFand NCSC are planning to buildon Seila experience inthe early stages of fiscal decentralization to communes, but there has been no concrete action to date. Initial sources will likely include a betterment levy-type assessment structured as a contribution for local infrastructure improvements paid by the beneficiaries."' A second near-term source i s likely to be user charges for selected local services.l12One short term option here would be the transfer of the animal slaughter (abattoirs) tax from the provinces to the communes, to give the communes a small source of revenue and allow them to develop collection ~apacity.''~ Inaddition, communes are to be allowed to impose fees currently charged by the centrallprovincial government for administrative services, such as birth, marriage and death registration, and to share in land 'lo The allocationto administrativeexpenses in 2002 was closer to 45 percent of total transfers from the center, but the portion over and above the 30 percent allocationis mostly funded from the provincial budget line that usedto finance chiefs when they were MOIemployees. '" Betterment levies will be most beneficial for urban areas, but they could also be used to finance certain types of rural infrastructure as well. The Municipality of Phnom Penh has already tested with some success this funding arrangement for the rehabilitation of streets and pavements. The CSF is supposed to require some minimum local commitment of funds for infrastructureinvestments, which may vary by type of public service. The CS will then have flexibility to charge more than the minimumandwill decide, givencentral guidelines, how to apportion the assessment amongCS residents. '"The Municipality of PhnomPenh, for example, charges for water and requires households to pay for the cost of sewage service through a 10percent charge assessed on the water consumption bill. Some of the Seila programrural communes havedevelopeduser fees for selectedservices. '13 The animal slaughter tax constituted 1.2percent of provincial OSRs in 2002 (budget estimate). Reforming Institutions to Improve Service Delivery III:Decentralization and Deconcentration 123 transfer fees."4 National guidelines for setting user charges need to be developed, but CS will require some flexibility in implementing them. Further analytical work on alternative commune revenue bases should be undertaken beforefurther developing the OSR legal framework. Communeplanning,publicexpenditure,andfinancialmanagement 6.22 The decentralization legislation requires all commune councils to prepare a simple 5- year Commune Development Plan (CDP) and a 3-year rolling Commune Investment Program (CIP). The commune Planning and Budgeting Committee (PBC), the only committee required to date, assists the chief with preparation. Participation i s facilitated through a broad- based Planning Forum, which comprises all council members, two representatives (one male and one female) from each village, and one representative from each NGO registered with the c~uncil."~The commune council i s required to monitor and prepare annual reports on CDP implementation for its constituents, the RGC, and donors who fund the CSF (the report must be prepared by the CS chief and the PBC and approved by the council. A comprehensive report on the results and impact of the CDP i s required six months before the expiration of the council's mandate. 6.23 Commune councils are required to actively promote and coordinate mechanisms for consultation with residents, civil society organizations, and community groups. Councils are expected to provide training, printed materials (such as handbooks), and education to the public. Commune residents can attend council meetings, and the chief i s responsible for encouraging them to do so. They cannot vote, but they can ask questions and make written complaints or suggestions to which the council i s obligated to respond. Commune councils are required to set up information boards at headquarters and every constituent village to display official council notices, including plans and budgets, as well as other relevantnews and information. 6.24 The commune budget, which is the responsibility of the commune chief, assisted by the PBC, is to be based on the CDP and CIP.l16 The budget i s intended to be formulated through an extension of the participatory planning process, but it i s not yet clear how genuinely participatory the process i s in practice. Budget amendment i s allowed, but specific procedures must be followed unless the amendment simply involves reallocation of resources within an approved expenditure category (and under its ceiling) or altering investment projects. The governor i s responsible for ensuring compliance with budgeting rules, but has no right to demand alteration of the budget if legality requirements are met.117,118 6.25 A separate commune accounting and payment system (cash-based with double entry recording) has been established to record transactions, produce required reports, and to provide a basis for monitoringbudget implementation. The chief i s the budget officer and has responsibility to authorize revenue collection, commit expenditures, and order payments in accordance with the approved budget. Only non-tax revenues can be collected directly by the communes (by a council-appointed revenue collection officer), and only minor payments are made directly through a petty cash system. The PTs, which serve as commune accountant, handle most commune financial activities, including: (a) collection of revenues, (b) execution of This has already been approved by NCSC, as has a joint MEF-Ministry o f Land Management, UrbanPlanning and Construction (MOLMUPC) prakas on sharing landtransfer fees with the CS. '15Over time coordination between planning at the commune andprovincial levels will have to be addressed. 'I6The CIP i s to cover the projects to be financed in full or in part with the CSF allocation and to include documentation of co-financing from own sources and any other sources. The annualcomponent of the CIP i s supposed to be reflected inthe capital expenditure section o f the commune annualbudget. '17The March 2002 Sub-decree on the Commune/Sangkat Financial Management System and subsequent budget- related guidelines issued by MEFlay out the details of the recurrent and capital revenue and expenditure budgets. 'I8Ifthere is an irresolvable conflict between the governor and the commune chief, the Minister of Interior mediates andmakes the finaldetermination on the case. 124 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review payment orders issued by the commune chief, (c) release of funds, (d) recording of transactions, (e) preparation of periodic financial reports, and (f) management (viz., filing and safeguarding all commune financial documents and records. PTs are also charged with monitoring commune finances. The councils are subject to audit by the NAA (and nongovernment auditors are required by CSF donors). The commune chief, accountant, and others involved in budget implementation are legally liable for undertakings that result inexcess expenditures, unauthorized adjustment of appropriations, and any other act detrimental to commune finances or property. Disciplinary procedures allow the council to deal with councilors who do not meet obligations, and sanctioned councilors may be removedand replaced according to MOIprocedures. 6.26 The procurement of works, goods and services must be undertaken in a manner consistentwith the general principles followed by the RGC. Small purchases can be evaluated by the chief or a delegate. For larger transactions, a procurement committee appointed by the chief manages the process. The committee must include the chief or a delegate (chair), two council members, and a clerk (secretary). Prior to procurement for a development project, the commune needs technical clearance from the relevant provincial department through the Provincial Office of Local Administration (POLA), although clearance i s assumed if the department does not respond within fifteen days of submission. Details of the bidding and award process have been clearly specified. Under unusual and limited circumstances, the provincial governor may authorize a waiver of competitive procurement. At the commitment stage, a contract or agreement consistent with the budget is negotiated. The verification process requires the chief to ensure that goods have been deliveredservices rendered and the supplier invoice i s correct. At this point, the chief issues a payment certificate, which i s processedby the PT. 6.27 The communes are required to prepare various financial reports. These include: (a) monthly and mid-year budget implementation reports in the form of a revenue and expenditure statement consistent with the budget classification system; and (b) a set of annual financial statements, including an annual revenue and expenditure statement, reconciliation of the commune's PT deposit account, and a fixed assets statement. The commune accountant at the provincial treasury must prepare these closing financial statements according to specified procedures, and the council must approve them. 6.28 The system of commune planning and budgeting is generally well developed at the conceptual level, largely because it benefited from six years of Seila experimentation, though a few aspects of the system require further consideration. Concerns have been raised about the possible complexity of the commune planning process, which includes five phases and eleven steps that require substantial time and input from broad segments of the local community. To the extent that this process could be simplified-without compromising community participation-it would be less taxing of local capacity. Second, the process for channeling funds to communes in three installments that require certain conditions to be met has been criticized as overly burdensome. A two-installment process mightmake more sense. 6.29 The picture is less clear regarding the implementation of financial management practices at the commune level. The major concern is that commune financial management procedures are different from the regular (national and provincial) procedures, which further fragments the system. The PTs act as cashiers but authorization of expenditure rests solely with commune chiefs. There is no pre-audit requirement and it i s not clear what other audit arrangements will be in place, raising concerns about fiduciary accountability. A related issue i s the flow of funds to the PTs during the year, which are likely to be either erratic or back-loaded, given systemic weaknesses. Concerns have also been raised about the complexity and demanding reporting requirements of the commune level PEFMprocess. Thus, it is important for the NCSC, presumably with MEFinthe leadrole, to monitor the implementationof planning, budgeting, and financial management processes, and to be willing to adjust and/or simplify the systems and Reforming Institutions to Improve Service Delivery Ill:Decentralization and Deconcentration 125 procedures, if necessary. Close scrutiny may require special monitoring of a sample of communes in order to gain a greater understanding of the system's actual operation. These concems are likely to be especially relevant in the provinces and communes that did not participate in Seila Phase I. Institutional development and monitoring 6.30 The Law on Commune/Sangkat Administrative Reform gives the Ministry of Interior (MOI) responsibility for determining procedures for monitoring and control of the general activities of the communes/sangkats. Three major types of monitoring and control are expected. The first i s to ensure that communes are meetingall legal requirements inperforming their duties. The secondi s to assess the degree of development of commune capacity. The third i s to assess the nature andextent of interventions and assistanceneededto develop commune capacity. 6.31 The MOI has set up the Department of Local Administration (DOLA) to perform these functions. DOLA currently has one or two staff ineach province and will eventually set up a Provincial Office of Local Administration (POLA) ineach province. MOImay also delegate to the province or district level any functions related to monitoring and control of the communes. Finally, the presence of the commune clerk, who i s a direct-hire employee of MOI, gives MOI a window into the day-to-day operations of the commune councils. 6.32 The provincial governors are required to prepare quarterly reports on the operation of the CSF for the MOL These, at a minimum, must include information on: (a) communes entitled to transfers, (b) communes complying with conditions of access and receiving transfers, (c) amounts transferred to each commune, (d) allocation of CSF resources in each commune for recurrent and investment expenditures, (e) specific types of investments supported by CSF resources, and (f) levels of own-source revenues mobilized by the councils. Based on this information, DOLA i s required to prepare comprehensive quarterly reports for the CSF Board (the CSF Board Secretariat i s required to prepare comprehensive reports for the RGC and all donors and intemational financial institutions that support the CSF Fund). Supporting DOLA's ability to establish a commune monitoring and evaluation system will be a key strategic measure for safeguarding the accountability of the decentralized system. Recommendations P Operationalize the Commune/Sangkat FundBoard. > Undertake further analytical work on commune expenditure assignment and alternative commune revenue bases-as inputs to the upcoming commune boundary review-before further developing the legal framework. P Provide greater support to DOLA for the establishment of a commune monitoring and evaluation system in order to safeguardthe accountability of the decentralized system. > The NCSC, with MEF in the lead role, should closely monitor the implementation of the planning, budgeting, and financial management processes with a willingness to adjust and/or simplify the systems and procedures, if necessary. DECONCENTRATION TO PROVINCESAND MUNICIPALITIES 6.33 Developing stronger deconcentrated sectoral offices at the provincial level is necessaryfor improving servicedelivery. Basic public services are being provided inadequately inmany parts of the country, and national poverty alleviation goals cannot be met without better performance. Enhancing the capacity of the provinces will improve delivery of national priority services, some aspects of which might eventually become formal commune responsibilitiesas the 126 Cambodia: IntegratedFiduciary Assessment and Public Expenditure Review system matures. Improving the functioning of the provinces, however, i s not simply an exercise in capacity building. Presently, many of the structures, systems and procedures in place for provincial operations are underdeveloped, and some are inneed of reform. A careful and detailed evaluation of the broader provincial fiscal and managerial structure i s needed, as i s special attention to the specific needs of particular sectors. As with the commune system, the provincial system cannot be developed in isolation, which appears to be a danger under the present RGC approach to reform. The relationship of the provincial to national public sector reforms and the emergingcommune system mustbe carefully considered. 6.34 At the same time, there is a need for an overarching deconcentration framework and analytical work on options for Cambodia. The RGC has not been clear about the objectives it hopes to achieve through deconcentration. I s the main intention to improve provincial and district capacity to perform functions on behalf of the RGC, or will provinces be given more autonomy? If the former, which functions should be the responsibility of the provinces, and how and when will they be developed? If the latter, how will political accountability to service users be ensured? Clarifying the direction of the reform will also inform the design of the provincial budget system. CAR, which i s developing a framework for deconcentration, needs to provide a clear set of signals and guidelines to the sectoral ministries for developing approaches and mechanisms, and needs to coordinate closely with line ministries and others on the development of the framework. This does not, however, mean that all ministries have to approach deconcentration in exactly the same way. CAR needs to do adequate background work to ensure that the principles it develops are broad and flexible enough to accommodate the likely diverse requirements of different sectors. 6.35 The legal foundation of deconcentrated government is straightforward, and is composed of the 1993 Law on State Financial Systems and the 1998 law on provincial finances, known as the "provincial budget law (PBL)." The Law on State Financial Systems defines vertical lines of accountability for sectoral departments and authorizes MEF and the NT/PTsto manage most provincial finances. The PBLdefines provinces as legal entities headed by governors and allows them to prepare annual budgets, which must be balanced, to make specified expenditures and collect specified revenues. Governors serve as budget officers for provincial departments financed through the RGC budget, and they also maintain a separate budget for their functions (the "Salakhet," or governor's budget). The Governor's budget 6.36 On the expenditure side, the greatest share of the governors' budget is spent on operating expenditures, primarily administrative operating costs (78.8 percent of the total). The Salakhet i s subject to the control of MOI, and as of 2001 is reflected in Government expenditure reports (viz., the TOFE). In2001, the Salakhet was equivalent to about 20 percent of the provincial component of the RGC budget. On the revenue side, tax revenues account for 48 percent of the total, with the tax on motor vehicles (17.4 percent), the excise on public lighting (12.3 percent), the wealth transfer tax (7.0 percent), and the business tax (6.8 percent) the most important (Annex A Tables A15 and A16 provide revenue and expenditure information from the 2002 Salakhet). Subsidies from the RGC budget are the second most important category of provincial revenues at 41.3 percent. There is, however, great variation in the importance of revenue sources across provinces. The second most important category i s provincial and municipal subsidies (9.2 percent), followed by salaries and allowances (7.6 percent). As with revenues, there i s considerable variation across provinces, but non-salary operating costs dominate inall cases. 6.37 Given the scarcity of resources, the Salakhet is not an insignificant amount of expenditure. Inorder to make sure that Salakhet expenditures are aligned with RCGpriorities as Reforming Institutions to Improve Service Delivery Ill:Decentralization and Deconcentration 127 outlined inthe NPRS, an efficiency analysis of these expenditures should be conducted (based on monthly reports provided to MEF), as it appears that nearly 79 percent of total spending i s for "administrative management." If and when, over the medium term, the role of the provincial administrationi s redefined, the need for and role of the Salakhet would need to be revisited. Provincial department expenditures 6.38 The provincial budget of the central government i s financed primarily from transfers from the central government. Transfers to the provincial budget have increased in recent years, from US$ 270,000 in 1998 to US$ 6.4 million in 2001, but still only account for 1.5 percent of Treasury-executed expenditure, with further increasesprogrammed for 2003 to 1.7 percent. 6.39 While it is difficult to give a complete picture of the structure of deconcentrated agency spending by level of administration, since expenditures of central agencies will include goods and services provided to provinces and sub-provincial administrations, the breakdown of central and provincial expenditures points to a high degree of centralization. Table 6.3 presents the provinces' share of civil administration spending allocated to institutions, excluding charges to central government, such as debt servicing. This shows a marked decline in the provinces share of Treasury-executed expenditures from 41.9 percent in 1996 to 28.6 percent in2001 (while risingindollar terms from US$ 69 million to just over US$ 90 million in2001), followed by a recovery to 39.1 percent inestimates for 2002. Most of this increase is accounted for by increases in provincial expenditures for just two functions: general administration and public works. Incritical service delivery sectors such as health, education and agriculture, where resources have to be dispersed to service providers at the field level, there i s a trend since 1996 toward the further concentration of resources in the hands of the central agency, indicating a lack of progress since the 1999 PER. Estimates for 2002 and programmed expenditures for 2003, however, suggest a slight reversal of this trend inagriculture and health. Taking into account PAP spending ineducation, which i s posted under central departments but actually spent by provinces and districts, there i s a slight improvement inthe level of deconcentration from 1999-2002. Table 6.3: Expenditure by Provincial Administrations as a Share of Total Expenditure by Function Function 1996 1997 1998 1999 2000 2001 2002E 2003B Core Government 5.9 5.3 3.9 3.0 3.1 3.9 8.2 11.3 GeneralAdministration 21.2 20.4 10.1 7.9 5.7 7.1 16.7 23.0 Judiciary 40.2 46.2 41.0 47.7 48.2 46.2 46.1 44.8 EconomicServices 18.0 11.6 13.2 11.6 12.0 9.8 23.4 22.1 Agriculture 37.3 27.4 31.7 40.3 28.3 20.0 31.2 30.9 Transport 19.8 10.5 17.0 16.0 7.0 5.6 40.9 34.4 Public Works 23.3 16.6 24.2 18.9 7.6 6.7 55.6 51.1 Other transport Other EconomicServices 8.9 6.0 5.9 4.4 8.6 6.9 12.4 12.1 EnvironmentalProtection 37.5 39.0 34.9 33.0 31.0 19.0 25.5 25.3 RuralDevelopment 22.7 29.3 19.4 11.9 32.7 19.0 39.5 40.4 SocialServices 73.9 68.9 73.1 57.0 62.6 60.0 56.2 54.7 Health 38.8 28.0 28.9 16.3 30.8 31.1 34.4 34.6 Recreation,Culture& Religion 36.3 31.5 30.0 28.7 12.2 11.7 19.7 17.9 Education 83.0 83.8 84.8 76.9 78.1 73.2 63.2 61.6 SocialProtection 95.5 93.0 95.6 94.4 86.9 87.4 83.8 85.0 Total 41.9 38.1 32.4 30.3 28.3 28.6 39.1 40.1 6.40 Provinces spend a much larger share of their budgets on personnel than the central agencies, and a commensurately smaller proportionon operations and maintenance. TOFE reports do not identify provincial capital expenditures, because they are centrally-executed, 128 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review though a significant share of the Government's capital expenditures are made at the provincial level, primarily for road rehabilitation. There has been some improvement in education and health, where PAP has facilitated the deconcentration of resources to cover operational costs at the field level. 6.41 Calculations show that there is no positive correlation between poverty incidence, or the number of poor, and provincial budget allocations for health, education, or other sectors (see Table 6.4). In fact, correlation coefficients are negative for 2002 (note that these calculations do not take into account PAP spending, which may improve the correlations). Allocations are largely a function of population size: the larger the province, the larger the allocation. However, per capita allocations for smaller provinces are much larger than those for larger provinces, presumably because the costs of running a provincial administration are much the same whatever the size of the province. While true of bothhealth and education, the impact of the administrative overhead i s much more apparent in other sectors, presumably because administrative overhead representsa larger share of total expenditures inthe other sectors. Inboth the health and education sectors, the ratio between budgeted personnel and operation expenditures varies by a factor of two or more between provinces. Table 6.4: Allocation of Health, Education and other Provincial Agency Budget Allocations by Province, 2002 (CR) Province Population % Health Per Other Salaries: Operations 000 Poor Capita Health Education Educ. PhnomPenh 919 12 2,849 18,179 37,123 1.97 0.17 Kandal 1045 18 4,816 15,723 281527 3,84 0.10 KampongCham 1575 12 4,277 12,504 23,110 4.26 0.07 Battambang 752 26 5,828 19,880 39,879 2.33 0.09 PreyVeng 929 53 4,758 14,987 27,003 4.83 0.15 SiemReap 664 54 5,343 11,392 27,552 2.82 0.05 KampongThom 547 29 5,280 14,256 32,806 3.29 0.14 Takeo 775 15 4,498 16,526 30,152 3.66 0.09 Svay Rieng 472 43 5,319 14,648 31,436 3,80 0.08 Pursat 344 41 5,612 15,549 40,628 2.99 0.16 KampongChhnang 403 45 5,755 14,691 36,336 2.79 0.14 KampongSpeu 579 18 4,911 13,750 30,824 3.25 0.09 Kampot 521 19 6,086 14,657 35,073 3.05 0.09 KrongPreahSihanouk 148 34 7,846 17,481 53,271 2.32 0.16 KohKong 106 8 8,570 10,947 49,463 3.77 0.5 1 PreahVihear 113 29 8,979 15,024 52,035 4.58 0.30 Kracheh 253 39 7,718 15,961 38,318 4.01 0.19 Ratanakiri 92 9 12,569 14,227 60,754 3.97 0.75 Mondol Kin 31 20 17,045 28,859 124,037 3.45 0.99 BanteayMeanChey 560 41 5,532 12,566 29,740 3.20 0.11 Stueng Treng 77 16 15,784 20,997 80,740 3.37 0.43 KrongKeb 20 49 20,325 58,943 213,415 4.35 0.86 Pailin 6 97 73,713 103,101 678,357 3.70 0.78 Otdar MeanChey 63 39 10,493 18,884 87,810 3.32 0.16 Total 10,995 28 5,315 15,189 33,610 3.33 0.12 Source:MEF budget data. 6.42 For the time being, key services, such as education, health and infrastructure, will continue to be managed by the provincial administrations of central agencies and receive their resources through the state budget. While the creation of the communes provides an opportunity for local initiatives and a framework for coordination of development efforts at the field level, they will not have a significant impact on the structure of spending. In this context, Reforming Institutions to Improve Service Delivery Ill:Decentralization and Deconcentration 129 further deconcentration of sector budgets and the expansion of streamlined budget execution mechanisms, such as PAP, are critical if additional resources are to be mobilized for service delivery. At the same time, improvements in resource targeting are needed so that the poorer provinces and those with lower than average service coverage, benefit from a larger share of resources. 6.43 Responsibility for the formulation and execution of sector budgets is divided between central ministries and their provincial departments. Ministries allocate the budget ceiling between their central agency and their provincial delegations in consultation with the MEF's DBFA.For the most part, the provincial department's ceiling is determined incrementally, taking into account increases in staffing levels, with little adjustment inthe regional structure from year to year. Provincial administration budgets are then prepared in conjunction with the Provincial Departments of Finance (PDAFs). Provincial budgets are compiled, executed, and monitored separately by the DBFA's provincial budget unit.At the beginning of each year, the line ministers delegate their authority as financial officers to the respective provincial governors, who are appointed by the King on the advice of the OPM. Consequently, while governors may intervene informally at the formulation stage, they have formal authority during budget execution. As all the players know that governors have an impact on actual budget execution, it is likely that gubernatorial influence extends back into the formulation stage, as it is in the interest of line ministries to negotiate agreementsduringformulation. 6.44 The logic of the present system is to balance central agencies' policy and technical leadership with the need to adapt to local conditions. However, the lack of clarity regarding the relative competencies of provincial governors and ministries under the system of dual subordination, as well as poor communications with the capital, tends to undermine coordination between provinces and central agencies during the budget process. Since funding for budget execution i s released by PDAFs, rather than the central agency, scarce cash is allocated on the basis of locally-defined priorities, leaving priority sectors starved of resourcesin some cases. This undermines the rationale for targeted budget support to the Government's priority sectors through conventional mechanisms, since there i s no guaranteethat allocations for the priority sectors will actually be implemented at the provincial level. It also makes it difficult for central agencies to hold their provincial delegations accountable for performance, since the resources needed to implement agreedprograms may not be made available. 6.45 The problem in education and most other sectors is that provincial spending, which accounts for a large share of total spending, might not be as tightly aligned with sectoral policy priorities. Nor is provincial spending as targeted to pro-poor service delivery as it ought to be (see Chapter 3). Two different approaches are beingpiloted ineducation and health to address the problem of technical inefficiency with respect to geographic resource allocation (see Box 6.1). Given the current state of centralization in Cambodia, it does not make sense for provincial budgets to be divorced from sectoral planning and budgeting processes. Improving technical efficiency in resource allocation will necessarily require better integration of provincial and central planning and budgeting. 6.46 The technical solution to these problems is to clarify the competencies of different levels of administration. The PBL went some way toward this by assigning provinces and municipalities exclusive competence for local services, including local roads, municipal services, and their offices. However, a formal division of competencies in other areas, such as education and health, would require policy decisions on the appropriate degree of cross-governmental decentralization that are likely to take many years to resolve. 6.47 Inthe meantime, progressis best achieved at a sectorallevel. Introductionof PAPhas enabled the health and education sectors to channel funds directly to provincial administrations, 130 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review by-passing the provincial budget execution system. Health has also made arrangements to earmark funding for provincial health delegations under the conventional budget mechanism. Similar arrangements could be made, administratively, for the remaining priority sectors. This would allow MEF and sector agencies to set targets for and monitor the deconcentration of resources. Sector earmarking at the provincial level would also allow line agencies to improve targeting of spending, by applying allocational formula based on indicators of need and/or performance. Box 6.1: ImprovingGeographicalTargeting inHealthand Education Inthe health sector, funds for operations, drugs, and supplies under Chapter 11& 13 are allocated between provinces on a formula basis. Fixed allocations are assigned to district health offices, district referral hospitals, and health centers. Provincial health offices receive a percentage mark-upbased on total operational district (OD) funding, and specific activities are allocated funds according to the number of patients using historical data. These criteria determine the block allocation for each province and OD. It i s important to note, however, that facilities in individual provinces do not necessarily receive these allocations and are dependent on the budget management of each OD office, and that resources to individual operational districts within a province are often divided up on the basis o f a local agreement. The advantage o f this formula-based approach to resource allocation i s that it i s clear and easily understandable. It also implicitly recognizes that many of the costs associated with health care provision are fixed and are incurred in sparsely populated areas to almost the same extent as in areas with large populations. However, the problem is that allocations are not strongly related to the quantity of health care need as measuredby either the size of the population or health status indicators. There are two main ways o f addressing this issue. The first is to ensure that the distribution of physical facilities better matches the needs and size of a population in a particular area. Infact the historic mismatch between population and facilities i s gradually being addressed through the 1996 health coverage plan, the recent sector strategy and supporting infrastructure development financed by government and external funding agencies. Inprinciple, if facilities are equally distributed standardized recurrent allocations can then be allocated to each facility. The problem i s that needs vary across populations of similar size and also fluctuate over time. Basing recurrent allocations on physical infrastructure may not be sufficiently responsive to these changes. A second approach i s to ensure that physical infrastructure i s as equitably distributed as possible but then allocate recurrent funding mainly on the basis o f population weighted for other need factors such as cost and health status. Education MOEYS has made sure that PAP (Chapter 13) i s derived from the estimated recurrent cost of the priority programs in the rolling ESSP. The PAP chapter i s prepared by MOEYS at the central level and there is a strong alignment between the education sector objectives and Chapter 13 budget allocations. Breakdowns by priority programs, and program allocations by Budget Management Centers (BMCs) and individual schools are made. Once the budget i s approved, each B M C implements its program. The PAP has introduced allocations to line managers and increased the programmatic transparency of budget allocations. The budget information i s widely available to all managers in the MOEYS and the donor community (although less so at the district and school levels). Although the ESSP estimates the capital and recurrent expenditure requirements to achieve the ESP objectives, linkages between the budgeting for Chapter 13 (PAP) and other budget chapters are not strong and they overlap. The problems with the capital budget have already been discussed. In addition, however, there i s no link between the ESP objectives and the budgeting of Chapters 11and 31. Inthe ESP cost estimates, a certain percentage increase in these chapters i s assumed. However, linkages are not made between these expenditures and the strategies and targeted outcomes of the ESPIESSP. PAP, as opposed to the other budget lines, seems to have a much greater responsiveness to sector priorities at both the central and provincial levels. PAP has clearly had a positive impact on the budget formulation process in the education sector. Recommendations > Develop a framework for deconcentration based on further analytical work and clarity of reform objectives. Reforming Institutions to Improve Service Delivery Ill: Decentralization and Deconcentration 131 9 Allow sector earmarking of some provincial expenditures to improve pro-poor targeting by applying an allocational formula basedon indicators of need andlor performance. KEY OPPORTUNITIESAND CHALLENGESINMOVING FORWARD :STRATEGY, CAPACITYBUILDING AND IMPLEMENTATION 6.48 Perhaps the greatest potential problems with the development and implementationof decentralization and deconcentration lie in the lack of attention to the coordinated development of a clear vision of where the system should be heading. Part of the problem i s that CAR has primaryresponsibility for deconcentration, while NCSC (chaired by MOI) takes the lead on decentralization. International experience suggests that such fragmentation of responsibility inevitably leads to conflicts and policy inconsistencies. Deconcentration and decentralization are closely linked, and this i s best reflected in an overarching and empowered coordination mechanism. Closer formal links between CAR and NCSC-which might take the form of ajoint work programwith key milestones for joint review-are highly desirable. 6.49 A related issue is the extent to which commune, provincial, and national planning are integrated effectively in pursuit of the WRS. Provincial and Commune Development Plans (PDP and CDP) necessarily have their own identity. PDPs are part of the national planning system and focus on the provincial attainment of national goals. The CDPs represent local priorities expressed through elected councils. This does not mean that national priorities are irrelevant for CDPs, but there must be some independence of development plans of different levels in a multi-level government system. Despite their separate identities, PDPs and CDPs are inescapably linked (e.g., Seila provinces have held planning integration workshops to provide an opportunity for provincialldistrict and commune representative to review and maximize the consistency of their development activities). Communes can also pass to provinces requests for priority services beyond their authority, means, or capacity. Similar planning integration, with appropriate modification, will be needed in the evolving RGC system, and this should extend to budgets. Presently provincial (gubernatorial) budgets barely cover basic operating expenses, so that provinces have little budgetary flexibility. As deconcentration proceeds, closer coordination of provincial and commune budgets will make more sense (with the caveat that they should remain legally separate). 6.50 The problem of capacity at all levels remains a great challenge to designing and implementing decentralization and deconcentration reforms. Provincial staff need skills for their own functions as well as to support the commune system. The commune councils and staff, including counselors, sub-committee members, and clerks, must also be trained to meet their obligations. The costs of capacity buildingand technical assistance are likely to be enormo~s."~ Further work i s needed to establish the reliability of these estimates, to determine their implication for the implementation strategy, and to create a credible basis for securing additional donor resources. 6.51 At the provincial level, RGC plans for the decentralizationsupport structure are not completely clear. Provincial Rural Development Committees (PRDCs) will continue to coordinate commune capacity buildingand technical assistance.However, the longer term role of the PRDCs i s far from clear in light of the newly created Provincial Offices of Local Administration (POLAs), which have been created as the main interface between provinces and communes. Institutional restructuring based on clear functional allocations of responsibilities i s It was recently estimated that over the period 2002-2005, US$ 6 million will be required for national program support for decentralization, about $15 million for provincial program support for the communes, and about $4.5 million for commune level capacity building and public educationon decentralization.Of this total of $25.5 million, approximately$13.5 million is yet to be coveredby either domestic contributionsor committedexternalresources. In addition,the RGC estimatesthat $13 million is requiredfor basic communeoffice facilities andequipment. 132 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review an important part of broader public sector reform, given the scarcity of resources and the need to maintain a right-sizedcivil service. 6.52 Related to concerns about integration of reforms and structures is the role of the villages. Although not elected levels of government, villages are an important linkto localpeople and are expected to play arole indevelopment planning.The RGC i s still considering how to deal with villages in the context of the rural development institutional structure. It appears that Village Development Committees (VDCs) will continue to exist where they have already been created, but it is not clear that new ones will be developed. The decentralization regulations allow commune councils to appoint a village chief and the village chief to appoint deputies, raising concerns about whether people appointed in this way will feel accountable to villagers. Clarification of the role of villages and development of a joint position by MRD and M O I are necessary. 6.53 Lastly, bolstering and safeguarding the democratic integrity of the commune system is an important challenge for the Government. The councils in principle have a fair degree of autonomy provided they satisfy certain legal requirements. There is, however, considerable potential for interference. The only technical staff person at the commune level, the clerk, i s an employee of MOI, and the role of council accountant i s played by a staff member of the PT. In addition, many communes will for the foreseeable future be dependent on technical assistance from provincial sectoral departments. Finally, governors have substantial commune oversight powers. Given the hierarchical mentality prevailing in the civil service and the institutional weaknesses in the system, even with defined legal safeguards and the very best of intentions, there are ample opportunities for interference in commune affairs. Higher-level oversight is clearly appropriate, but care must be taken to balance the vertical accountability of the councils with their accountability to the people who elected them. Existing legalprovisions do this to some extent, but whether the rules can be enforced i s less clear. This is a delicate and complex matter in practice and requires careful consideration and monitoring. Recommendations 9 Undertake further work to improve the reliability of capacity buildingrequirement costs, to determine implications for the implementation strategy, and to create a credible basis for securing additional donor resources. 9 Rationalize strategy for institutional support of communes, clarifying division of labor betweenPRDCs and POLAs as necessary. 9 Clarify and develop ajoint M R D M O Iposition on the role of villages. 7. CONCLUSION: TOWARDA STRATEGY OFREFORM 7.1 This review has argued that Cambodia has to meet a number of key challenges- fiscal, fiduciary, and institutional-over the medium term if it is to make significant progress on its poverty reduction and development plan. First, Cambodia will have to improve resource mobilization to ensure aggregate fiscal sustainability. Second, to reduce fiduciary risk to public funds, the Government will have to engage in comprehensive reform of budget execution, cash management, and public financial control systems. Third, the Government will have to rationalize public expenditure policy and management further to carry out its National Poverty Reduction Strategy (NPRS). Last, Cambodia will have need to undertake comprehensive civil service reform-focusing on pay and employment issues-in order to deliver poverty-reducingservices. 7.2 In broad terms the fiscal, fiduciary, and institutional challenges identified in this review have been on the Government's reform agenda for some time. To place this analysis in context, this brief section assesses progress made to date in implementing the recommendations of the 1999 World Bank Public Expenditure Review (PER), based on the premise that understanding progress as well as obstacles to progress may shed some light on how to craft areformstrategy (see Annex C. Cambodia 1999PER-Summary of Achievement). 7.3 There is no doubt that the Government has made progress-significant in some areas, less impressive in others-in all major reform areas flagged by the World Bank's 1999 PER: domestic resource mobilization, rationalizing public expenditure policy, and enhancing public expenditure management.12' The most notable areas of progress are those in which central oversight agencies and line ministries partnered to implement reform. The best example of this i s the Priority Action Program in education and health, which has moved line agencies toward program budgeting, improved the targeting of resources to the poor, and may also have had a positive impact on outcomes. To be sure, further improvement i s required, but there has been progress. Less impressive progress has been made in forestry, for example, where reform seemed to proceed well to the point of policy but floundered when it came to implementation. Other reforms inthe areas of civil service and tax policy and administration fall between the two extremes. The Government undertook some civil service reform, but overall progress has been too slow, relative to the importance of the problem. On tax policy and administration there was some progress inrevising the Law on Investment and strengthening tax administration, but many would argue that measures fell short of resolving the core problems. 7.4 One factor which may explain the differential in success is the level of consensus about the importance and nature of the reform. On PAP there was broad agreement between the Government (central and line ministries) and the donors that getting resourcesto the front line of service delivery was crucial. On civil service reform there was not such agreement between either central and line agencies or between Government and donors on the nature of the reform needed. In the area of tax policy there was also less consensus between Government and donors. For example, there were legitimate differences of opinion about the benefits of tax incentives. It It shouldbepointedout that the 1999 PER i s usedto ground this brief assessment, thoughit i s understoodthat many other donors (including the IMF,ADB, UNDP, andthe bilaterals), documents (including the SEDP and SEDP 11), and instruments(includingthe PRGF, SAC, andother projectandprogramloans) are integralpartsof thereformprocess. 133 134 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review seems that to the extent that technical issues are unresolved in a convincing manner, reform will be undermined. Another factor that might explain success in the case of PAP i s the relative simplicity-at least at the conceptual level-of the reform itself, which was especially important given weak bureaucratic capacity. This suggests the critical importance of building technical consensus as a foundation for reform, and of keepingreformsimple. 7.5 Forging consensus means taking a participatory approach to analysis and problem solving, as inthe case of this review. For example, Government, the Bank, the ADB, and others donors, notably the IMF,DFID,and SIDA, participatedinreviewing the concept paper, the terms of reference for individual studies, and the subsequent drafts of individual sections (some also provided financial support). Several workshops were also held to facilitate discussion between Government and donors, and within Government (some line agencies seemed to place great value on the workshop sessions in which they could communicate directly with central oversight ministries). More, however, remains to be done. First, donors themselves should take greater responsibility for coming to consensus. A review such as this offers the donor community the chance to explore options and arrive at consensus. Fragmentation in the donor community reduces the probability of reform. The Government itself needs to do more to forge consensus between central oversight (e.g., CAR and MEF) and the line agencies. This includes improving the sharing of information between agencies, and with donors. 7.6 The second hypothesiethat the Government tends to rely on ad hoc reform approaches for systemic problems-reflects the weak capacity of the public sector to undertake change, which has been noted throughout this review. The principal examples of the ad hoc approach are the Priority Action Program-as a parallel budget systemoperating in a small set of ministries-and the Priority Mission Group program-as a parallel salary system in which relatively highsalary supplements are provided to small groups of individuals working on priority tasks. The limitedexisting capacity makes these types of approaches inevitable. They are not without their potential costs, however. The first potential cost i s that these ad hoc approaches are not embedded in long term reform strategies. To the extent that these measures become isolated from the bigger reform picture, instead of paving the way toward more comprehensive reform, they could ultimately undermine or postpone systemic reform. Paradoxically, the more successful these ad hoc measures are in solving critical problems, the less urgent systemic reform becomes. The second potential cost i s that these limited reforms drain existing capacity from more comprehensive reform. 7.7 The suggested solution, then, is to make sure that ad hoc measures are part of a larger reform program. This i s now what the Government, together with donor partners, is developing for PAP. This review, for example, suggests a reformstrategy for mainstreaming PAP and consolidating the different budget systems. Ideally, ad hoc measures would be part of a comprehensive and sequencedreform strategy from the start. The strategy would then be updated periodically to reflect recent experience. If, for example, the Government decides to move forward with a semi-autonomous revenue authority, it should make sure that the reformshould be embedded ina larger mediumterm tax reformstrategy. 7.8 Though the IFAPER focuses more on institutional issues-public expenditure and financial management, the civil service, and decentralization-the broad themes identified in the IFAPER and the 1999 PER are consistent.This is not surprising, since weak capacity means that reforms will have to be gradual. It does mean, however, that to be most useful to the Government, donors will have to assist inthe development of strategies to overcome institutional obstacles, which entails participatory approaches and consensus building.Ultimately, this review will only be successful to the extent that it builds on the existing participatory process as donors move forward to support Government inimplementingthis set of recommendations. Conclusion toward a Strategy of Reform 135 7.9 The analysissuggeststhree guidelines for movingforward with a public expenditure policy and management reform agenda. First, donors and Government should continue the participatory approach with which this review began. Donor-donor and Government-donor consensus is critical for ensuring success, and a premium should be put on negotiated consensus. Second, any ad hoc or pilot reformmeasures must be placed, from the outset, inthe context of a broad reform strategy. Third, reforms must be simple-and tailored to bureaucratic constraints and political incentives. Greater attention to the politico-bureaucratic context of technical reforms should enable reformers to craft measuresmore likely to be successful. ANNEX A Table A1 Cambodia: Medium-TermMacroeconomicFramework (inpercent of GDP,unlessotherwise indicated; basedonthe 2002National Accounts) 1999 2000 2001 2002 2003 2004 200.5 2006 2007 Projections Real Sector RealGDP (percent change) 6.9 1.7 6.3 4.5 5.0 6.0 6.0 6.3 6.5 RealPer capita GDP (inCR, percentagechange) 4.2 5.O 3.7 2.5 2.9 3.9 2.9 3.2 3.4 CPI Inflation (average: percentagechange) 4.0 -0.8 0.2 3.3 3.8 3.5 3.5 3.5 3.5 National saving 13.6 9..2 12.7 13.7 13.9 14.1 14.1 14.4 14.6 Government saving 1.6 1..5 1.2 0.9 1.1 1.3 1.3 1.3 1.2 Private saving 12.0 7.7 11.5 12.8 12.8 12.8 12.8 13.1 13.5 Domestic investment 15.9 13.5 17.9 16.6 16.9 17.4 17.7 17.8 18.0 Government investment 5.7 6.7 7.2 7.7 7.1 7.0 6.9 6.8 6.8 Private investment 10.2 6.8 10.7 8.9 9.8 10.4 10.8 11.0 11.2 Of which: change in inventories 1.4 -1.7 2.0 -- _ _ __ _ _ __ -- Fiscal Sector Revenue 10.6 11.2 11.7 12.1 12.7 13.4 13.9 14.3 14.5 Tax revenue 7.7 8.2 8.4 8.5 8.5 8.7 8.8 8.8 8.6 Of which: Domestic Taxes 1.1 1.8 1.8 1.9 1.8 1.8 1.8 1.8 1.8 Int.Taxes 6..3 6.2 6.3 6.5 6.6 6.8 6.8 6.8 6.1 Nontax revenue 2.8 2.8 3.2 3.5 3.6 3.6 3.4 3.5 3.6 Of which quota auctions 0.7 0.2 0.3 0.4 0.5 0.5 0.0 0.0 0.0 CapitalRevenue 0.1 0.2 0.1 0.1 0.2 0.2 0.3 0.3 0.3 Revenue Measures _ _ _ _ _ _ -_ 0.4 0.8 1.4 1.7 2.0 Expenditure (cash basis) 14.7 16.4 17.7 18.9 19.0 19.1 19.4 19.7 20.2 Current 8.9 9.5 10.4 11.1 11.4 11.9 12.2 12.6 13.1 Interest payments 0.1 0.1 0.2 0.2 0.5 0.7 0.7 0.7 0.7 Other 4.6 5.4 6.6 7.0 7.0 7.1 7.2 1.5 7.8 Memo item: social spending li 2.0 2.3 2.9 3.6 3.8 4.3 4.4 4.5 4.5 Capital 5.8 6.9 7.3 7.8 7.3 7.2 7.1 7.1 7.1 Current balance 1.6 1.5 1.2 0.9 1.1 1.3 1.3 1.3 1.2 Overallbalance -4.0 -5.2 -6.0 -6.8 -6.3 -5.7 -5.5 -5.5 -5.6 Overallbalance (including grants) -1.3 -2.3 -3.1 -3.3 -3.4 -2.9 -2.7 -2.6 -2.8 Domestic financing -0.3 -0.1 0.1 -0.3 -0.2 0.0 0.0 0.0 0.0 Externalfinancing ' 4.4 5..3 5.9 7.1 6.5 5.1 5.5 5.5 5.5 Disbursements 4.4 5..3 6.0 7.2 6.8 6.0 6.0 6.0 6.0 Amortizations 0.0 0.0 0.0 0.1 0.3 0.3 0.5 0.5 0.5 ExternalSector Domestic exports (percentage change) 34.5 30.9 7.4 10.7 9.6 9.7 10.9 10.3 10.5 Retainedimports (percentage change) 27.7 36.5 6.8 5.8 8.6 9.8 9.8 9.9 10.1 Retainedimports (excl. garments, pct. change) 16.8 29.8 3.8 6.6 9.2 12.3 12.1 11.9 12.1 Current account balance(excluding transfers) -9.0 -12..3 -13.4 -10.7 -10.4 -10.5 -10.8 -10.4 -10.0 Current account balance(includingtransfers) -2.3 -4.3 -5.3 -2.9 -3.0 -3.3 -3.6 -3.4 -3.4 Foreign direct inv. (inmillions of US$) 146 135 95 98 101 110 120 131 144 Overallbalance -2.8 -1.7 0.6 1.4 0.4 0.4 0.9 1.8 2.0 Gross official reserves (inmillions of US$) 422 485 550 642 696 742 810 904 1,011 (inmonths of imports ofgoods & services) 3.2 2.8 2.9 3.3 3.3 3.2 3.2 3.3 3.4 Net capital flows 3i 8.5 10.8 11.9 10.6 11.4 10.2 10.1 10.1 9.6 Financing gap (in millions of US$) 0 0 0 0 31 20 15 0 1 Extemal debt 4/ 68.7 67.6 66.2 65.6 35.2 36.0 37.0 38.3 39.4 Extemal debt (NPV) 4i 56.1 56.0 55.5 52.9 21.8 22.0 22.2 23.0 21.6 Externalpubl.debt-serv. ratio, accrual basis 11.1 8.5 3.3 3.0 2.6 2.6 2.4 2.2 2.3 Externalpubl. debt-serv. Ratio, cashbasis 0.7 0.8 1.1 1.1 2.1 2.6 2.4 2.2 2.3 Sources: Dataprovided by Cambodian authorities and Fund staff estimatesandprojections. "Includes components of wagespaid in the social sectors. Includes disbursementsfor thefinancing gap in 2003. Net oficial disbursement,exceptionalfinancing and oficial transfers. 4/Figures include bilateral debt with the Russian Federation and the United States and reflects the impact of completing rescheduling agreementson Naples terms with these creditors in mid 2003. " A s percent of domestic exports of goods and services. The declines in 2000 and 2001 reflect the tailing off of payments to the Russian Federation. 6/As percent of domestic exports of goods and services. 137 138 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review Table A2. Public Expenditure Cambodia and International Comparators, % GDP Comparators Total Health Education East Asia & Pacific 15.0- 1.8 4.0 South Asia 16.7 1.2 3.O Low Income 18.4 1.2 3.8 Vietnam 21.2 1.o 2.8 Lao PDR 2.3 2.4 Cambodia, incl. External 25.3 2.9 3.0 Cambodia, Government only 12.8 1.o 1.7 Source: World Development Indicators, 2002. Table A3. Treasury-ExecutedSpending by Function, % Structure Function 1996 1997 1998 1999 2000 2001 2002E 2003B Core Government 62.6 63.2 69.4 54.7 58.9 51.0 41.0 37.4 General Administration 17.0 15.7 26.0 19.0 29.1 25.7 18.7 17.5 Defense 32.5 32.7 29.5 25.2 20.2 16.5 14.2 12.6 Security 12.8 14.6 13.7 10.1 9.3 8.5 7.6 6.8 Judiciary 0.3 0.3 0.2 0.3 0.4 0.4 0.5 0.5 Economic Services 10.8 14.1 9.3 14.2 12.6 16.9 8.3 8.0 Agriculture 2.1 2.6 1.7 1.9 2.3 3.9 3.0 2.7 Transport 3.1 4.1 1.8 2.6 5.1 6.4 1.1 1.o Public Works 2.6 ,2.6 1.2 2.2 4.7 5.4 0.8 0.7 Other transport 0.5 1.5 0.5 0.4 0.4 1.o 0.3 0.3 Other Economic Services 5.5 7.3 5.7 9.5 4.8 6.1 3.9 3.8 EnvironmentalProtection 0.2 0.2 0.1 0.2 0.3 0.4 0.4 0.4 RuralDevelopment 0.5 0.4 0.6 1.2 0.8 1.8 1.1 1.o Social Services 19.5 21.2 18.8 26.4 24.8 28.1 31.8 30.2 Health 4.8 5.9 4.2 9.4 6.7 7.9 9.4 9.5 Recreation, Culture & Religion 0.6 0.8 0.6 0.8 1.8 1.9 1.4 1.4 Education 9.1 9.1 9.4 11.4 10.7 12.9 15.8 14.6 Social Protection 5.0 5.5 4.6 4.8 5.7 5.4 5.3 4.7 Other 6.5 1.0 1.9 3.5 2.9 2.2 17.8 23.5 Debt 6.5 1.0 1.9 2.5 2.0 1.8 1.3 2.5 Other not classified - 1.1 0.9 0.5 16.4 21.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: MEF TOFEReports. AnnexA 139 Table A4: Treasury-Executed Civil AdministrationExpenditureby EconomicCategory, as % of GDP Budget Chapter 1996 1997 1998 1999 2000 2001 2002E 2003B Sub-TotalCurrentExDenditure 4.6 4.3 4.2 5.0 5.9 7.4 8.2 8.6 10: Salaries andAllowances 1.5 1.5 1.3 1.5 1.7 1.7 2.2 2.1 11: OperatingCosts 1.9 1.8 1.7 2.1 2.8 3.5 3.2 3.0 12: Subsidies for Prov.Admin. 0.2 0.2 0.1 0.1 0.2 0.2 0.4 0.7 13: SpecialProg.Agreements 0.0 0.0 0.3 0.0 0.1 0.6 1 .o 1.o 20: InterestonLoans 0.2 0.1 0.1 0.2 0.2 0.2 0.2 0.2 30: Econ.Transfer Payments 0.2 0.1 0.1 0.1 0.2 0.3 0.3 0.2 31: SocialTransfer Payments 0.6 0.7 0.5 0.7 0.7 0.9 0.8 0.7 32: Transfer to Int.Org. 0.1 0.0 0.0 0.0 0.0 0.1 0.1 0.1 40: Miscellaneous 0.0 0.1 0.1 0.1 0.1 0.6 41: Contingencies 0.1 Sub-totalCapitalExpenditures 1.3 1.o 1.1 1.8 2.4 2.2 2.0 2.3 50.1: Cons. & Equipment 0.6 0.8 0.7 1.5 1.7 1.4 1.5 0.9 50.2: Counterpart Funds 0.1 0.2 0.3 0.2 0.6 0.4 0.4 0.4 50.3: Inv. by ForeignFunds - 0.3 0.1 0.8 52: FinancialOperations 0.0 53: Debt Amortization 0.5 0.1 0.0 0.1 0.1 0.1 0.2 Total 5.9 5.4 5.3 6.8 8.3 9.6 10.2 10.9 Source: MEF TOFEReports and Bank estimates. TableA5: Ratioof OperatingExpendituresto Salaries andAllowances by Function Function 1996 1997 1998 1999 2000 2001 2002E 2003B Core Government 1.67 1.94 3.01 2.57 3.21 4.32 2.52 1.99 GeneralAdministration 1.67 1.93 3.05 2.54 3.19 4.32 2.52 1.97 Judiciary 1.71 2.22 1.26 3.73 4.02 4.29 2.53 2.59 Economic Services 2.78 2.35 2.44 3.26 3.02 4.03 2.87 3.37 Agriculture 1.53 1.07 1.15 1.67 2.15 3.23 2.52 2.78 Transport 0.54 0.53 0.42 0.84 0.92 1.24 0.92 1.49 Public Works 0.41 0.34 0.28 0.84 0.95 1.30 0.96 1.22 Other transport 0.94 1.10 0.80 0.81 0.85 1.14 0.84 2.21 Other EconomicServices 7.10 6.48 6.69 8.62 6.71 7.22 5.03 5.43 EnvironmentalProtection 2.91 1.73 1.50 1.84 2.40 4.62 2.54 2.88 RuralDevelopment 1.43 1.12 1.19 1.54 4.25 7.35 6.63 6.56 SocialServices 0.80 0.69 0.58 0.80 0.99 1.36 1.47 1.50 Health 3.63 3.28 3.10 5.15 6.34 8.31 6.20 6.20 Recreation, Culture& Religion 0.78 0.99 0.75 1.34 2.58 2.92 2.79 3.29 Education 0.34 0.24 0.23 0.30 0.29 0.53 0.71 0.70 SocialProtection 1.44 1.29 0.77 1.49 2.36 3.59 3.04 3.09 Total 1.29 1.24 1.42 1.44 1.73 2.36 1.91 1.85 Source: MEF TOFEReports and Bank estimates. 140 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review Table A6: External Assistance as Percentageof Total Sector Expenditure Summary Function 1996 1997 1998 1999 2000 2001 Core Government 19.8 23.3 20.8 9.2 7.8 9.0 Economic Services 37.5 27.6 25.9 26.2 27.0 24.6 Agriculture 14.5 4.8 2.9 6.8 10.3 8.3 Transport 13.5 10.0 11.0 9.1 11.0 14.0 Other Economic Services 8.7 11.2 11.1 9.5 5.2 2.0 Environmental Protection 0.8 1.6 0.8 0.8 0.5 0.2 A r e a u r a l Development 17.6 18.2 14.8 15.6 15.7 14.5 Social Services 22.3 26.6 36.2 36.5 33.8 42.5 Health 9.8 8.6 14.8 19.0 15.8 15.5 Education 7.8 13.0 13.7 10.8 9.5 10.6 Community & Social Services 4.7 5.1 7.8 6.6 8.5 16.4 Humanitarian Aid & Relief 2.8 4.3 2.3 12.7 15.6 9.4 Total 100.0 100.0 100.0 100.0 100.0 100.0 Source:Bank estimates based on CDCDatabase and MEF TOFE. Table A7: Treasury-ExecutedCivil Administration Expenditure by EconomicCategory, as % of GDP Function 1996 1997 1998 1999 2000 2001 2002E 2003B Core Government 1.67 1.94 3.01 2.57 3.21 4.32 2.52 1.99 General Administration 1.67 1.93 3.05 2.54 3.19 4.32 2.52 1.97 Judiciary 1.71 2.22 1.26 3.73 4.02 4.29 2.53 2.59 Economic Services 2.78 2.35 2.44 3.26 3.02 4.03 2.87 3.37 Agriculture 1.53 1.07 1.15 1.67 2.15 3.23 2.52 2.78 Transport 0.54 0.53 0.42 0.84 0.92 1.24 0.92 1.49 Public Works 0.41 0.34 0.28 0.84 0.95 1.30 0.96 1.22 Other transport 0.94 1.10 0.80 0.81 0.85 1.14 0.84 2.21 Other Economic Services 7.10 6.48 6.69 8.62 6.71 7.22 5.03 5.43 Environmental Protection 2.9 1 1.73 1.50 1.84 2.40 4.62 2.54 2.88 Rural Development 1.43 1.12 1.19 1.54 4.25 7.35 6.63 6.56 Social Services 0.80 0.69 0.58 0.80 0.99 1.36 1.47 1.50 Health 3.63 3.28 3.10 5.15 6.34 8.31 6.20 6.20 Recreation, Culture & Religion 0.78 0.99 0.75 1.34 2.58 2.92 2.79 3.29 Education 0.34 0.24 0.23 0.30 0.29 0.53 0.71 0.70 Social Protection 1.44 1.29 0.77 1.49 2.36 3.59 3.04 3.09 Total 1.29 1.24 1.42 1.44 1.73 2.36 1.91 1.85 Source:MEF TOFEReports and Bank estimates. e e-*- 3-m-*- N" AnnexA 143 TableA10. TotalGovernmentandExternalExpenditureon Education (CR millions) 1997 1998 1999 2000 2001 2002 2003 1-GeneralEducation 9,312 10,055 26,564 26,736 47,470 94,210 110,700 Recurrent 7,002 9,912 20,613 22,691 31,117 94,210 110,700 Ch 10Wage 1,621 1,849 1,991 2,202 1,229 1,450 2,140 PAPWage 10,752 12,400 Non Wage 5,381 8,063 18,622 20,489 29,888 82,008 96,160 O&M 4,605 7,866 18,486 20,29 1 17,551 22,300 24,420 PAP 144 12,163 59,383 71,300 Subsidies 776 197 136 54 174 325 440 Capital 2,310 143 5,951 4,045 16,353 2-Higher Education 4,362 5,156 8,550 9,001 11,732 15,000 18,768 Recurrent 4,011 4,986 8,550 9,001 11,732 15,000 18,768 Ch 10Wage 1,363 1,858 2,748 2,763 2,567 3,010 4,245 NonWage 2,648 3,128 5,802 6,238 9,165 11,990 14,523 O&M 2,412 2,726 5,331 5,826 6,562 7,200 6,700 PAP 2,000 3,920 7,000 Subsidies 236 402 471 412 603 870 823 Capital 351 170 3-Youth andSports 1,817 2,452 3,470 6,793 8,535 10,000 12,551 Recurrent 1,735 2,411 3,470 6,793 7,540 10,000 12,551 Ch 10Wage 358 420 577 450 553 580 831 NonWage 1,377 1,991 2,893 6,343 6,987 9,420 11,720 0&M 1,002 1,228 2,153 5,545 5,416 7,590 8,300 PAP 860 1,000 2,500 Subsidies 375 763 740 798 711 830 920 Capital 82 41 995 9-Provinces 70,755 84,716 117,148 127,330 161,816 166,990 190,981 Recurrent 70,655 84,716 117,148 127,330 161,816 166,990 190,981 Ch 10Wage 61,424 75,919 106,656 115,613 127,739 145,000 173,091 NonWage 9,231 8,797 10,492 11,717 34,077 21,990 17,890 O&M 8,180 7,762 9,148 6,074 18,391 18,000 13,900 PAP 2,572 13,065 Subsidies 1,051 1,035 1,344 3,071 2,621 3,990 3,990 Capital 100 Total Domestic 86,246 102,379 155,732 169,860 229,553 290,200 337,000 RECURRENT 83,403 102,025 149,781 165,815 212,205 286,200 333,000 Ch 10Wage 64,766 80,046 111,972 121,028 132,088 160,792 192,707 NonWage 18,637 21,979 37,809 44,787 80,117 125,408 140,293 Compositionof Non-Wage 18% 72% 18% 79% 57% 12% 0&Mfor Administration 16,199 19,582 35,118 37,736 47,920 55,090 53,320 PAP 2,716 28,088 64,303 80,800 Subsidies 2,438 2,397 2,691 4,335 4,109 6,015 6,173 CAPITAL 169,343 180,354 158,676 162,384 131,984 94,700 89,700 Govt. Capital 2,843 354 5,95 1 4,045 17,348 4,000 4,000 Extemally Financed 166,500 180,000 152,725 158,339 114,636 90,700 85,700 Total Govt. andExternal 252,746 282,379 308,457 328,199 344,189 380,900 422,700 144 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review Table All: Governmentand Private Recurrent Expenditure by Levelof Education 1997 1999 2000 2001 2002 Total Government andPrivate Total 231,908 337,435 341,441 394,948 481,638 Primary 142,490 217,327 209318 246,671 281,308 Lower Secondary 53,532 64,949 63,306 77,840 103,061 Upper Secondary 20,3 15 36,147 44,590 43,689 49,870 Post Secondary 7,967 13,551 15,032 18,532 31,514 Other 7,605 5,461 8,995 8,217 15,885 PrivateShare (%) Primary 65% 55% 50% 42% 37% Lower Secondary 67% 56% 55% 54% 50% Upper Secondary 80% 71% 70% 70% 67% Post Secondary 50% 37% 40% 37% 24% Other 0% 0% 0% 0% 0% Total Private 37% 38% 37% 38% 39% Recurrent 64% 56% 53% 49% 46% Primary 65% 55% 53% 47% 45% Lower Secondary 67% 56% 55% 54% 51% Upper Secondary 80% 71% 70% 70% 67% Post Secondary 50% 37% 40% 37% 25% Govt. RecurrentExpenditure Total 83,403 149,781 165,815 212,205 286,200 Primary 50,006 96,839 105,722 143,119 178,625 Lower Secondary 17,680 28,558 28,757 35,852 51,183 UpperSecondary 4,101 10,373 13,340 13,286 16,507 Post Secondary 4,011 8,550 9,001 11,732 24,000 Other 7,605 5,461 8,995 8,217 15,885 PrivateExpenditure Total 148,505 187,654 175,626 182,743 195,438 Primary 92,484 120,488 103,796 103,552 102,683 Lower Secondary 35,851 36,391 34,549 41,988 51,878 UpperSecondary 16,214 25,774 31,250 30,403 33,364 Post Secondary 3.956 5,001 6,03 1 6,800 7,514 Source: Bank calculations using MEF TOFE for 1997-2001; Budget proposal for 2002; 1999 Poverty Assessmentfor distribution of enrolment by quintilefor base year 1997. Sources: Bank calculations using MEF TOFE for 1997-2001; Budget proposal for 2002; 1999 Poverty Assessmentfor distribution of enrolment by quintilefor base year 1997. 4 4 4 3 AnnexA 147 TableA14. Cross Country Comparisonsof Coverage,MostRecentYear Primary Secondary PostSecondary Gross Net Gross Gross Enrolment Enrolment Completion Enrolment Enrolment Country Ratioa/ Ratiob' Rate Ratio Ratio Cambodia 124 86 60 24 1.4 Lao PDR 111 72 64 29 2.8 Mongolia 88 82 56 36.0 Vietnam 110 94 99 61 6.9 China 120 100 100 69 5.7 Indonesia 114 97 91 11.3 Malaysia 91 91 90 58 11.7 Philippines 117 100 92 79 35.2 Thailand 88 84 57 22.1 Hong Kong 96 91 28.0 Japan 102 100 100 92 42.7 Korea, Rep 94 92 102 60.3 Sources: World Bank Edstats. UNESCO 1999 Yearbook. MRY data are between 1997-1999. Cambodia is 2001/02. Mongolia -Public Expenditure and Financial ManagementReview, World Bank, 1999. The gross enrolment ratio is the total enrolment divided by thepopulation of oficial school age (e+. 6-11 in primary, 12-17 in secondary) b/ The net enrolment ratio is the enrolment of youths in the oficial school-aged range divided by the oficial school-agedpopulation. 3 .3 : Q '5 k c, Q) rr 0 rr 0 Y a &J P 1 .I 3 m 1 W m aa r% Y W rr 0 95c '5 s 150 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Reliew Table A17. BudgetVariance and Backloading2000 and2001 ~~~ ~ EconomidSectorStructure 2000 2001 2002 Central Provin- Central Provin- Central Provin- cial cial cial Recurrent-Defense & Security 99.1 99.1 94.4 Recurrent-Civil Administration 86.7 93.2 100.1 97.8 75.9 94.7 Ch. 10: Salaries 104.9 94.8 99.5 92.6 120.5 103.6 Ch. 11.Operating Costs 99.7 82.1 112.5 92.7 76.2 80.2 Ch. 13: Specific Program Activities 95.4 76.3 98.5 67.4 96.0 Ch. 30: Economic Interventions 66.3 95.8 81.8 91.6 116.7 88.6 Ch. 31: Social Interventions 88.5 91.9 83.1 97.2 125.4 90.3 Capital 94.8 87.1 82.6 Total Treasury- Executed Expenditure 92.4 96.1 84.6 Education 64.2 98.5 61.3 105.3 43.5 91.6 Health 86.8 78.8 93.9 86.5 47.0 97.9 Agriculture 91.4 88.2 84.8 87.6 81.7 87.1 Rural Development 95.1 91.6 94.8 86.2 77.8 81.2 Source: Bank estimatesbased on MEF TOFE. . . . . . rc 0 E . . . 0 E . . . . 0 . 5G2s 3 i -c,V % cd c e, e -B e e e e e e 0 I; ANNEX C CAMBODIA Country FinancialAccountability Assessment East Asia and Pacific Region Central Operational Services Unit World Bank, Washington D.C. 161 162 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review PREFACE The Country Financial Accountability Assessment (CFAA) is a diagnostic study of the financial management of a country. It provides an objective assessment of the strengths and weaknesses of financial management operation, identifies the key risks and capacity gaps, and assists the government in developing a strategy for reforms and an action plan to improve financial management capacity inthe public and private sectors. The findings and recommendations of the CFAA are in agreement with the diagnostic and recommend actions of other Bank work that has provided assistance to Government in the specific area of public finance, in particular the Country Assistance Strategy (CAS) and the HIPC. The substantial part of the findings and recommendations contained inthe CFAA provide strong support to the Government efforts to improve the management of public finance and to the preparation of the Poverty Reduction Strategy Credit (PRSC). This CFAA was carried out by a World Bank mission in July 2002 under the umbrella of the Integrated Fiduciary Assessment and Public Expenditure Review (IFAPER). An earlier note was discussed at the first IFAPER workshop, held in October 2002 in Phnom Penh, attended by representatives of the Ministry of Economy and Finance, the line ministries, the International Monetary Fund, and the Asian Development Bank. This report was prepared on the basis of the findings of a World Bank mission to Cambodia from July 29 to August 8, 2002, carried out by Mr.Wijaya Wickrema, Senior Financial Management Specialist and Ms. Agnes Albert-Loth, Senior Finance Officer. Mr. David Shand and Mr. Pierre Messali provided comments as peer reviewers. Comments were also received from Ms Nancy Chen. The team i s grateful for the cooperation and advice provided by concerned officials of the Ministry of Economy and Finances as well as those of other public institutions inCambodia. This CFAA complements and forms part of the IFAPER. It draws on the experience of previous studies such as the IMF Review entitled "Public Expenditure Management Framework," September 2001, and the results of various Asian Development Bank as Technical Assistance projects. It goes further in evaluating the effectiveness of accounting and auditing at the national and sub- national levels as well as the oversight functions. The accountability framework and the important aspects of financial accountability inthe country are thus covered. AnnexC 163 Abbreviations/Acronyms ADB Asian Development Bank ADD Accelerate District Development BMD Budget Management Department CDC Council for Development of Cambodia CIDA Canadian InternationalDevelopment Agency CFAA Country FinancialAccountability Assessment CMC Cash Management Committee COM Council of Ministers C/SF Commune/ Sankat Fund DBFA Department of Budget andFinancialAffairs DFI Department of FinancialInspection EEPPMU Economic and Financial Planning, Policy and Monitoring Unit FBC Finance and Banking Commission FC Financial Controller FCU Foreign Currency Unit FTB Foreign Trade Bank IMF International Monetary Funds IFAPER IntegratedFiduciary Assessment and Public Expenditure Review MEF Ministryof Economy andFinance MOEYS Ministryof Education, Youth and Sport M O H Ministry of Health MOP Ministryof Planning NAA National Audit Authority Nl3C National Bank of Cambodia NT National Treasury OPM Office of the Prime Minister PAP Priority Action Program PDAF ProvincialDepartments of Finance PDOEYS ProvincialDepartments of Education, Youth and Sport PER Public Expenditure Review PIP Public Investment Programs PT ProvincialTreasury RGC Royal Government of Cambodia TA Technical Assistance TOFE Table of State Financial Operations TSA Treasury Single Account UNTAC UnitedNations Transitional Authority for Cambodia WB World Bank 164 Cambodia: IntegratedFiducialy Assessment and Public Expenditure Review Table of Contents PREFACE............................................................................................................................................ 162 EXECUTIVE SUMMARY ................................................................................................................. 165 Legal framework.............................................................................................................................. 165 Context............................................................................................................................................. 165 167 Control Environment ........................................................................................................................... Conclusion ....................................................................................................................................... 169 LegalFramework............................................................................................................................. 169 169 BudgetFormulation............................................................................................................................. InternalControl Environment.......................................................................................................... 170 170 Analysis ........................................................................................................................................... Current System ................................................................................................................................ 172 Expenditure Control............................................................................................................................. 174 Recommendations............................................................................................................................ 173 Current situation .............................................................................................................................. 174 Analysis ........................................................................................................................................... 175 Treasury Operations and CashManagement ....................................................................................... 178 Recommendations............................................................................................................................ 177 178 Analysis ........................................................................................................................................... Current system................................................................................................................................. 180 Public Accounting and Reporting........................................................................................................ 183 Recommendations:........................................................................................................................... 182 Public Accounting............................................................................................................................ 184 Public Chart of Accounts ................................................................................................................. 184 Reporting ......................................................................................................................................... 185 Recommendations............................................................................................................................ Analysis ........................................................................................................................................... 185 186 Auditing............................................................................................................................................... 189 190 InternalAudit................................................................................................................................... External Audit.................................................................................................................................. 191 Analysis ........................................................................................................................................... 191 Recommendations............................................................................................................................ 191 Recommendations Summary ............................................................................................................... 192 Annex C 165 EXECUTIVESUMMARY Context 1. The main objective of the Country FinancialAccountability Assessment (CFAA) i s to analyze the quality and reliability of the public expenditure management mechanisms in Cambodia and to assess the level of fiduciary risk associated with public expenditure. The second objective of the CFAA i s to formulate recommendations designed to improve financial management in the public sector. The analyses and recommendations contained in the CFAA are critical to the Government's efforts already under way to improve public finance management. The CFAA will also make vital contributions to the future CAS and the PRSP. 2. The conclusion of the CFAA i s that the legislative and regulatory framework governingpublic finances in Cambodia i s generally acceptable, with the exception of the control function, which requires a complete overhaul. However, the implementation of public expenditure i s very weak, resulting in a lack of transparency and a high level of risk concerning the reliability of public expenditure management. In addition, the financial management systems and procedures cannot be relied upon to provide any assurance that the public funds are used for authorized purposes. The system i s all in cash based and most receipts and almost all payments do not pass through bank accounts. Such a system provides many opportunities for corruption and misuse of public funds. The action program proposed in the CFAA would in the long run, and with the support of other donors, improve substantially the reliability of public finance in Cambodia. Specific measures (in particular those relative to public accounting and control mechanism) should be implemented immediately and be fully operational by the end of 2003 (see action plan) in order to reduce the adverse impact of weaknesses identified and improve the transparency and efficiency of public finance management. 3. The Government with the assistance of IMF, has already started implementation of the recommended reform measures. The mission commends the government's efforts but note that progress i s very slow due to many constraints faced by the government, serious weaknesses still remain. The Government has embraced an extensive reform initiatives sponsored by its development partners. These include major reforms such as performance based budgeting, budget decentralization, medium term budget framework, standardization of accounting and others. The agenda i s totally beyond government capacity and misses fundamental weaknesses particularly in the payment and reporting systems. However, there i s an urgent need to prioritize government's efforts if the reforms are to be successfully completed. The findings and recommendations of the CFAA, inparticular the 15 actions of the attached priority action plan, should facilitate the government's selection of specific targets for improvement. The Bank has taken note that implementation of several priority actions, particularly those relating to accounting, was already in progress. The Bank would like to encourage the government to actively continue with these reforms, in particular the creation of an integrated accounting and budgetary information system. Legalframework 4. The Cambodian system of public finance management i s governed by a series of laws and ministerial decrees. This law i s based in part on the 1993 Constitution, and more specifically on the financial system law adopted by the National Assembly, promulgated by Kram No 01 N S of 166 Cambodia: Intearated Fiduciary Assessment and Public Exuenditure Review December 28, 1993. The Cambodia's budgetary and accounting law which i s largely derived from the French system, constitutes a coherent and solid framework for the Government's budgetary policy and public finance management. Budgetformulation 5. The public expenditure system in Cambodia i s governed by two basic rules, namely: separation of the responsibilities of the authorizing officer (Ordonnateur) and of the accountant (Comptable Public), and the rule of certification of services rendered. The budgetary procedure which i s heavily centralized i s the sole responsibility of the Budget Directorate (Directeur du Budget) who plays a pivotal role. The spending ministries are not yet sufficiently associated with the budget execution. Expenditure Control 6. There are at least four systems of budget execution in operation. The rules governing the recording of expenditures for budget management reporting are mix of cash and accrual accounting. Under IMFrecommended reforms, a treasury single account (TSA) was established with the National Bank. However, this account i s not used as intended. Use of cashwithdrawal orders to effect payments exposes these accounts for significant risksof misuse of funds. 7. MEFalso maintains severalbank accounts which aredollars denominated. These accounts are not under National Treasury control. There are no transparent rules for operation of these dollar denominated bank accounts. It i s not clear whether receipts and payments from these accounts are fully reflectedinthe budgetary accountsprepared by the NationalTreasury. 8. The practice of offsetting tax and customs dues against government obligations to suppliers continue despite IMF recommendation to discontinue. Review by the mission revealed that there are more than 100 accounts still in existence. Operation of these accounts is another possible source of misuse of funds. Treasury operationsand Cashmanagement 9. The management of the government's treasury needs to be improved: (i) National Treasury has no control over revenue deposit which continue to come through the bank account of collecting agencies; (ii)different divisions of the National Treasury still operate independently with separate data base to manage their operations; (iii) foreign currency accounts are held in the National Bank and operated by a separate department of MEF without any review by the National Treasury; and (iv) monitoring of the government's cash accounts take place exclusively on a daily basis, with no short or mediumterm vision. 10. Public The public accounting system suffers from numerous weaknesses, some of which are serious to affect the reliability of all government accounts. The problemidentified are generally linked to poor management and lack of accounting procedures. There i s no clear accounting procedures manual. The accounting system has evolved over the years and a combination of Soviet, French and Anglo-Saxon systems. Financial statements are not produced. The accounting system produces only a monthly trial balance. Numerous reports produced outside the accounting system are used for managerial reporting. Auditing 11. The control system inCambodia appears to lack the necessaryfocus on efficiency and it needs substantial improvements to provide satisfactory safeguards against inefficient and unauthorized use of public funds. The control system i s heavily accentuated on ex-ante controls, mostly at the level of Annex C 167 budget execution. Ex-post controls are uncoordinated, and include several units working independently. These include: (i) Inspection General des Finances (IGF) which operates mostly on demand, (ii) Auditor General does not have sufficient capacity and resources to operate effectively, (iii)National Assembly does not have the capacity and resources to support the control function it i s expected to assume, (iv) MEFFinancial Controller which operates mostly on demand, and (v) internal audit function within the sector ministries are very weak. Management and information systems 12. The management and information systems at the budgetary and accounting level suffer from serious flaws. The government has been attempting to remedy this situation for some time, but its efforts still need to be improved. Although an integrated financial management system should be the ultimate goal, the technical competency of staff and current development of accounting rules and regulations do not warrant embarking on a complex systemfor a time being. Capacity building 13. The human resource situation in the MEF and other sector ministries i s a major concern for the Government in its efforts to successfully implement the present budget reform. The staff inMEF and other sector ministries i s also insufficiently trained in the computerized budget and accounting systems, and staff inthe sector ministries need training inbasic budget management as well. Conclusion 14. In spite of the multiplicity of controls and the centralized nature of the PEM system, weak financial management practices pose serious fiduciary risks. The fundamental problems result from a weak control environment, severe weaknesses in National Treasury (NT) operations, especially in the area of cash management, inadequacies in the public accounting system, and weak internal and external auditing capacity. These problems are so serious that the Government's budget and reporting systems cannot be relied upon to expend resources in an accountable manner. These systemic weaknesses thus raise concerns about fiduciary malfeasanceand possible corruption. 15. The current budget reform undertaken by the government constitutes a major initiative to ensure the long-term efficiency, effectiveness and transparency of the management of public funds. The impact of the reform so far encouraging but much remains to be done before implementation of the reform i s fully effective, especially the accounting systems and capacity issues. 16. The measureproposed under the CFAA reinforce actions already under way to implement the reform as well as additional recommendations for the improvement of current actions, which should be included in the Bank's dialogue with the government. The short-term actions mentioned below, which should be completed within the next two years, are improvements on the existing government action plan recommended by the IMF.The actions include: (i) improved budget management capacity at the MEF, sector ministries and local government; (ii) develop accounting standard and reporting systems; (iii) establish an integrated computerized treasury payment system for accurate recording of all public revenues and expenditures; (iv) rationalization and integration of internal and external control processes. A development action plan together with an indicative timeframe i s set out in Appendix 1to this report. 17. The above measures would help ensure an efficient implementation of the public finance reform over a two year period would include the following action: (i) sharper Bank focus in its dialogue with the Government of Cambodia through a single public sector finance action program which would take into account the different instruments CAS, PRSC, HIPC; (ii) maintenance of tighter coordination by the MEF over all the donors' programs concerning public finance; (iii) dissemination of the CFAA and other donors' recommendations in country through a series of 168 Cambodia: Integrated Fiduciary Assessment and Public Exuenditure Review workshops that would include the borrower and donors to ensure buy-inby bothborrower and donors alike; (iv) assessment of the need for additional financial support to the borrower to facilitate implementation of the budget reform. 18. Specific areas of concern requiring additional work in FY 2004 include the following: (i) status of local government accounting and budget management; (ii)capacity issues and training requirements at all levels of the MEF, sector ministries and local government; (iii)dissemination of government accounts to the public and raising public awarenessto public finance issues; (iv) review of the dysfunctional system of audit inpublic finance. AnnexC 169 CONTROLENVIRONMENT LegalFramework 19. Cambodia i s a very poor country with GDP per capita US$260 and population 10.3 million. After 25 years of war and internal conflict, the Paris Peace Accords were signed in October of 1991. The Accords established the Supreme National Council (SNC), representing the four major Cambodian political factions as the transitional vehicle of Cambodia's sovereignty. They also established the United Nations Transitional Authority for Cambodia (UNTAC) to maintain peace and develop a neutral political environment in which free and fair elections could be held. The elections took place in May 1993 and a national government was established in September 1993 within a framework of parliamentary democracy, under a new constitution, with a constitutional monarch as head of state. Internationalassistanceresumed. 20. The Cambodian system of public finance management law is based in part on the articles 57, 90 and 91 of the 1993 Constitution. The public finance reform was initiated in early 1993, based on the studies that had been undertaken since July 1992, and led to the adoption by the National Assembly of the law on Financial System (Budgetary), promulgated by KramNo 01N S of December 28, 1993. A series of legal and administrative instruments have been developed and implemented for the reformof public finance. The law on the Financial System (Budgetary) was first applied with the 1994 National Budget. Data analysis for budget execution were initiated during the budget execution in 1994. By 1995 and subsequently, these reforms had to be expanded, given greater clarification, accompanied by the necessary implementation measures. The overall rule of law from the financial and economic standpoint hadto be completed with new legal and regulatory provisions. 21. Cambodia's budgetary and accounting law which i s largely derived from the French system, constitutes a coherent and solid framework for the Government's budgetary policy and public finance management. The law defines the roles and responsibilities of the National Assembly, the Ministerof Economy and Finance (MEF)and other public officials inthe administration of revenue collection and expenditure management. The public finance in Cambodia integrates the budgetary principles, rules and practices relating to: (i) Separation of the responsibilities of the authorizing officers (Ordonnateurs) and of the payment officers (accountantskashiers); (ii) Unity: according to this rule, all budget appropriations must be gathered in a single budget document. It i s prohibited to make any collection of revenue or any payment outside of the budget. But in fact, the budget law i s made of many different types of budget accounts; (iii) Universality: all revenues and all expenditures must be shown completely in one document. All receipts must not be contracted with expenditures; (iv) Non assignment of receipts: receipts must not be assigned to some expenditures. But in fact, many processesallow this affectation such as "fonds de concours" and "budget annex" or "comptes speciaux du tresor"; (v)Annuality: appropriations are available duringonly one fiscal year. But infact, unspent funds can be carried over under certain conditions and, above all, capital expenditures are voted in a multi-year perspective, authorization program (AP)and Credit de Paiement (CP); (vi) Specialization: appropriations must be specialized by nature, destination or service. But in fact, some chapters are non-specialized and more frequently chapters are globalized. 22. The National Budget becomes the Government's financial program and the basic instrument for realization of four key financial functions of the State, namely: (i) resource mobilization, (ii) resource allocation, (iii) distribution function, and (iv) macro-economic stabilization. Internal ControlEnvironment 23. A sound internal control environment requires the existence of laws, rules and regulations, a leadership that sets the ethical and professional standards, a well motivated staff who understands the internal control rules, open communications and reliable monitoring and performance measurement. 170 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review Notwithstanding the existence of acceptable laws, rules and regulations, the Cambodian internal control environment lacks many of the above requirements. A poorly paid and a poorly trained civil service i s not motivated to achieve high standards of performance in discharge of their duties. There are no minimumqualification requirements for technical staff. Inthe field of accounting and financial management, there are no recognized professional qualifications. Average civil service salaries are about $28 (2003) per month and are not sufficient to attract qualified and experienced staff.' 24. A highly centralized public expenditure management system does not provide incentives for line ministry staff. Line ministries often maintain off-budget revenues to finance urgent expenditures given uncertainty about timing of cash release. This practice, and others like it, invariably provide opportunities for illegal fees and financial leakages. Moreover, there i s no reliable monitoring and performance measurement system. Budgets are monitored in aggregate by MEF for compliance with overall budget ceilings. Individual ministries, MEF financial controllers, and National Treasury officers maintain separate databases to serve their individual control requirements. No attempt i s made to reconcile these reports or relate expenditure to performance of individual spending ministries. 25. Furthermore, in some cases the principles embodied in the law are simply not followed. For example, the principle of separation of authorizing officer and payment officer i s not applicable for some parts of the budget (viz., PAP). Certain revenue collections are not remitted to the budget, resulting inoff-budget revenues and expenditure (thisproblem has been cited inthe forestry case). The requirement that all payments above Riels 2.5 million be paid by check or bank transfer i s not followed at all. These practices have been adapted over time inorder to allow the systemto respond to present needs. However, they contribute to a weak control environment by allowing opportunities for misuse of public funds. BUDGETFORMULATION 26. For the most part, Cambodia's budget system continues to follow a traditional approach: resources are allocated to institutions through budget ceilings, the institutions prepare bids which comprise a detailed breakdown of the inputs they require, these are then negotiated with the Ministry of Economy and Finance's (MEF) Department of Budget and Financial Affairs (DBFA). The approved budget presents a detailed breakdown of inputs by chapter, distinguishing allocations for salaries and allowances, operating costs, economic and social interventions, and capital. Input allocations are further broken down into sub-chapters, which are used for the purposes of budgetary control, and line items, with a total of forty-three line items detailed in agency budgets for operating costs. Agencies have limited authority to alter the composition of inputs during execution - virement within chapters has to be approved by the Minister of Economy and Finance, and between chapters by the Office of the Prime Minister (OPM). Weaknesses in the budget execution and cash management processes mean that the strategic prioritizations that are made in budget preparation are often not delivered. CurrentSystem 27. Budget calendar and macroframework: There i s a sound well-established budget calendar in Cambodia. It begins inMay/June with the derivation of the macroeconomic aggregates of the budget. Revenue projections, which are mainly based on GDP growth projections, determine the resources available for the forthcoming year. The Royal Government of Cambodia does not, at present, borrow from the domestic banking sector to finance current budgetary expenditures, a provision that i s enshrined in the annual budget law. In the macroeconomic planning of the budget, the MEF aims to operate a small current surplus to finance capital expenditure (largely counterpart funds for donor- financed projects). The capital budget i s financed from concessional project support loans. 'Most civil servants admit that they have to do a secondjob to earn a living income. Annex C 171 28. Departmental ceilings: Once revenue availability has been determined, a process of prioritization allocates revenues between current and domestically financed capital budgets and between broad economic sectors. These ceilings are primarilyincremental but adjustments are made at the margin to take account of high-priority policies (e.g. increases in the sector allocations are approved by the Council of Ministers). 29. Budget circular: The approval of the Council of Ministers signals the beginning of the construction of line department ceilings, which are transmitted to line ministries with a detailed budget circular in late July. The circular details the general policy stance of the budget and the ceiling for domestically financed expenditure for each ministry. The envelopes for each ministry are derived with reference to the current year's budget and expenditure performance to June when the f i s t budget review i s conducted with adjustments for the policy priorities included in the sectoral planning process. The ceilings contain an implicit planning reserve; they are set lower than total projected resources inanticipation of new expenditure requirements being introducedinto the budget process. 30. Spending agency submissions: Ministries are required to provide detailed estimates for their expenditure by chapter, subchapter, and line item by the middle of August. These are prepared with reference to the needs of each ministry's various provincial and district offices (this i s in contrast to budget execution when the provincial departments of the ministries fall under the control of the provincial administration). The level of sophistication varies by line ministry.The ministryof Health, which has had substantial technical assistance, has developed a simple formula based on fixed and variable costs to determine the funding needs for operational costs for each of its provincial departments and major health institutions. Other ministries have less sophisticated methods of delivering their budget bids. 31. Negotiation process: Following the receipt of ministry bids (which generally exceed the predicted ceilings), a process of discussion and negotiation takes place in September/October. These discussions are often highly detailed, with the Department of Budget questioning individual line item bids. Decision are often referred to the level of Finance Minister or Prime Minister. Upward and downward adjustments are made to take account of competing priorities. A draft budget finally emerges and i s submitted for approval to the Council of Ministers and onward to the National Assembly for approval by the last week of December. 32. Dual budgeting: The capital budget i s almost entirely foreign financed; only the small domestically financed component (mostly counterpart funds) i s subject to normal budgetary discipline. During the 199Os, considerable effort was put into the establishment of a rolling, three-year Public Investment Program, managed by the Ministry of Planning (MOP). This sought to match top-down allocations for investment, based on a macro-economic framework for the budget agreed by the Council of Ministers, with programmed allocations for on-going projects and a prioritized listing of new investments presented by the line agencies. MOP i s supposed to appraise line ministries' proposals to determine whether they meet government's development priorities and whether costing is realistic. 33. Unfortunately, the PIP has had little impact on the expenditure management process or expenditure outcomes. Project proposals are subject to cursory review. The result i s a shopping list of projects. Allocations for capital expenditures, through the budget's Chapter 50, are supposedto reflect the domestic financing requirements identified in the PIP. In practice, no such allocations are made. Instead, the investment budget i s managed by the MEF's investment department, with funds allocated to projects approved by the Council of Ministers or the OPM directly. A listing of investment projects i s presented in the annual budget, though additional projects are approved in-year. Usually expenditures are posted against the nominally responsible agency, though this is not always the case. Allocations of counterpart funds are made in much the same way. As a result, the recurrent budget process i s largely divorced from that of the capital budget. 172 Cambodia: Intearated Fiduciarv Assessment and Public Exvenditure Review 34. The PIP has now lost credibility as a planning tool. Agencies recognize that the PIPplays no meaningful role in investment decision-making, submissions are tardy, with scant information on proposed projects and virtually none on actual disbursements, while preparation and Council of Ministers approval of the annual PIP tends to slip outside the budget calendar or-as in 2001-is dropped altogether. As proposed in the previous PER, with the introduction of the MTEF attention should turn to the integrationof the programming instruments. 35. Legislation: The 1993 Organic Law and subsequent further decrees govern the annual budget law. The annual budget law grants the executive the right to collect revenues up to amount specified in the budget and sets expenditure ceilings for the two budgets. Details of functional ceilings and broad economic category of spending (split by central operations and provincial operations) and project investment are annexed to the budget. The National Assembly does not vote detailed line item allocations. The annual budget law specifies the rights and responsibilities of MEFand line ministries in budget execution. It also sets out borrowing limits, the manner of management of commercial enterprises, and the accountability for budget execution. Analysis 36. It could be argued that the framework for budget preparation and management, though characterized by a high degree of central control, has many desirable features for a country that faces significant macroeconomic challenges ,in particular from a narrow and fragile revenue base, a poorly developed financial system, and weak governance structure. Budgetary expenditures are constrained to available revenues and borrowings tightly controlled to make use of concessional expenditure and avoid inflationary financing. 37. However, weaknesses in budget preparation and execution mean that the budget often performs poorly on technical and allocative efficiency grounds. The budget preparation process does not facilitate appraisal of spending plans in relation to medium term policy priorities; in many instances, it is overly fragmented and lacks transparency; it does not provide enough information for cash management purposes to facilitate the implementation of expenditure programs; and the budget execution process can reprioritize budgetary decision on an ad hoc and nontransparent basis. 38. Incremental single year expenditure planning frustrates the efficient and effective realization of the government's policy priorities and impedes the ability of line ministries to plan service delivery. Owing to the concentration of human resources on expenditure control, there i s a lack of capacity in the MEFto assess the efficiency or effectiveness of expenditure programs of line ministries. Most line ministries, with the notable exceptions of education and health, are not able to effectively manage resources and plan for the medium term, owing to the centralization and short term focus of budget decision making. The MEF i s conscious of a wide range of expenditure pressures over the coming years, and has initiated work on a medium term expenditure framework (see IFAPER main report - chapter 4). 39. The functional separation of the current and capital budgets frustrates coherent implementation of the government's policy agenda. The separate budgeting and execution systems distort the budgetary prioritization process. The allocation of counterpart funds from domestic resources i s effectively determined outside of the MEF, which further complicates cash management within the budget year. 40. Insome instances, revenues that are budgeted as fungible state revenue are infact earmarked for discretionary expenditure during the budget year. The budget that i s formulated, therefore, i s not a fully transparent statement of the spending intentions of the public sector. This practice contributes to the cashflow problems experienced by the main service delivery ministries inbudget execution. AnnexC 173 41. The budget classification structure does not allow meaningful analysis of the economic impact or policy focus of public spending. It reflects a narrow control focus, which only allows highly aggregated analysis of the economic impact of public sector activities. For expenditure reforms to be effectively pursued it need to be reformed to facilitate the planning control, and monitoring of expenditure inrelationto its policy goal. Recommendations 42. Improving expenditure planning and programming: In order to make this significant advance in budgeting in Cambodia a transition must be made from the single year, control focused expenditure planning currently conducted in the MEF to a more transparent forward looking process based on policy objectives. This is particularly important given the increasing number of expenditure pressures anticipated by the Ministry over the coming year, for example, from commune and national elections, increasing allocations to social sectors, ASEAN obligations, administrative reform, and military demobilization. 43. Given the limited technical and physical capacity in the Economic and Financial Planning, Policy and Monitoring Unit (EFPPMU) and the Department of Budget, inthe short term this exercise will be largely limited to increasing the amount and quality of information submitted to the MEF duringthe budget preparation process. This will enable an analytical basis to be slowly built up that facilitates expenditure prioritization according to mediumterm policy objectives. The parallel reform inbudget execution (moving from pre-audit to post-audit) should eventually free resources within the Department of Budget to undertake more expenditure programming and assessment functions. A key principle of the reforms i s to increase the planning and programming capacity in the line departments so that the MEFcan concentrate on overall budget policy rather than detailed control. 44. Decreasing fragmentation and improving transparency in the budget process: Ultimately, there should be a single consistent budget preparation exercise that clearly exposes and determines trade-offs between all public operations. All inflows, including donor-funded project expenditure should be included inthe exercise. At present, there i s serious fragmentationinbudget preparation and execution that need to be resolved. The most serious problem i s the disconnect between current and capital expenditure. In order to reassert the unity of the government budget, it i s crucial to reform institutional relationships inthe preparation on the investment budget and integrate it with mainstream expenditure planning. 45. The MOPand the CDC should remain responsible for the long term development plans, including contributing to the broad strategic development priorities. The MEF should have sole responsibility for decisions on the aggregate allocation and commitment in the annual budget which includes investment funds. Ultimately, one would expect that a fully unified system for preparation and execution of public expenditure would exist irrespective of the funding source. Inthe immediate term, however, integrating the execution system i s not a reasonable expectation inCambodia. 46. Another element of fragmentation i s the practice of pre-assigning some revenues to particular accounts in the NBC, which are not accessible for normal budget execution. At the time of budget preparation, all these revenues taken into consideration as available for budget financing, and the aggregate budgetary expenditure is determined accordingly. Some of these accounts may be used for unbudgeted discretionary expenditures outside of the budget control system. 47. Zmproving budget classification and establishingaprogram budgetframework: The current budget classification system i s a mix between highlevel aggregates and overly detailed descriptions of activities. The budget i s organized according to organizational structure, which broadly corresponds to functional classification. Within an organization, it is organized according to chapters, which provide economic classifications. It is, however, not organized in any kind of program format, which makes it difficult to determine the cost of a specific objective or program, and the final output. As such, it does 114 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review not allow meaningful economic or policy analysis of public sector operations. In addition, the expenditure planning and programming reforms envisaged above require line ministries to be able to plan and budget according to individual policy outputs (or programs) and for the MEF to be able to allocate, challenge, and monitor on the same basis. The ministry has expressed a desire to better understand the extend to which nonsocial sector ministries are actually implementing social programs. Thisrequire moving toward a systemof programoriented budgeting. 48. The priority in Cambodia i s to reform the classification system to allow coherent financial reporting. The new system should allow classification of expenditure by function and economic category as described inthe new Government Finance Statistic annual. A budget classification system usually consists of five dimensions to retrieve the information necessary for public expenditure management: (i) Source of Funds; (ii) Organization; (iii) Functional; (iv) Programs (activities and projects); and (v) Economic. EXPENDITURE CONTROL 3.1. The process of budget execution has been revised and reformed in a number of ways during the last decade. Partly as a result, the current system combines a small number of recent, post United Nations Transitional Authority for Cambodia (UNTAC), modifications with elements of the French system and centrally managed (Soviet systems). Ingeneral, initial attempts in the first half of the 1990s to devolve responsibility for budget implementation have been rescindedin an effort to regain tighter central control of finances. 49. As inmany developing and transition economies, the current and capital budgets inCambodia are executed through largely separate institutional procedures. In2001, financing for approximately 70 percent of capital expenditures i s expected to come from donor sources (grants and loans). Most of this expenditure is madeoffshore (inthe form of goods and services), henceonly a small proportion of these funds i s expected to be channeled through the Government budget implementation processes. The execution of current and capital expenditures, therefore, needs to be considered separately ineach case. They each have particular strengths, as well as some very considerable weaknesses. 50. At present, the institutional responsibilities for current budget implementation in Cambodia are divided between the MEF (in conjunction with the line ministries) for expenditure authority and the National Treasury (NT)for payments systems. Although, the NT is technically a department of the MEF, it is physically, a separate institution, operates with relative autonomy from the rest of the ministry,and has its own distinct institutionalfeatures andproblems. Current situation 51. A key feature of the process is the ex-ante (pre-audit) control of spending. Approval by the National Assembly of the annual budget law relates only to the global appropriation amount, major aggregates (revenues/expenditures, capitaVcurrent spending commitments and disbursements, and provinciallmunicipal budgets), and the associated borrowing and amortization arrangements. Aggregated tables are attached to the Budget Law specifying estimates for revenues (by chapter, article, paragraph, and subparagraph), for recurrent expenditures (by spending agency and chapter), and for capital expenditures (by progradproject). 52. The MEF i s primarily responsible for overseeing the implementation of the current budget. Two departments: the Budget Department and the Treasury Department (NT) have the key roles. Line ministries and provincial administrations are also directly involved but have limited powers and accountability. There i s widespread dissatisfaction with several features, and consequences, of the current procedures. Notification of spending agency appropriations and approval to begin AnnexC 175 implementation of the Budget Law i s issued by the MEF in January. However, this notification only permits spending agencies to begin to seek approval for individual expenditure commitments. There is no system or authority by which the MEF issues quarterly or monthly cash releases to spending agencies. 53. The budget implementationprocedures are simple and clear according to the law. The normal procedures require four phases: (i) commitment; (ii) liquidation; (iii) authorization; and (iv) payment. The commitment phase require submission of a proposal by the requesting ministry with a proforma invoice, verification of budget availability, signature of the requesting minister and pre-audit of the proposal by the financial controller of the MEF.This i s the most time consuming part of the process and has allowed considerable power to MEF to exercise control over the spending authority of the ministries. The spending agencies must seek MEF approval throughout the year for individual purchase requests. The principal initial documentation for this process comprises the "commitment visa", which is authorized at the appropriate delegated level in the MEF. The Financial Controller assigned to each spending agency, but located inthe BudgetDepartment (MEF) administers this pre- approval (pre-audit) process. Each commitment visa may require up to about ten signatures, depending on the requiredlevel of authority within the MEF.The time requiredto effect this process may extend from a few days to a few months. As of 2003, however, financial controllers are being out posted in MOEYSand MOHon a pilot basis. 54. The liquidation phase establishes the delivery of goods or services and the amount of the liability to the suppliers. Following above verifications, a payment order i s prepared and passedon for authorizationof payment. The MEFfinancial controller is then requiredto examine the payment order for completeness and authorize the payment. The authorized payment order i s then passed on to the treasury to effect payment. The Provincial Treasury (PT) follow a similar procedure inthe case of both provincial and central payments. Once this control procedure i s complete, spending agencies may proceed to implement the required budgetary operatiodactivity, though there i s still no guarantee that this can be completed without considerable further delays. A secondkey feature of the current process i s that a very high proportion of current expenditures, including salaries, are executed in cash. Thus, irrespective of the signed commitment visa, payments may only be executed if there are sufficient cash resourcesavailable at the required location. Analysis 55. The strengths of the current budget execution system are few, though not inconsequential. In particular, tight control of spending i s maintained in an environment characterized by severe budget constraints, major shortfalls in financial management skills, and frequent lapses in public administration ethics. On the whole, positions and responsibilities involving the execution of budget expenditure are tightly defined and the basic systems and practices for accounting are widely observed. The above process has led to many problems in budget implementation. There are no standards of performance that requires various actors of the process to complete their respective tasks. It is also not clear what criteria i s used for prioritization of commitment requests when budget constraints are faced. It i s alleged that, given chronic cash shortages faced, the system allows extractionof fees to process authorizations bothat commitment and payment stages. 56. As the normal procedures have led to delays in budget implementation, special procedures have been designed to overcome these difficulties. One such special procedure i s the Priority Action Program (PAP) wherein funds are released on a quarterly basis to fund operational expenditures inthe ministries of education and health. The PAPbudget for fiscal 2002 for education was not spent as of July 2002 due lack of agreement between MEF and the Ministry of Education on objects of expenditure. Although the PAP budget for health was released, there has been very little oversight exercised by the MEF or the Treasury on actual spending. The accounting and reporting arrangements 116 Cambodia: Intearated Fiduciarv Assessment and Public Exvenditure Review are not very clear. The Budget Strategy and Enforcement Center, a special unit created within MEFto manage the PAPbudget, has stoppedfunctioning. 57. Another special procedure is advance of petty cash to pay operational expenses under the Accelerated District Development scheme (ADD). Essentially, these procedures shift expenditure management from pre-audit to post audit. Although they are both acceptable models to avoid excessive ex-ante controls and shift accountability to spending ministries, the experience so far indicate that they have not fully met their objectives. 58. A recent addition to this multiplicity of budget execution procedures i s the Commune/Sankat Fund (C/SF). The C/SF is created to finance operating and investment costs of the recently elected commune councils. These number about 1,600 units throughout the country. The C/SF has entirely different expenditure management procedures compared with regular or PAP procedures. The provincial treasuries act as cashiers but authorization of expenditure i s solely with the commune chiefs. There i s no requirement of "pre-audit" and it i s not clear what other audit arrangements will be inplace. 59. The evolution of different procedures of budget execution has resulted in weak institutional and organization environment as staff are constantly faced with confusing and contradictingprocesses. The current budget execution and reporting systems cannot be relied upon for accurate measure of accountability and poses serious fiduciary risks.The alternative procedures devised clearly contradicts the Organic Budget Law. There are inadequate intemal control procedures to ensure that devolved funds are used for purposes intended. Reliable intemal or external audit systems are not inplace. The ability and skills of financial management staff at the ministerial level are not sufficient to manage such a shift inbudget responsibility and accountability. 60. Even though aggregate targets have been met, there i s considerable variance in budgetary performance against initial allocations at the agency level. Central agencies tend to overspend. For agencies' provincial administrations, the problem i s rather one of under-spending. In 2000, twelve of the twenty agencies with provincial administrations reported shortfalls against provincial administrations' budget allocations, amounting to 40 percent of initial allocations. Provincial execution performance improved markedly in 2001 and 2002, but provinces' execution rates, the proportion of provincial administrations under-spending, and the overall level of under-spending were still worse than for central agencies. Implementation constraints are one factor contributing to the provinces' poorer execution performance, but this may also be a consequence of the lower priority accorded to provincial governments in allocating available cash. Execution rates have tended to be lowest for provincial operating expenditures, since salaries and social interventions are generally given higher priority when funds are scarce. In contrast, execution rates for PAP have been significantly higher bothin comparison to fundingfor provincial agencies' operating costs and inrelation to special programactivities at the central level. 61. Additional factors contribute to the problem. Firstly, delays in the communication of budget allocations to agencies after the budget law i s passedby the National Assembly hinder the preparation and clearance of mandates in the early months of the year. Secondly, agencies often batch expenditure mandates for submission to MEF. Once submitted, MEF may delay the mandates' approval pending verification of availability of funds. Third, like agencies inmost countries, there is an end of year rush to submit mandates for any remaining budget balances, and corresponding rushon the part of the MEF to clear mandates so that they can be expensed. This in turn reflects the limited program implementation and budget execution capacity at the agency level. The rapid increase in agency budget allocations in recent years has aggravated these problems, as suggested by the steady increase end of back loaded expenditures seen in 1999, 2000 and 2001. A further contributing factor has been the introduction of PAP, which has required the release of substantial mandates directly to education, health and agriculture, thus aggravating the liquidity constraint. AnnexC 177 62. Lastly, the environment for budget execution i s heavily influenced by the legacy of central control. O&M expenditures are prescribed by MEF in fine detail (43 items) in Chapter 11 of the b ~ d g e tVirement (to transfer funds within budget chapters) requires the approval of the Minister of . ~ Economy and Finance. Transfers between chapters require the approval of the OPM (and transfers between ministries require parliamentary approval). This inhibits cost effective and flexible program management. 63. A key feature of budget execution (other than for the Priority Action Program) is pre-audit of all but minor spending at national and provincial level by MEF financial controllers. Multiple signatures arerequiredfor anything but trivial spending approvals while procurement committees slow disbursement markedly. Expenditure approval processes in the provinces are equally complex. A single disbursement can require signatures from the spending line ministry, the PDAF, the provincial governor, and the Provincial Treasury (7'). The irony is that these centralized, bureaucratic control procedures do not seem to provide for fiduciary accountability, though they do undermine effective and efficient expenditure management. 64. The weaknesses of the current system, however, are substantial. These can be summarized as follows: 0 Budgetexecution is often extremely slow, inparticularthrough the pre-audit process. 0 Multiple steps/ signatures appearto strengthen "control", but diminishaccountability. 0 Many pre-audit procedures (signatures) for expenditure are perfunctory and add no value to the final decision. 0 The process is labor intensive and encourages spending agencies to adopt strategies for managing/facilitatingof flow of "their" papers through the MEFaheadof others. 0 Budget expenditures are very uneven throughout the year and tend to surge in November and December inan attempt to implement delayed programs. 0 Lack of cashfrustrates implementationof the approved budget. 0 A very long"complementary period" to February 28 annually is requiredto finalize (liquidate) expenditures committed duringthe final two months of the fiscal year. Recommendations 65. The IMF Report discussed several budget management models and recommended that the budget executing responsibilities be realigned between MEF and the spending ministries by transferring the Financial Controller (FC), currently resident with MEF, to the spending ministry. The FC will report to the Secretary of State of the ministry but will also have a reporting responsibility to Head of Budget Management at MEF.The mission fully endorse this recommendation. The principal features of the recommended model are given below, though this model should be considered an ideal to be attained over the longterm. 66. There will be a single budget execution procedure (replacing current normal, PAP, ADD and C/SF procedures). After the budget i s approved, the MEF informs the line ministriedagencies about the approved annual appropriation. The line ministries and MEF agree on a monthly and a quarterly cash plan for cash management. MEF issues authorizationto draw money from the Treasury and start the spending process. This could be by an issue of a warrant or notice of cash allocation for a specific period (month or quarter). 67. The spending ministries prepare expenditure commitment, liquidation and authorization following the normal procedures currently in place. The FC, after verifying that spending request i s MEFusually fixes ceilings at the sub-chapterlevel, with agenciesallocating within these ceilings. 178 Cambodia: Integrated Fiduciary Assessment and Public Exvenditure Review within the budget and cash allocation, authorizes commitment for spending and the final payment. The Treasury makes payment and adjusts the cashand budget balancesof the spending ministry. 68. Under the above model, MEF will exercise control over budget spending by approving the original appropriation and by issuing cash allocations. Within these authorizations, considerable discretionary powers would be devolved to spending ministries to prioritize expenditures within available cash allocations. A system of issuing cash allocations would also provide certainty to spending ministries and improve budget execution. TREASURY OPERATIONSAND CASHMANAGEMENT 69. The institutionaldesign of the current treasury system in Cambodia has much to recommend a strong NT supported by a network of Provincial Treasury branches; comprehensive coverage of all public transactions at levels of the government; and consolidation of all custodial functions, including accounting, of public money in the NT. It also functions as the cash manager of the government and provides liquidity to all the line agencies and the provincial governments. 70. However, shortage of cash to finance the budget during the fiscal year has been a recurrent theme at nearly all levels inthe government, from line agencies to the provincial authorities. The MEF has tried to address this issue inmany different ways: cashrationing; establishing different systemsfor implementing the budget to ensure that priority expenditures are met; preparing annual and monthly cash plans; and regular across-the-board ad hoc cuts in cash allocation. All these initiatives have succeededin improving cash management only partially (if at all). Such ad hoc solutions have created multiple systems for executing the budget. The main problem is that cash shortage is symptomatic of many system-wide basic flaws in expenditure management. Any improvement in cash management will depend upon the reform of the other critical systems: establishing a rule-based and transparent system for cash allocation; makingpayments; accounting system; and treasury reforms. Current system 71. The actual functioning of the NT, however, has many features of a centralized command economy. Its main focus has become control of liquidity and cash in the system, and, as a result, its sole task seems to be to collect cash and provide liquidity to its clients the line agencies and the provinces. The NT, in addition to being the principal cashier for the government, carries out banking functions for some private sector suppliers to the government. One of the notable features of the current financial system in Cambodia i s the near absence of viable commercial banks, except for FI'B (a subsidiary of the NBC), and two or three other banks (mainly in Phnom Penh all of which do not undertake retail banking operations). The NT, for example, maintains deposit accounts on behalf of private sectors supplier of goods and services to the government. After the goods and services have been supplied in the fulfillment of a contract, and the NT receives payment order, it transfers the amount of payment to the deposit account of the supplier maintained in the NT. The supplier can use this account for payment of his dues to the government such as taxes and other revenues and payment of government obligations. These transactions are done by cash deposit and payment, or by carrying out off-setting accounting entries to reflect collection of revenues or payment of expenditures. However, the supply of cash to the supplier, for larger amounts of payment, requires another step in the process specific authorizationfrom the MEF. 72. The National Treasury (NT) and its network of provincial treasuries are the maincustodian of the government's financial assets and the centers of public accounting. Organizationally, the NT i s part of MEF. However, inpractice it operates very independently.The foundation of NT operations is the budget law, which requires that all public monies collected be deposited with the NT and all payments on behalf of government be made by the NT. Thus, all payment orders, after having been approved by the requesting ministry and pre-audited by MEF, are forwarded to the NT, which reviews AnnexC 179 the documentation for completeness. Payments are effected in two ways: by cash or by crediting the account of suppliers that maintain deposit accounts with the NT. The paid payment orders are then passedon to the NT's accounts division for accounting purposes. The NT maintains the Government's books of accounts and i s the only place where accounting records are kept based on double entry accounting principles. There are no accounting records maintainedeither by the spending ministries or financial control department of MEF.4 73. The main power of the NT comes from its power to provide or to deny cash to an agency. All the treasury transactions collection and payment of money are in cash. Providing liquidity to the line agencies and the provincial treasury involves moving a considerable amount of cash daily from one location to another. Eachday the NT transports currency to different regions. Central ministries collect cash from the treasury office inPhnomPenh. The provincial treasury remit surplus cash (after meeting their requirements) by currency transfer to the NT. Owing, however, to the lack of adequate communication facilities inthe information system, and (most important) uncertainty surrounding cash allocation, the NT often does not have a full picture of cash available inthe provincial treasury or with the line agencies. 74. The cash management system as currently designed has certain strengths. In theory, cash allocation should be made according to acashplanthat the MEFprepares at the beginning of the fiscal year. The MEFprepares annual as well as monthly cashplans covering both the central and provincial levels. However, in practice this system does not work. Many factors affect proper management of liquidity: the daily movement of cashfrom one location to another; the lack of modem communication facilities; the absence of a banking system; and uneven cash inflows leading to seasonal cash shortages. These factors have, in turn, introduced distortions: retention of cash by the line agencies; lack of information by the NT on the exact amount and timing of inflows; netting of liabilities by the private sector and autonomous agencies before remitting any money to the NT; and a general lack of transparency in all financial transactions. 75. The Government's cash resources are not managed from a single fund as provided for under the law, but are fragmented so that Treasury only has access to a small part of Government's total deposits. At end June 2002 about 10percent of the RGC's financial assets were available to Treasury for the purposes of budget execution (liquid assets then amounted to about 6.6 percent of the total Treasury executed budget, riel liquid assets just 2.8 percent). B y the end of December, liquid assets had dropped to 5 percent of total assets. 76. The management of cash resources of the government is fragmented. The government has some accounts in the National Bank of Cambodia (NBC) and some further accounts in the Foreign Trade Bank (FTB). These accounts cover both current and capital budget revenues and expenditures. The NT manages some of the accounts inthe NBC, while some others (foreign exchange accounts) are outside its control. The special accounts for some of the donor funds are maintained in the NBC, though these also remain outside the NT control. The NT, however, consolidates all the financial accounts of the government, including the operations of the government bank accounts outside its control inthe NBC. Inthe absence of any independent oversight system, the accuracy of the financial information on these bank accounts cannot be verified. Inthe case of accounts outside the control of the Treasury, it was not clear if these accounts were being reconciled in a timely manner. This i s a huge source of nontransparent management. 77. The NBCprovides the NT with the daily balance inthe mainTSA. The NBC does not seemto play any other role in liquidity management. Some other government deposits seemto be operated on instructionfrom other authorities inthe government. The cash inflow inthe NT i s received from many sources money collected directly by the tax departments and remittedto the NT; cash deposited inthe Financial management arrangements for the communes are addressedinChapter 6. 180 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review NT or the provincial treasury; cash deposited directly in the NBC; cash deposit by the private sector and public sector autonomous agencies in the NT on account of dues to the government; and foreign currency deposits by the external donors for project funds. Inaddition, the government holds many U S dollars denominated accounts inthe NBC from the revenues collectedinUS dollars. Analysis 78. The NT organization and functions, in theory, provide a strong institutional framework for efficient management of cashin Cambodia. However, inpractice this has not taken place. There are no policies and procedures that require revenue collection agencies to transfer funds to NT on a regular basis. Line ministries that collect revenues retain them as long as they could. Tax and customs revenues are accounted inthe NT only after collecting departments deposit them with NT..There i s no requirement for large tax payers to pay taxes directly to the TSA. Forestry revenue collection is another area where misuse of funds were reported recently. Lack of a revenue forecasting system and lack of automated information systems incollecting agencies and NT have weakened the ability of NT to effectively managecashinflows. 79. Some of the basic system that should underpin any financial management system are completely absent from the NT. There are no standards to govern the treasury operations or public sector accounting, except some decrees and ad hoc circulars issued from time to time; no audit and no requirement to audit the accounts; and no automation with voluminous accounting transactions is being carried out by using hand-held calculators or manually. As in the case of most procedures governing financial operations in Cambodia, the focus of the NT remains on compliance and control through multiple regulations and procedures. The problem i s seriously compounded by the absence of regular audit either internal or external. Some oversight function is, however, performedby more than one unit in the government through physical inspections of the line agencies. According to anecdotal evidence, this has also resulted in rent-seeking behavior at different levels of financial control, thus encouraging the agencies to flout established rules. Excessive controls seem only to have resulted in strengthening these systematic aberrations. There is also a serious capacity problem of lack of basic skills. This i s reflected in constant revision to the accounting data during the year. 80. The third component of the explanation i s the lack of an automated data recording system in both the NT and its provincial branches. Although provincial cash balances are reported daily, the accuracy of this data i s questionable. Book balances and cash flows are rarely reconciled with stocks of cash. Thus, the lack of a reliable information system i s a major obstacle to efficient cash management. The situation in the provincial treasuries are very much the same with all of the transactions taking place in cash. Provincial treasuries are responsible for payment of salaries for government workers, particularly teachers and healthworkers. There was one reportedcase of fraud in the Kampong Speuprovince. This demonstrate the risk of misappropriationdue to cashbasedpayment practices. Cash shortage thus results from a vicious cycle: agencies are not sure of adequate cash allocation and retaincash; and becauseagencies retaincash, the NT does not have enough cash. 81. The above deficiencies have resulted in a system which i s not only inefficient but also lends itself to rent seeking. Lack of transparent procedures to allocate available cash and predominantly cash based collection and payment systems allow opportunities for corruption and misuse of funds. IMF has recommended establishment of a Treasury Single Account (TSA) with the NBC to conduct NT financial transactions. A TSA was established and used for deposit of government revenues in Riel, TSA is not used for payment of suppliers and no payment checks are issued. Instead NT withdraws cash from TSA and brings into NT for payment of creditors incash. 82. Further,fragmentation of funds occurs becausethe Government's foreign exchange reserves are not under the control of the NT. The MEF's Foreign Currency Unit (FCU) i s responsible for managing foreign currency revenues, although the law requires that foreign currency earnings be AnnexC 181 deposited with NBC and converted to Riel, which does not happen. Foreign currency accounts typically have enough resources to meet current obligations, but remain unused to fund the budget in times of Riel shortages. Thus, there are no transparent rules for management of the foreign currency accounts. At the time of the visit of the mission at the end of July, NT did not have sufficient funds in Riel for payment of July salaries. However, the balances inthe foreign currency accounts maintained by MEF'sForeignCurrency Unitfar exceededthe months requirements of salary payments. 83. Treasury's weak control over revenue collection provides only part of the explanation for the severe cash management problem. Another part of the problem i s a result of poor coordination between Treasury and MEF's Budget Management Department in the allocation of cash to fund the budget. Although RGC established a Cash Management Committee (CMC) chaired by the MEF's BMD,CMC does not seemto havehadmuchinfluenceoncashmanagementby the NT as yet. One of the more serious flaws in the current system i s the lack of institutional coordination in the MEF. Organizationally, the NT i s a part of the MEF.Inreality, however, it does not generally take directions from the MEF. There i s a near absence of formal information flows for managerial purposes between the NT and other departments in the MEF (especially the budget department). The NT does, however, provide some consolidated accounting information for statistical and external reporting purposes to the MEF. Decisions to allocate cash are made by the NT without apparent reference to either budget priorities (see section on PAP below) or payroll obligations. Indeed, the process by which the NT allocates cash to budget heads i s unclear (some payment orders are paid immediately, while others are delayed until cash becomes available). The government faces severe cash shortages due weak controls inrevenue collection. It is unclear how the NT allocates available cashresources to fund government obligations. When payment orders are received at the NT, they are paid if cash is available or left unpaid untilcash becomes available. NT has an accounting of unpaid bills only of those suppliers that maintaindeposit accounts with NT.The other unpaidbills are not recorded. 84. There are other avoidable distortions in the system that affects cash management. The budget i s approved before the end of the fiscal year (calendar year). But the regulations governing authorization to spend are issued after the beginning of the fiscal year. The proper financing of the budget starts only around the beginning of February or later. There i s a complementary period, which sometimes extends unofficially into the second quarter, when the accounts of the previous year are kept open by the MEFto clear payment for goods and services receivedbut not paid for by the end of the previous year. The line agencies use this complementary period to spend money from the unspent balances in their budget allocations by accounting for these expenditures in the previous year. This distorts the picture of actual spending both inthe previous and the following year. Owing to delays in adjusting the previous year's accounts, expenditures incurred in the previous year are also sometimes adjusted against the next year's allocation, affecting the agency budgets for the next year. 85. The main features driving the current situation of cash shortage are: (i) distrust among The various agencies (MEF, line agencies, tax departments); (ii) A feeling that the budget of the line agencies will not be fully financed in a timely manner; and (iii) A distribution problem due to cash retentioninrevenue surplus provinces. The distrustof the system i s based on nontransparent allocation of cash. The line agencies know that the annual or the monthly cash plans prepared by them have little impact in the actual allocation of cash. The mission also heard anecdotal evidence from the line agencies of the incentives inthe system to ensure adequate cash allocation. Cash shortage thus results from a vicious cycle: agencies are not sure of adequate cash allocation and retain cash; and because agencies retain cash, the NT does not have enough cash. 86. The MEF has also accepted that the approved budget will not be fully financed and this is reflected in its decision to establish a parallel system for financing the priority expenditure (PAP) in four ministries: health, education, agriculture and rural development. There are other systems, such as the ADD, to ensure adequate financing for some other priority programs. The result i s that the line agencies not covered by these special financing systems are convinced that the flow of even partial 182 Cambodia: Integrated Fiduciary Assessment and Public ExDenditure Review financing of their budgets will be uneven and uncertain. This has resulted in a great incentive for line agencies, including the provincial treasury to retain government revenues that should have been deposited in the NT. Under the current rules, revenue i s to be remitted to the NT by the line agencies as soon as it i s collected. Owing to the lack of banking system and logistical problems of cash distribution, the NT allows line agencies to retain cash for their projected liquidity needs. Inpractice, however, the line agenciestend to retain a major part of the revenue collected by themunless forced to deposit it inthe NT. 87. With all the above weaknesses, the government has managedto avoid serious budget deficits or excessive domestic borrowing. This i s a mystery that baffles many analysts. Unpaid accounts recorded in the accounting system (equivalent of domestic borrowing) at the end of July 2002 amounted to 168,426 million Riels (US$ 43.2 million). This amount may be understated as these reflect debts to those suppliers who maintain deposit accounts. But the understatement may not be that material. The explanation of the phenomena i s more likely to be that many cash transactions both revenue and expenses are off budget, that i s line agencies retaining revenue to meet expenditures. In addition, NT releases payments only as revenue flows thereby avoiding serious deficit spending. Although, expenditure authorizations are made without reference to cash availability, it i s likely that suppliers do not honor purchase orders when payment backlog buildup. This probably i s the cause of back-loading or under spending of the budget. Recommendations: 88. The Government adopted two measures recently on the recommendation of the IMF in an attempt to improve control of revenue collections and expenditures. The RGC established a Treasury Single Account (TSA) in the NBC to conduct NT financial transactions. Some revenues in Riel are being deposited in the account and progress i s being made. However, the TSA i s not used to pay suppliers by either check or transfer. Rather, the NT withdraws cash from the TSA and injects it into the Treasury to pay creditors incash. 89. The requirement that all government revenues shouldbe deposited inthe NT and all payments are centralized in NT should provide the basic foundation for efficient cash management. Under the budget execution model proposed, the periodic cash allocation function could be assigned to the Cash Management Committee (CMC) who will regularly review forecast of cash inflows and issue cash allocations to the spending ministries. Clear rules could be established for allocating limited cash to spending ministries. The transfer of current cash allocation authority from the NT to CMC would increase the transparency of this function as well as provide more certainty to the spending ministries. 90. The IMF Report makes number of recommendations to reform treasury operations and improve cash management. Some of these, such as the Cash Management Committee, opening of Treasury Single Account and stock taking of domestic debt (payment in arrear) have been completed. However, as noted above these measures needs further strengthening to be fully effective. The mission recommends that the following actions be under taken to move forward with the reformprogram: Issue a directive requiring all payments above an agreed amount (the law states Riels 2.5 million) be made only by check or bank transfer. 0 Issue a directive requiring all large tax payers to pay taxes directly to TSA. Issue guidelines for revenue collection agencies on deposit of revenues to TSA within specified time limit of collection. Establisha simple computer system to record all financial transactions inthe NT and provincial treasuries. Integrate management of foreign currency accounts to NT inaccordancewith the law. Improve coordination between NationalTreasury and Budget department of MEFby empowering CashManagement Committee. AnnexC 183 0 Require all spending unitsto submit quarterly cashforecast to CashManagement Committee and update them monthly. 0 Require all cashtransaction to be channeled through Treasury Single Account. 0 Improve management information by automation National Treasury and Provincial Treasury transaction. 91. Inmakingrecommendations for reforming the treasury system, the missionwas guidedby the main objective of ensuring that the NT i s the only manager of the government accounts so that it can manage all financial and non-financial assets and liabilities of the government. On of the most important structural reforms should thus be to move the NT away from its primary role as the banker to the government, and gradually devolve this role to the commercial banking system as the commercial banking sector improves with better regulatory framework including banking supervision. The NT should: (i) manage, and account for all inflows and outflows in the government's Control, accounts in order to ensure government resources are managed in a unified fashion to gain the maximum return; (ii)Manage all government financial and non-financial assets and liabilities (including domestic and external debt); (iii) Prepare the final accounts of government; present these accounts to the National Audit Authority; and present the audited accounts to the National Assembly for approval. PUBLICACCOUNTING AND REPORTING 92. The financial accounting system has evolved over the years and i s primarily based on French accounting principles. However, over time, other accounting conventions have been adopted either due ideological reasons (soviet conventions during Khmer Rouge regime) or to respond to changing fiscal management requirements. The system, as currently practice, does not meet with any acceptable accounting standards and does not provide accurate and timely information required for financial management of a modem state. Over time the system has deteriorated owing to the successive crises and incompatible systems imposed on this basic model. The lack of attention given to the system over the year has led to an inability to respond to the increasing and changing demands of budget management in particular, because basic constitutional systems such as a standardized accounting methodology do not exist. Also, the system i s manual, and there i s an extremely low level of automation in the National and provincial treasury. The result i s the system that provides late, incompleteand inaccurate information. 93. The financial management was ranked as the third most important area in need of improvement by a questionnaire survey for implementing agencies (behind project management and planning and programming) at the Country Portfolio Performance Review (CPPR) workshop which was held in August 2000. The financial management capacity of the Project implementing Agencies suffer in three dimensions. First, the national procedures for budget preparation are weak and the overall financial situation makes the availability of counterpart funds a constant challenge. Second, Bank financial management procedures are poorly understood and not easy to deal with. Third, as project are decentralized into the regions and local areas, the issues mentioned above are simply compounded, unless there are simple and clear systems that can be taught quickly and coherently, as through the institute of Economics and Finance. The financial management systems across all agencies i s very weak. There are lack of training, coherent and consistent policies, procedures and guidelines in all aspects of financial management and internal control systems. Of particular concern is the lack of common accounting procedure for World Bank projects inCambodia. Closely linked to this is the lack of training and training manuals and lack of assistance when important financial management issues are raised. The focus group concluded that improving financial management skills was the most important area of concern to national and local implementing agencies. I84 Cambodia: Intearated Fiduciarv Assessment and Public Emenditure Review PublicAccounting 94. The Cambodian financial accounting i s centralized in the NT and its provincial branches. There are no accounting records maintained at the line ministries or at the other branchesof MEF.The Chart of Accounts used i s highly aggregated, basically showing control accounts with very little or no subsidiary ledgers for analysis. A trial balance is produced monthly but no financial statements are produced. There is no accounting manual detailing policies, procedures and rules except for several decrees or sub-decreesthat govern expenditure management. 95. The financial management of the government i s based on regulations that establish principles for public finance management, most importantly the separation of ordonnateurs and public accountants in charge of payment. They define powers, obligations, and accountability for the public officials inbudget execution and divided the process into two distinct stages: 0 The administrative phase involving ordonnateurs inline ministries and the budget department inMEF. 0 The accounting and payment phases involvingthe treasury and the ForeignCurrency Management Unitfrom the MEF. 96. The focus of the accounting and reporting system i s firmly on the final encashment stage of payment. All government transactions are meant to flow through this system. However, the reality of the situationi s that line ministries retain many revenues and spent outside of this formal system. Public Chartof Accounts 97. The current Public Chart of Accounts was initially established by a 1994 Prakah, the initial objective was to precisely identify services, location, and nature of budget operations through a specific system of budget codification. This codification has, however, been undermined and abandoned under subsequent finance laws. Currently, the Public Chart of Accounts consists of 63 accounts updated on March 19, 2001, for the state and four accounts updated on August 31, 2000 for the province. Some suppressed accounts, continue to be used, particularly in the provinces. There are two main types of budgetary accounts, which do not contain detail on individual transactions: 0 Account NO900 (revenue of state at central and provincial level) and 902 (province) to record total budgetary revenue; and 0 Accounts NO 910 (expenditure of state at central and provincial level) and 912 (province) to record total budgetary expense. 98. Most operational accounts are either transfer accounts or entity accounts. Accounting procedures for recording transactions are not defined by any standard guidelines or methodology and are therefore open to different interpretations indifferent provinces. 99. Accounting information i s compiled by the Treasury from manually recorded individual transactions into intermediate aggregations, which are consolidated on excel worksheets into monthly central and provincial balances. These activities result in the final account balance (balance generale des comptes du tresor) for the year and form the basis of the L o i de Rbglement (Finance Law on final budget execution), which i s approved by the National Assembly, usually four or five months after the end of the year. Before spending, the availability of budget allotments i s also controlled through a manual register. Payment orders and other transactions are aggregated in hand-written daily trial balances. These daily balances are then summarized in the ledger. The semi-automated ledger aggregatesthe daily operations on a monthly basis with a separatereference to the account number. AnnexC 185 Reporting 100. The main financial statement (called TOFE- Table of State Financial Operations) for expenditure management i s produced by the Budget Department of MEF. The TOFE is produced based on information supplied by NT. However, there i s no reconciliation between the TOFE and the ledger balances maintained by NT on revenue or expenditure accounts. The TOFE i s the only report meant mainly for external users (such as W)and i s of little internal practical use. This i s compiled in the MEF from the information provided by Treasury monthly balances. It is highly aggregated and shows above-the-line information at the chapter level; it does not disaggregate by economic classification or functional unit. Analysis 101. The basic objective of public finance accounting i s to provide comprehensive, timely and accurate financial information on the financial position of the government by recording budget execution and revenue collection transactions in accordance with the established accounting system and procedures. The accounting system should: Be comprehensive and capture all transaction of the public sector, recording each transaction, inflow and outflow, assets, liabilities, incomes, and expenditure to comply with legal and other requirements using a standard and internationally accepted accounting methodology. 0 Provide information on a daily and as neededbasis and by transaction, and by any other analytical format. 0 Enablemanagerial analysis to ensure that funds are usedefficiently and effectively. 0 Produce timely, accurate, and comprehensive financial accounting statements for information and effective control of finance. 0 Be an instrument for management of cashby the treasury; and Maintain the bookkeeping for bank account of the government at the NBC. 102. Judged against these principles, the current system in Cambodia falls short in a number of dimensions. The combination of computerized and hand-written systems, an obsolete methodology, and the lack of precision from an outdated Chart of Accounts result in unreliable and partial information make it difficult to understand and verify the validity of fiscal information. The main shortcomings are: 103. Incomplete information: The recording system generatesmonthly balances five days after the close of every month. However, this information i s limited to the operations carried out by the Treasury, and it does not contain detail on transactions outside of its control, for example, transactions carried out with cash retained by line ministries and through the extra budgetary funds (e.g. the road maintenance fund). 104. Unreliable information: The mission reviewedaccounting records maintainedboth at NT and provincial treasuries. The mission concludes that the Chart of Account i s inadequate to reflect all government financial transactions (assets, liabilities, bank and cash accounts, revenues, expenses, end- of-year results), the information i s unreliable due lack of reconciliation and incomplete recordingof all government transactions (e.g. foreign bank accounts maintained at MEF), and the accounts are not subject to independent audit. Given the above, it's the conclusion of the mission that the accounting system cannot be relied upon to provide any assurance that public funds are used for purposes intended. 105. Lack of Detail: The accounting system does not provide easily accessible information on individual transactions. Most accounting data i s confined to aggregate information in various tables, 186 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review manual trial balances, and balance sheets not directly from accounting registers of individual transactions. The aggregates refer to different accounting documents and complementary tables are needed for consolidation. This restricts the ability of the system to provide adequately detailed management information. 106. Lack of Reconciliation: There is a lack of basic reconciliationsystems to verify to accuracy of the information held within the accounting system. Whilst information i s held at an aggregate level on revenue collection there i s no reconciliation between the Treasuries and Revenue Collecting Departments on the detailed composition of the flows to the Treasury Line department keep records of their expenditure, but these are not reconciledat any point with the accounting records inthe Treasury Line ministry records of expenditure are not reconciled with MEFlTreasury information to develop agreedrecords of expenditure reports. There i s no systemic reconciling of accounting information with the banking sector. Inthe case of advance payments made by the treasury for programs under the PAP, the advance i s adjusted on the basis of expenditure accounts rendered by the Budget Management Centers. It was, however, not very clear to the mission if the Treasury played any significant role in the scrutiny of these accounts beyondcompiling these. 107. Financial Reporting does not facilitate public sector management: The result of the weaknesses in the accounting system is that expenditure reporting (TOFE) does not provide sufficiently accurate or detailed information for effective implementation. Information i s provided by chapter without detail on the functional source of the activity, this i s to a certain extent a function of the mismatch between the existing budget classification and Chart of Accounts structure. The reporting system, does not allow the MEFto identify on a timely basis the volume or source of current and future expenditure pressures. These functions are important to enable effective cash management. The lags and inaccuracies in the system mean that reports are often revised substantially throughout the year. 108. Unfortunately, the problems are myriad. First, the system i s centralized in the NT and its provincial branches, and there are no accounting records maintained either by the line ministries or other MEFdepartments. Second, the chart of accounts i s highly aggregated, basically showing control accounts with very little or no subsidiary ledgers for analysis. Third, though a trial balance i s produced monthly, no financial statements are produced. Fourth, there i s no accounting manual that details policies, procedures, or rules with the exception of several decrees and sub-decrees that govern expenditure management. Fifth, the TOFE (Table of State Financial Operations), which i s the main financial statement for expenditure management, i s produced by the EFPPMUbased on data supplied by the NT, though there i s no reconciliationbetween the TOFE and the ledger balances maintained by the NT on revenue or expenditure accounts. 109. Given the severity of the problems the inadequacy of the chart of accounts for reflecting all government financial transactions (assets, liabilities, bank and cash accounts, revenues, expenses, end- of-year results), the unreliability of basic information due to the lack of reconciliation and incomplete recording of all government transactions (e.g., foreign bank accounts maintainedat MEF),and the fact that accounts are not subject to independent audit-it must be concluded that the accounting system cannot be relied upon to provide assurance that public funds are used for the purposes intended. Though a Treasury Accounting Working Group has been formed to address these problems, progress has been slow. Addressing the lack of uniform accountingpolicies andprocedures as well as the lack of adequate systemsfor timely and accurate data recording and reporting should be considered the two most urgent initiatives. Recommendations 110. As most government transactions relate to revenue collections and payment of recurrent costs, devising accounting policies and procedures should not involve adopting all the currently available Annex C 187 international public accounting standards.The accounting working group should be requiredto review the relevant standards and come up with simple and clear policies and rules without getting too involved with complex issues that may have only distant future applicability. 111. Although an integrated financial management system should be the ultimate goal, the technical competency of staff and current development of accounting rules and regulations do not warrant embarking on a complex system for the time being. A computer system should be installed to computerize accounting transactions. At the NT, where bulk of the transactions take place, an integrated software package could be installed to undertake all transactions. For provincial treasury operations, a stand alone PC with a single user capability i s more than adequate. Data transfer from provincial treasuries to NT system could be through disks (as currently practiced in some of the Bank's projects with provincial operations). 112. The basic principles underlying the recommendations for the accounting and financial reporting system are to establish a system that closely links budgeting, spending, and reporting within a standardized single accounting framework. At the same time, it should be rationalized and computerizing by the recording of each transaction in a ledger (general and subsidiary) based system, item by item, in order to establish trial balance and balance sheets in accordance with established rules. In the medium term, these reforms will also improve the effectiveness of the National Audit Authority. A concurrent step should be to improve the quantity, quality and timeliness of information by designinga system of financial reporting. Certain immediatesteps needto be taken. These are: 0 Accounting system shouldbe reviewedto ensure that the distortions such as netting are disallowed. 0 Accounts capture all the transactions; all inflows and outflows inthe public sector. 0 Compilation of accountsfollow a standard system that are consistent with generally accepted accounting principles so that the accuracy of information i s ensured; and The accounts are timely. 113. Review and standardize the methodologyfor public sector accounting: Currently, the public accounting system in Cambodia does not follow a standardized international accounting system and methodology. It i s guided by numerous ad hoc decrees, laws and instructions. In the process, many nontransparent accounting practices are been followed. Offsetting transactions, for example, mainly owing to the shortage of cash, has become a common feature of the money management system. Government tax dues from a private sector supplier are offset against his claims against the government. Apart from the fact that these transactions are nontransparent and result in the loss of detail, this practice can also result inthe loss of revenues owing to the lack of transparency. Offsetting transactions should be completely avoided, and all flows should be treated distinctly. The money due to the government should be collected in a separatetransaction and deposited inthe NT. The payment due from the government should be paid as a distinct transaction. Offsetting i s more than an accounting distortion; it affects boththe transparency of operation and cashmanagement. 114. Ensure comprehensive coverage by the public accounts: The public accounting system should record all the revenue and expenditure transactions, item by item, with supporting documents, including transactions inforeign currencies and user fees and charges for all government agencies. It i s necessary to prepare an inventory of all the spending units supported by the budget and where budgetary inflows are collected. At the same time, it i s important to document different types of transactions arising inthe public sector and ensure that these are captured in the budget classification and the chart of accounts. 115. Review and standardize the Chart of Accounts: Ina well structure classification system, the budget i s organized in a program structure that provide identifying each item of expenditure under specific objectives of spending. Each function may, however, fall under more than one organizational category. 188 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review 116. Zmprovement of Reporting System: The reporting system in Cambodia, should at the most basic level, be able to accomplish the following: Enablefinancial managersinthe MEFand the spending agencies to control commitments and expenditures pressuresto amounts that can be financed within budgetary resource availability. Utilize information provided on all stages of the budget execution system and by all agencies, record and track expenditure from appropriation, through commitment to payment by economic classification, functional unit, and program. Enablethe MEFto prioritize cash allocation, especially inthe case of development spending. Make an accurate forecast of all revenue inflows. 117. Prepare a master plan for the integrated computerizationof Budget Execution: The mission understand that each line agency provides a monthly statement of above-the-line expenditures to the NT within five days of the end of the month. This is one of the strengths of the system. The monthly report should contain, inaddition to budgetary spending, the following: All the commitments visas issuedto the line agency under eachbudget lineitem. All the payment liability incurred where goods and serviceshave beenreceived but payment order has not been issued. Where request for commitment visa has been made but the visa has not been issued, and A projectionof requestsfor the issue of commitment visas inthe next three months. 118. This information should be received both for the recurrent and the investment budgets. This information should be sent to the NT and to the budget department, especially the Cash Management Unit.This will enable the NT and the budget department to assess the expenditure pressures in the next three months and prioritize the spending accordingly. 119. Improve the skill level of accounting staff: Reinforcement of management capacity in the National and Provincial Treasuries i s an essential prerequisite for improving public finance management in Cambodia. In the short term, the mission recommends an implementation of basic level of automation for the National Treasury and Provincial Treasury, this will involve computerizationof the following specific functions: Production of general accounts balance by consolidation of central and provincial balancesby central accounting office. Recordingof each separateexpenditure transaction inthe expenditure office. Recordingof each separaterevenue transaction inthe revenue office; and Recording of each separatecash operation inthe cash office. 120. Inthe provincial treasury, the recordingof all state andprovincebudget transactions shouldbe computerized. One important issue i s that even basic computerization of the accounting system at the NT or its branches will require a coherent and consistent accounting policy and methodology. The current accounting framework and policy, which consists of many unorganized and seemingly ad hoc accounting directives issued from time to time, need to be reviewed and put together in a coherent framework. The basic framework of such an accounting policy should be developed following generally acceptedaccounting principles. 121. Technical assistancewas provided under an IMFprogram and ADB. However, the progress of reformi s very slow and nothing of significance have been achieved since commencement of TA some twelve months ago. The mission i s of the opinion that the Accounting Reform Group set up to steer the reform program i s involved intrying to resolve too many theoretical and practical issues which has AnnexC 189 bogged down the reform effort. The mission proposed a very pragmatic approach to get the reform programon track. This would involve paying attention to basic principles and accuracy of information before tackling more complex issue of accounting principles and standards. 122. The main objective of the f i s t phase of the reform program should be to produce accurate accounting information based on a unifiedchart of accounts. There should be no changes to the current budget classification, budget execution rules or the payment system. Based on the above assumption, the following implementationprogrami s recommended: Require Treasury Accounting Working Group to establish an acceptable set of accounting policies and procedures to be used inpreparation of government financial statements. These principles and procedurescould be basedon internationalpublic accounting standards but adopted to suit conditions currently prevalent inCambodia; The Working Group to submit a revisedchart of accounts combining the current budget classification coding of revenue and expenditure and treasury accounts and expanding the asset and liability accounts to include all government transactions. Develop a simple accounting manualfully explaining the linkage of new chart of accounts to budget classifications and providing clear rules for coding of financial transactions. Require that all government financial statements are generatedthrough the computer system. Addressing the lack of uniform accounting policies and procedures andlack of adequate systems for timely and accurate data production are the two most urgent weaknesses that need urgent attention. As most government transactions relate to revenue collections and payment of recurrent costs, devising accounting policies and procedures should not involve adopting all the currently available international public accounting standards.The Accounting Working Group should be requiredto review the relevant standards and come up with simple andclear policies andrules without getting too involvedwith complex issues that may have only distant future applicability. 123. Although an integrated financial management system should be the ultimate goal, the technical competency of staff and current development of accounting rules and regulations do not warrant embarking on a complex system for the time being. The Bank has successfully assisted in installing stand alone PC based computerized accounting systems for project accounting inCambodia. These systems have used simple software (such as Peachtree) or more complex software (such as Solomons) and are operated by trained Cambodian staff. A more complex project, such as the Flood Emergency Rehabilitation Project, uses Solomons software and i s operated by Cambodian staff with the assistanceof one expatriate accountant. 124. For provincial treasury operations, a stand alone PC with a single user capability i s more than adequate. Two or three person teamcould be trained to operate the system with support from NT. Data transfer from provincial treasuries to NT system could be through disks, as currently practiced in some of the Bank projects with provincial operations. AUDITING 125. The function of control of the public expenditures i s still premature. It subsists the organizational and coordination deficiencies between the different structures intervening inthe internal and external controls. The annual accounts of the government are not audited and certified to be correct. There are also no private sector auditors, though there are some international audit firms mainly to audit the accounts of the multinationals operating inCambodia. 190 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review ExternalAudit 126. Though there has been progress in improving the audit function, the public oversight of government financial management remains extremely weak. While there is a banking and finance committee inthe National Assembly, it i s unclear to what extent the committee i s active inoverseeing government finances. For example, though the TOFE is submitted to the committee, it does not appear that the committee publishes its deliberations. As Cambodia does not publish government financial statements or audit reports, there i s a complete absence of public debate on government financial affairs. The National Assembly doesn't exercise inpractice any control on the public finances and that the finance section has a very limitedhuman and material resources. 127. The government passed a law CS/RKM/0300/10 in January 2000 establishing a National Audit Authority (NAA).The NAA is independent of the executive branch of the government and is empowered to carry out external auditing function of the government. The law provides an appointment of an Auditor-General and two Deputy Auditor-Generals by Royal Decree at the recommendation of the Government and approved by a two-third majority vote of the whole National Assembly members. The Auditor-General shall have a rank and privilege equivalent to that of a Senior Minister. The legal foundation of the NAA i s sound and i s consistent with good international practice. The financial independence of the NAA and its authority to determine the scope of audits are guaranteed by law. The law also requires that each internal audit unit of government departments shall report its findingto the NAA. 128. The NAA i s responsible for executing the external auditing function of the Government. The Auditor General should be empowered to conduct an audit of transactions, accounts, controls, operations, and programs of government institutions, in accordance with generally accepted auditing standards. This includes the National Treasury, all ministries, public financial institutions, municipalities, provincial and contractors or other organizations that have received financial assistance from the government. According to the law, the NAA i s an independent, public entity and shall report directly to the National Assembly, Senate and the Government (for purpose of giving information). The reports issued by the Auditor General shall be considered as a public documents. The NAAhas its own separatebudget provided by the National Budget. 129. NAA staff are hiredon higher salariesthan that paidto civil servants. This has enabled NAA to attract quality staff and require higher ethical standards from its staff. However, NAA acknowledges that a considerable amount of on-the-job training i s required for it to be fully effective. As a young organization with a mission to accomplish, NAA appears to be well motivated. The NAA i s fully staffed (100) and became operational at the beginning of the year 2002. The staff of the NAA has its own separate statute. This superior institution has to be not only competent concerning judgment of the accounts but also on the judgment of the regularity and the quality of the management of the authorizing officers. Staff recruited to NAA will need to develop skills in conducting financial performance and compliance audits. Audit manuals for performance and compliance auditing will be developed to help staff to conduct such evaluations. Audit procedures for reporting and administration will also be developed under the ADB technical assistance. 130. NAA is not yet on the radar screens of many of the donors. NAA badly needs technical assistance to upgrade its staff skills and provide it with equipment (computersflaptops) to effectively carry out its duties. Under the original program put together by ADB, training in theory and practice spanning 12 and 16 months respectively were planned. However, due lack of TA resources, this has not taken place. Actually, the external control i s almost non -existent because the NAA doesn't have competent and well trained staff who can assume these responsibilities. There i s an urgent need for donor funding of training activities and equipping of NAA. AnnexC 191 131. The NAA carried out an audit of the year 2001 budget implementation as its first audit assignment. The audit was completed in September 2002. A report was issued to National Assembly butthis is not yet published. As its first audit report is yet to be published, it is not proper to make any judgments on capacity of NAA. InternalAudit 132. The Public Expenditure Management System in Cambodia are excessively compliance oriented. There i s a unit in the MEF that monitors compliance through periodic on the spots visits to the various units.This unit, however, does not follow a standard internal audit system, which though internal to the MEF, i s an external unit for the line agencies. There i s no internal audit system in Cambodia, and there i s a need to establish internal audit capacity inthe line agencies, especially as the authorities plan to delegate greater expenditure managementto the line agencies. There i s also a need to start to prepare a manual for internal audit, the hierarchical and reporting relationship of internal audit unit inthe line agencies, audit calendar and strategy, and training of the audit staff. 133. The internal control is represented in Cambodia by the internal auditing function and the Department of Financial Inspection (DFI). According to the law CSRKM/0300/10 of January 13, 2000, the Internal Audit Department shall be established within each ministry and state enterprises. The internal audit department shall report to its head of ministry, institution and shall submit its report and conclusion to the National Audit Authority (NAA). However, according to the NAA, there is no report submitted to them from the InternalControl Department. Analysis 134. The ADB technical assistancei s supposed to design the communication and information flows between the Auditor General's Office and internal audit in line ministries, the Finance and Banking Commission of the National Assembly, DFI,provincial governments and agencies, and the Ministryof the National Assembly-Senate Relations and Inspection. The TA will focus on developing the framework within MEFand MOHfor effective internal auditing. By having an effective internal audit function, management in a ministry is assured that the control framework i s adequate and i s working as intended. NAA, by reviewing the work performed by the internal auditor, can form an opinion on the adequacy and quality of the internal audit work done. Where the external auditor i s satisfied that the work of internal audit meets professional standards, the external auditor can than rely on this work and reduce the extent of review that i s required. If this pilot test i s successful, the concept and methodologies can be extended to the other line ministries and state enterprises as a model to be followed. NAA's effectiveness inthe longer term will be enhanced. 135. In addition, a framework will be developed for strengthening the Finance and Banking Commission (FBC) of the National Assembly. This will include streamlining the reporting process between NAA and FBC and preparing a mandate to enable FBC to function as a key participant inthe framework of accountability. FBC's role will be expanded to establish a closer working relationship with NAA in the areas of audit planning, budget funding and following up key audit findings. The activity will build on training to be provided by the Canadian international Development Agency (CIDA). FBC can be a conduit for dialogue between NAA and parliament; have a clear role in the method of selecting and appointing the auditor general and deputy auditor generals; and enhance the auditor general's independence by supporting the auditor general's requests for budget funding to parliament and ensuring that the office of the auditor general i s made accountable for its activities. Recommendations 136. The government has also embarked upon creating internal audit units in key ministries. The mission believe that internal auditing i s an important function required to improve internal control 192 Cambodia: Integrated Fiduciary Assessment and Public Expenditure Review environment of the line ministries. However, this i s not a priority at the moment as government capacity to manage multiple improvement activities i s limited. This effort i s undertaken due donor pressure and i s another case where the government undertake more than it can handle. Government's scarceresources are better spent in strengthening NAA at this stage. 137. With the establishment of the NAA, the role of MEF's DFIwill be redefined. DFIcould be the internal auditor for MEF and could monitor the implementation of the budget of agencies funded by the Government's budget. It i s thus essential that DFI be given a clear role, recognizing the responsibilitiesof NAA,internal audit, and the National Assembly-Senate Relation and Inspection. To effectively perform in its new role, DFI will require further training and methodologies to be developed for internal audit and budget inspection. The TA will help DFI develop an internal audit charter and manual of procedures, abudget inspection manual, and staff training. 138. The NAA i s fully staffed and became operational inearly 2002. The NAA completed an audit of the 2001 budget in September 2002 and issued its report to the National Assembly. The report has not yet been made public, but it i s imperative that NAA reports become public documents after approvalby the NationalAssembly. In2003 the NAA i s programmedto carry out nine audits of public bodies, including provinces, communes, and line departments. The NAA seems to be moving forward purposefully, yet it i s clear that a considerable amount of on-the-job training and technical assistance will be required for it to be fully effective. RECOMMENDATIONSSUMMARY 139. The government has embraced many reform initiatives sponsoredby its development partners. These include major initiatives such as performance based budgeting, budget decentralization, medium-term budget framework, standardization of accounting, treasury reforms and many others. The agenda i s totally beyond government capacity and misses fundamental weaknesses particularly in the payment and reporting system. Therefore, it i s in the intention of this report not to add to this long list but provide a framework for step-by-step approach to reformstarting with the most basic. 140. The mission opines that reliable, accurate and timely information i s the foundation of any meaningful fiscal management improvement program. Therefore, the mission recommends that the government focus on building this foundation as the first priority. This should be accomplished without making any fundamental changes to the current budget execution system or the payment system. 141. The following i s a summary of the mission's recommendations: 142. Budget Process Evaluatethe operational efficiency, compliance, and performance of PAPby undertaking a review and tracking survey before further expansion (focusing on education, health, and agriculture) in order to gradually eliminate current multiple procedures. Refinejob descriptions, institutionalarrangements, and authorizationprocedures for outposted financial controllers. Outpost all financial controllers from the Budget and Financial Affairs Department (BFAD)to the spending mini~tries.~ 143. Treasury Management Define and apply threshold for National and Provincial Treasury expenditure operations to be executed by check or bank transfer, and use the banking systemfor all intragovernmentaltransfers. Reformunderway: financial controllers outpostedto MOHand MOEYS. AnnexC 193 Issue a directive requiringall payments above an agreedamount (law states CR 2.5 million) be made only by check or bank transfer; Issue a directive requiring all large tax payers to pay taxes directly to Treasury Single Account with the NationalBank; Undertake an inventory of all revenue accounts and draw up a schedulefor the elimination of revenue retentionaccounts that have no legalbasis and consolidate revenues into the Treasury SingleAccount with legalbasis. Integrate management of foreign currency revenuesto National Treasury inaccordance with the law. Empower the CashManagement Committee to play a greater role incash allocations to spending ministries 144. Public Accounting and Reporting 0 Require Treasury Accounting Working Group to design accounting policies and procedures based on applicable internationalpublic accounting standards; 0 Require Treasury Accounting Working Group to submit a revisedchart of accounts combining the current budget classificationcoding of revenue and expenditure and treasury accounts and expanding the asset and liability accounts to include all government transactions. 0 Develop a simple accounting manual fully explaining the linkage of new chart of accounts to budget classifications andproviding clear rules for coding of financial transactions. 0 Review requirements for computerizationof National and Provincial Treasury operations. 0 Train adequate staff ineach of NT and provincial treasury offices inthe computerized accounting system. 0 Require that all government financial statements are generatedthrough the computer system. 145. In order that the above reforms are managed efficiently, it is recommended that Priority Management Group be established with staff drawn from the National Treasury, Provincial Treasuries, MEF Budget and Financial Affairs Department and key line ministries. This group should work full time with assistancefrom external technical experts inimplementing the above program. 146. Control Functions Establishan effective Parliamentary oversight on public finances by providingnecessary training and equipment. 0 Establishan effective internal audit function and control framework inline ministries. 0 Strengthen the National Audit Authority by conducting training on International Auditing Standards and providing necessaryequipment. Financial statements to be audited annually by the NAA. Publishaudited financial statementsand make themaccessibleto the public. AppendixB Cambodia Country FinancialAssessment Accountability - Prioritizedlist of Recommendation Responsible Technical Time Frame Objectives Recommendation Unit Assistance Medium Short Medium Short Term Term Term Term 2004- 2003 2004- 2005 2003 2o05 1. For effective budget management. a. Evaluatethe operationalefficiency, compliance,and performanceof PAP by undertakinga review and tracking survey before further expansion (focusingon education, health, MEF and NT X X and agriculture)in order to gradually eliminate current multiple procedures. b. Refinejob descriptions, institutional TOR for arrangements,and authorization proceduresfor consultant to be X outpostedfinancial controllers. by the WB c. Outpostall financial controllersfrom the Budget and FinancialAffairs Department(BFAD) to the MEF spending ministries 2. To improve treasury operations and cash management. a. Define and apply thresholdfor National and ProvincialTreasury expenditureoperations to be executed by check or bank transfer, and use the MEF and NT X banking system for all intragovernmental transfers. b. Undertakean inventoryof all revenue accounts and draw up a schedule for the elimination of revenue retention accountsthat have no legal MEF X X basis and consolidate revenues into the Treasury Single Account with legal basis. c. Integratemanagement of foreign currency revenuesto NT. MEF X d. ImproveCash ManagementCommittee and bank account reconciliation processes. MEF and NT X 195 196 Cambodia: IntegratedFiduciary Assessment and Public Expenditure Review Prioritizedlist of Recommendation Objectives Recommendation Responsible Technical Time Frame Unit Assistance Medium Short Medium Short Term Term Term Term 2004- 2o03 2004- 2003 2005 2005 3. To improve Government's financial accountability. a. Revised Chart of Accounts combiningthe current budget classificationand treasury National X X accounts. Treasury b. TechnicalAssistanceto be provided for computerizationof NationalTreasury Operations NT andWB X X X and training. c. Prepare financial statements in accordance National with adopted standards. Treasury X d. Developmentof implementationplan for provincialtreasuries. NT andWB X X 4. To strengthen Government's control functions and transparency. a. Establish an effective Parliamentary oversight National on publicfinances. Assembly X X b. Establishan effective internalaudit function and control framework in line ministries. Line Ministries X X c.Strengthenthe NationalAudit Authority (NAA) by conductingtraining of their staff on NationalAudit InternationalAuditing standardsand providing Authority X necessaryequipment. d. Financialstatementsto be audited annually by NationalAudit the Auditor General. Authority X e. Publishauditedfinancial statementsand make National them accessibleto the public. Assembly X REFERENCES Adachi, M.2002. "Cambodia-Outstanding External Debt and Rescheduling Scenarios." IMF (International Monetary Fund) Staff Report, EBS/O2/13. ADB (Asian DevelopmentBank) TA 3257-CAM: 2001 Surveys. . 3257-CAM. "Strengthening the Maintenance Planning and Management Capabilities at the Ministryof Public Works and Tranport." Agrifood Consulting International.2002. "Rice Value Chain Study: Cambodia." World Bank. Bank Estimates Based on MEF, TOFE report, December 2002. Bannock Consulting. "Developing Strategies for Priority MissionGroups towards Sustainable Civil Service Reform." November 2001. Bitran, R., C. Espinosa, J. Knowles and V. Turbat. 2002. "Cambodia Health Financing Study PromotingPro-poor Health FinancingPolicies in Cambodia." World Bank. Cambodia. "National Poverty Reduction Strategy." December 2002. Phnom Penh. . SES, 1999. Cambodian Health Promotion and Primary Health Care Project. "A Training Needs Assessment on PAP." April 2001. CARS. "Civil Service Rationalization Strategy: MediumTerm Wage Projection, 2002-2006." June 27,2002. ."Operational Reviewof Service Delivery." October 2002. CSES. Socio-Economic Survey - 1997 and 1999. D e Loof, F., F. Bonnet, and J. Thome. 2001. "Takeo referral hospital 1997-2000: pioneering a hospital financing scheme inthe Kingdom of Cambodia." Swiss RedCross, Phnom Penh. DHS.Demographic and Health Survey. 2000, Dom, Catherine, Tim Ensor, and Leon Bernard Suy. 2003. "Results-oriented Public Expenditure Management in Cambodia." OverseasDevelopment Institute, Working Paper 201. London. Feenstra, Pieter. 2001. "Contracting and Similar Experiences inHealth inCambodia." Medicam. Phnom Penh. FisheriesAction Coalition Team, incollaborationwith the Environmental Justice Foundation. 2002. "Feast or Famine? Solutions to Cambodia's Fisheries Conflicts." London. 197 198 Cambodia: IntewatedFiduciarv Assessment and Public ExvenditureReview GRSP. (lst) Seminar (Association of SoutheastAsian Nations). March 2001. ASEAN H D S . Household Demand Survey. 1996. IMF(International MonetaryFund). Government Finance Statistics. Various years Kaufmann, Daniel, Aart Kraay, and Pablo Zoido-Lobaton. 1999. "Governance Matters." Policy Research Working Paper 2196, World Bank, Washington D.C. . Government Finance Statistics, SIMA Database, Cambodia (Budget Law, TOFE,and IMF staff estimates): a/ 1992-1993, b/ 1995-1996, c/ 1999-2000, d12001. Klugman, Jeni and R.Taliercio. 2003. "Cambodia: PRSP and BudgetLinkages." Manuscript. World Bank: Washington, D.C. Law on Taxation. 1997. L e Houerou and R.Taliercio. 2002. "Medium Term Expenditure Frameworks: FromConcept to Practice-Preliminary Lessonsfrom Africa." Africa Region Working Paper Series, No. 28. World Bank: Washington, DC. Litvack, Jennie and Jessica Seddon, editors. 2000. "Decentralization BriefingNotes." WBI (World Bank Institute) Working Paper, World Bank, Washington, D.C. MAFF."Annual Conference on Agriculture, Forestry and Fisheries." April 10-11, 2002, p. 15. . January 2003. "Report on the Situation ofAgriculture, Forestry, and Fisheries in2002." McLure, Charles and Jorge Martinez-Vazquez. 2002. "The Assignment of Revenuesand Expenditures inIntergovernmentalFiscal Relations." Manuscript. Medium-TermExpenditureFramework, 2003-2005. Draft document, 12 September 2002, p.1. Ministry of Health. Expenditure & BudgetBooks, 1998-2002. ."Health Center Manual," PhnomPenh, 1997 MOEYS. PAP Basic Education2000 and 2001 Update. MOP. "A Poverty Profile of Cambodia." 1997 and 1999. MRD."Policy for RuralRoad." October 1999. NGO Forum on Cambodia. "A Rapid Comparison of the NGO Statementto the 2002 CG Meetingandthe SecondDraft of the PRSP inCambodia." November 2002, p.8. Poverty MappingResults from WFP and "Estimation of Poverty Rates at Commune Level in Cambodia." RGC, October 2002. Royal Government of Cambodia. "MOEY S, Poverty Impact Analysis." 2002. References 199 Sakko, Jan. 1999.Access, Transport and Local Economic Development, ILO ASIST Songco, Dan 2001. "Accountability to the Poor: Experiences in Civic Engagement inPublic ExpenditureManagement." ParticipationGroup/Social DevelopmentDepartment, World Bank. Taliercio, Robert. 2003. "Designing Performance: The Semi-Autonomous RevenueAuthority Model inAfrica and Latin America." Manuscript. World Bank: Washington, DC. Theodore Valentine, October 2001. "Towards a MediumTerm Pay Reform Strategy for the Mozambique Public Service." Van Damme, W. and B.Meessen. 2001. "Sotnikum New Deal, the first year." MSF Holland/Belgium, Phnom Penh. World Bank, 2000. "Cambodia Governance and Corruption Diagnostic: Evidencefrom Citizen, Enterprise and Public Official Surveys." World Development Report. 2004. World Bank. Forthcoming.