Report No. 29421-KE Kenya Public Expenditure Review 2004 Report on the Structure and Management of Public Spending February 2005 Poverty Reduction and Economic Management 2 Country Department for Kenya Africa Region Document of the World Bank, co-produced with the Government of Kenya FSP Fiscal Strategy Paper GDP Gross domestic product GFS Government Finance Statistics GoK Government o f Kenya HCDA Horticultural Crops Development Authority HIPC Heavily IndebtedPoor Countries Initiative HIV Humanimmunodeficiency virus IFMIS IntegratedFinancial Management InformationSystem IMF InternationalMonetary Fund INTOSAT InternationalOrganization o f Supreme Audit Institutions KARI Kenya Agricultural Research Institute KDB Kenya Dairy Board KCC Kenya Cooperative College KEMFRI Kenya Marine and Fisheries ResearchInstitute KEPHIS Kenya Plant HealthInspectorate Service KETRI Kenya Trypanosomiasis Research Institute KEVEVAPI Kenya Veterinary Vaccines Production Institute KFA Kenya Farmers' Association KIPPRA Kenya Institute for Public Policy Research and Analysis K M C Kenya Meat Commission KRA Kenya RevenueAuthority KRB Kenya Roads Board K Sh Kenya shilling KVB Kenya Veterinary Board KWS Kenya Wildlife Service LA Local authorities LATF Local Authority Transfer Fund MDGs MillenniumDevelopment Goals MOA MinistryofAgriculture MoCDM Ministry o f Cooperative Development andMarketing MoF MinistryofFinance MoLFD MinistryofLivestock andFisheries development MoEST Ministryof Science, EducationandTechnology MPER Ministerial Public Expenditure Review MPND MinistryofPlanningandNationalDevelopment MTEF Medium-TermExpenditure Framework NARC National Rainbow Coalition NCPB National Cereals and Produce Board NGOs Nongovernmental organizations NHC National Housing Corporation NHIF National Health Insurance Fund NPV Net present value NSHIS National Social Health Insurance Scheme NSIS National Security Intelligence Services NWCPC National Water Conservation andPipeline Corporation NYS National Youth Service O&M Operations and maintenance (expenditures) OVP Office of the Vice-president and Ministryo fNational Reconstruction PAC Public Accounts Committee PEM Public Expenditure Management .. 11 PEM-AAP Public Expenditure Management Assessment and Action Plan PER Public Expenditure Review PETS Public Expenditure Tracking Surveys PIP Public Investment Programme PRGF Poverty Reductionand GrowthFacility PRSP Poverty ReductionStrategy Paper QBR Quarterly Budget Review RMFL Road Maintenance Fuel Levy Fund SAGAS Semiautonomous government agencies SWG Sector Working Group TTC Teacher training college TVERS Targeted Voluntary Early Retirement Scheme TVET Technical and vocational education and training TSC Teachers Service Commission U S A I D U.S.Agency for InternationalDevelopment VAT Value added tax WTO World Trade Organization ... 111 Table of Contents A B B R E V I A T I O N S AND A C R O N V M S .................................................................. i A C K N O W L E D G M E N T S ..................................................................................... x EXECUTIVESUMMARY ............................................................................................... xi 1. INTRODUCTION...................................................................................... 1 OBJECTIVESOF THE PER WORK ....................................................................................... 2 RECOMMENDATIONS OF THE 2003 PER ............................................................................ 3 ORGANIZATION OF THE PER WORK IN2003-04 ............................................................... 3 2 AGGREGATE FISCAL PERFORMANCE.............................................. . 5 RECENT ...................................................................... 5 RECENT FISCALPERFORMANCEAND TRENDS................................................................... MACROECONOMIC PERFORMANCE 6 Revenue.................................................................................................................... 6 Expenditure.............................................................................................................. 7 Deficit Financing and Public Debt.,........................................................................ 8 FISCALPOLICYOPTIONSINTHEMEDIUM TERM ............................................................. 11 Debt Reduction Targets and the Deficit: SomeProjections .................................. 12 Challenges and Risks............................................................................................. 13 Comments on the medium-term fiscal strategy in the ERS:................................... 14 3. COMPOSITIONOF GOVERNMENT EXPENDITURE ...................... 16 ECONOMIC 16 Recurrent Expenditure........................................................................................... COMPOSITION............................................................................................... 16 Development Expenditure...................................................................................... 20 Public Service WageBill ....................................................................................... 21 EXPENDITURE COMPOSITIONBY MINISTRIES .................................................................. 27 Signals of Policy Intent-Changes in the Budget over the Last Two Years ..........30 THEMTEFAND ALLOCATIVEBEHAVIOR ...................................................................... 33 Poverty Focus ofspending and Core Poverty Programs...................................... 35 LOOKING FORWARD-SPENDING PRIORITIES INTHE MEDIUM TERM............................. 37 Public Expenditure Cost of Reaching Millennium Development Planned MTEF Sector Ceilings and ERSPriorities ............................................. Goals (MDGs)........................................................................................................ 37 -38 Linking ERSPriorities with Budget Allocations.................................................... 39 4. PUBLIC EXPENDITURE MANAGEMENT......................................... 40 RECENT DEVELOPMENTS ................................................................................................ 40 INDICATORS OF PEMPERFORMANCE ............................................................................. 41 Composition of Outturn Expenditure versus OriginalBudget .............................. 41 Stock of ExpenditureArrears (Pending Bills) ....................................................... 44 Systemic Issues Raised by the Discussion on Deviations and Pending Bills.........45 iv BUDGET PLANNINGAND FORM~LATLON ......................................................................... 46 TheMTEF Objectives and Process in Kenya ........................................................ 46 OTHERISSUES MANAGEMENT INPUBLIC EXPENDITURE ................................................. 53 SUMMARY OFRECOMMENDATIONS AND ACTION PLANFORPEMREFORM ...................60 5.MINISTRY MINISTERIALEXPENDITUREREVIEWS......................................... 61 OF EDUCATION.SCIENCEAND TECHNOLOGY (MOEST) ................................ 61 Aggregate Expenditure .......................................................................................... 61 Primary Education.,............................................................................................... 62 SecondaryEducation ............................................................................................. Tertiary Education................................................................................................. 65 67 Technical and Vocational Education and Training (TVET).................................. 67 68 Overview ofNeeds in the Sector............................................................................ Other Types of Education ...................................................................................... 69 Expenditure Projections......................................................................................... 69 Other Areasfor Increased Spendingand Future Funding Sources ...................... 70 Budget Analysis, Execution, and Expenditure Management ................................. 71 72 Summary of Recommendations.............................................................................. ThePER Process.................................................................................................... 72 MINISTRY HEALTH OF ..................................................................................................... 73 73 Sector Background and Broad Policy.................................................................... Sector Performance ............................................................................................... 74 Aggregate Expenditure .......................................................................................... 75 Composition of Expenditure .................................................................................. 75 Budget Execution .................................................................................................. -78 Effectiveness of SpendingIssues............................................................................ 80 FutureFunding Requirements............................................................................... 81 Public-Private Roles in Financing.,....................................................................... 81 82 Conclusions.,......................................................................................................... TheNational Social Health Insurance Fund (NHSIF).......................................... -82 MINISTRIES AGRICULTURE, OF LIVESTOCKFISHERIESDEVELOPMENT, AND AND COOPERATIVE DEVELOPMENTMARKETING AND ...................................................... 83 Sector Background and Broad Policy.................................................................... 83 Aggregate Expenditure .......................................................................................... 84 Composition of Expenditure .................................................................................. 86 88 Development Expenditure..................................................................................... Transfers to Parastatals......................................................................................... -89 Budget Execution ................................................................................................... 90 Development Budget and Donor Funding............................................................. 92 Compliance with Expenditure Allocations ............................................................ 92 Effectiveness Issues............................................................................................... -93 FutureFundingRequirements............................................................................... 95 Conclusions.,.......................................................................................................... 96 MINISTRY ROADS, OF ..................................................... 98 Mission and Administrative Structure ................................................................... PUBLIC WORKSAND HOUSING 98 Aggregate Expenditure Trends.............................................................................. 98 V Roads.................................................................................................................... 100 Is Spending in the Roads Subsector SufJicient. and How Effective Is It? ............101 Conclusions.......................................................................................................... i04 MINISTRY WATERRESOURCESMANAGEMENTDEVELOPMENT OF AND ....................... 105 Structure of the Ministry and Current Reforms and Restructuring ..................... 105 106 Aggregate Expenditure Trends............................................................................ Challenges in the Sector and Expenditure Priorities .......................................... 107 Efficiency and Effectiveness of Expenditures ...................................................... 109 Conclusions.......................................................................................................... 111 6 SUMMARYOF KEY RECOMMENDATIONS AND THE PER . PROCESS IN2004/05 ........................................................................... 112 DEVELOPING IMPLEMENTINGA CREDIBLEMEDIUM-TERM FISCALSTRATEGY.....112 AND PRIORITIZING EXPENDITURE ACHIEVEERSOBJECTIVES TO ........................................ 113 STRENGTHENING PEM.................................................................................................. 114 &,ECOMMENDATIONSFOR SPECIFICMINISTRIES ............................................................ 115 Ministry of Education. Science and Technology.................................................. 115 116 RECOMMENDATIONSFORTHEPERPROCESS 2003/04............................................. Ministry of Health................................................................................................ IN 117 ANNEXES .................................................................................................. 119 Annex 1 Deficit Implications ofAlternative Debt Targets............................................ . 120 Annex 2. Medium-TermFiscal Framework and Strategy inthe ERS............................ 122 Annex 3. Changes inthe Pay andEmployment Structure o f Civil Service ..................... 128 Annex 4. Status o fMillenniumDevelopment Goals inKenya Annex 5.MTEF Sector Working Groups: Composition andTerms of Reference..........129 ....................................... Annex 6. List o fFundAccounts Establishedunderthe Exchequer and Audit Act.........131 134 Annex 7 .Action Plan to Upgrade PEM Capacity inKenya............................................ 135 STATISTICAL ANNEX............................................................................. 139 List of Figures and Tables Figures Figure 3.1, Relative Shares o fRecurrent andDevelopment Expenditure ......................... 16 Figure3.2. Shares inTotal Expenditure. 1996/97-2002/03 .............................................. Figure 3.3. Transfers to Parastatalsas Share of Total Expenditure................................... 17 18 Figure 3.5. ChangeinDistribution ofPublic Expenditure (Actual % of Total) ...............23 Figure3.4. Structure ofthe Public Service Employment in2002 ..................................... Figure 3.7. Budget Allocation and Actual Expenditure for Recurrent CPPs..................... Figure 3.6. Composition o fDevelopment Spending (Actual), 1998/99-2002/03 .............28 30 Figure3.8. Budget Allocation andActual Expenditure for Development CPPs ...............35 36 Figure 3.9. Relating Core Priority Programsto ERS Priorities ......................................... 39 vi 47 Figure 4.2. The KenyaMTEF: ProposedBudgetProcess andTimeline........................... Figure 4.1. The PER. MTEF. andBudgetPreparationCycle: Kenya............................... 54 Figure 5.1. Distribution of RecurrentSpendinginMinistry ofHealth.............................. 77 Figure 5.2. Ministry of Health-Actual versus BudgetedExpenditure............................. 79 86 Figure 5.4. DistributionofMoLFD Total Expenditure(Actual), 2002103........................ Figure 5.3. Distributionof MOA RecurrentSpending(Actual), 2002/03.......................... 87 Figure 5.5. Expenditureby Subvotefor the MoCDM(Actual), 2002/03.......................... Figure 5.6. Distribution of MoCDMTotal Expenditure.................................................... 88 88 Figure5.7 Distributiono fMOADevelopmentSpending(Actual), 2002/03.................... . 90 Figure 5.8. Recurrent Transfers to Parastatals(Actual), 2002/03...................................... 90 Tables Table 2.1. Revenue andIts KeyComponents(% o f GDP). 1999/2000to 2003/04............7 Table 2.2. Expenditure andIts KeyComponents(% of GDP). 1999/2000-2003/04 ..........8 9 Table 2.4. Extemal Loans andGrants. 1999/2000-2002/03 (K ShMillion)....................... Table 2.3. Deficit andFinancing(K ShMillion). 1999/2000-2003/04............................... 9 Table 2.5. Variability of BudgetResources. 1994/95-2002/03 ......................................... 10 Table 2.6. Public Debt (K ShBillion). 2000-04 ................................................................ 11 Table 3.1. Transfersto Parastatalsby FunctionalCategories-Share of Total Table 3.2. DevelopmentExpenditure,ExcludingNet Lending (K Shbillion).................21 Expenditure(%) ........................................................................................................ 19 Table 3.3. Public ServiceWage Bills Relativeto Other MacroeconomicIndicators: 22 Table 3.4. Employmentinthe Public Sector (Thousand), 1995-2003 .............................. Selected Sub-SaharanAfrican Countries (%)........................................................... Table 3.5. Evolution inTeachers' RealWages, 1996/97-2002/03 ................................... 24 25 Table 3.6. EstimatedAverage Real Earningsper Employee (K Shper Year), 1997-2003a ............................................................................................................... 25 Table 3.7. Total Ministerial Spendingby BroadFunctional Categories, 1999/2000- 2002/03 ..................................................................................................................... 28 Table 3.8. Actual Recurrent Spending, 2001/02 and2002/03 (Current K ShMillion) ....29 Table 3.9. Changes inBudgetedRecurrent Expenditure, 2002/03-2003/04 (Current K ShMillions) ............................................................................................ 30 Table 3.10. Underspendingofthe DevelopmentBudget, 1998/99-2002/03 (Current K ShMillion).............................................................................................. 32 Table 3.11 .DevelopmentBudgets Comparedwith Actual Spending (Current K Sh Million) ..................................................................................................................... 32 Table 3.12. Index of ChangeinShares inMTEF Ceilings, BudgetAllocations, and 33 Table 3.13. Shares of 2003104 MTEF and Budget (MTEF Sectors, %)............................ Actual ExpenditureRelativeto 1999/2000a............................................................. 34 Table 3.14. MTEF Sector Ceilings (% o f Total) ............................................................... 38 Table 4.2. Stock ofPendingBills (K Shbillion) ............................................................... Table 4.1. Budget Outturn. 2000/01-2002/03: Deviation of Actual fi-omPrinted (%).....42 44 vii Table 4.3. Summary of Budgeting andReporting Status of SAGASand Fund Table 4.4. Status of Audit Reports for Appropriation Accounts and Other Funds............55 Accounts ................................................................................................................... 58 Table 5.1. Total Actual MoEST Expenditure. 2000/01-2003/04 ...................................... 61 Table 5.2. Public Education Spendingaround the World. 2000 ........................................ 62 Table 5.3. Projected Education Expenditure Targets Modeled inthe MPER (Current K Shbillion) ............................................................................................... 69 Table 5.4. Total Public Spendingon Health. 1999/200O-2002/03 (Actual. K ShBillion) ............................................................................................... 75 Table 5.5. Recurrent Public Spendingon Health, by Level of Service, 2000/01-2002/03 Table 5.6. Total Public Spendingon Agriculture, 2000/01-2003/04 ................................ (Actual K ShMillion) ............................................................................................... 76 Table 5.7. Total Public Spendingon Livestock andFisheries, 2000/01-2003/04 ............85 Table 5.8. Total Public Spendingon Cooperative Development, 2000/01-2003/04 ........85 85 Table 5.9. Deviationbetween Budgeted and Actual Expenditures for Agriculture, IncludingLivestock, Fisheries, andCooperative Development Activities...............86 Table 5.10. Transfers to Parastatalsas Percentageof the MOA'SRecurrent Budget, 89 Table 5.11.PercentageDeviations between Budgeted andActual Expenditures for 2000/01-2003/04 ...................................................................................................... Table 5.12. Actual Expenditure ofthe Ministry, 1999/2000-2003/04 .............................. the MoCDM.............................................................................................................. 91 99 Table 5.13. Ministry o fRoads, Public Works andHousing: Breakdown of Total Actual Expenditure by FunctionK ShMillion), 1999/2000-2002/03 ................................. 99 Table 5.14. Ministry o fRoads, Public Works and Housing Shares: Total Expenditure byEconomic Classification, 1999/00-2002/03 (K Shmillion)................................ 99 Table 5.15. Expenditure on the Classified Roads Sector, 1991/92-2001/02 (US$ Million, 2001 Prices) ..................................................................................... Table 5.16. Disbursementsof RMLFRevenues, 2001/02-2003/04 (KshMillion) .........100 101 Table 5.17. The State o f Roads inKenya (Kilometers) ................................................... Table 5.18. Priority Capital Requirementsinthe Roads Sector, 2003-07 ...................... 102 102 Table 5.19. Trends of Combined Expenditure byAll Departments andAgencies Belonging to the Ministryfrom January 2003 Onward (K Shbillions) .................107 Table 5.20. Three-Year Breakdown of Expenditure (Actual) by Functionfor All Departmentsand Agencies Belonging to the Ministry from January 2003 Table 5.2 1. Total (Actual) Expenditure by Economic Classification, 2000/01-2002/03 Onward (K ShMillion) ........................................................................................... 108 (KShMillion)......................................................................................................... 108 StatisticalAnnex Tables Table A.l. Summary of KeyMacroeconomic Indicators................................................ Table A.2. RealGDP Growthby Sectors, 2000-03 (Constant 1982Prices) ...................140 140 Table A.3. Central Government Financial Operations 2000/01-2003/04 (KShMillion.......................................................................................................... 141 ... Vlll Table A.4. Central GovernmentFinancialOperations2000/01-2003/04 142 Table A.5. Shares inTotal ExpenditureandNet Lending............................................... (% of GDP) ............................................................................................................. 143 Table A.6. Transfers to Parastatalsand Statutory Bodies................................................ Table A.7. Ministerial RecurrentExpenditures(K ShMillion)...................................... 144 Table A.8. Ministerial Shares ofRecurrentExpenditure................................................. 147 148 Table A.9. Ministerial DevelopmentExpenditure(K Shmillion)................................... 149 Table A.10..Ministerial Shares of DevelopmentExpenditure......................................... 150 151 Table A.12. Ministerial Shares of Total Expenditure...................................................... Table A.11 Total Ministerial Expenditures(K ShMillion)............................................ 152 Table A.13.Total Ministerial 153 Table A.14. Actual Development Spendingby Functions.............................................. Expenditures(as % of GDP)............................................ 154 Table A.15 . Changes in the DevelopmentBudget2002/03-2003/04 (Printed Table A.16.MTEF, Budget, andActual DevelopmentSpendingby Sector...................154 Estimates, K ShCurrent Million) ........................................................................... 155 Table A.17. Budget versus OuttumbyKeyMinistries, 2002/03, and Comparison Table A.18. Forward-Looking MTEF Ceilings, 2003/04-2006/07 (K ShMillion) ........156 with 2003/04Budget............................................................................................... 157 Table A.19 . Civil ............................. Table A.20. Fiscal Outcomes, 2000/01-2002/03 ............................................................. ServiceEstablishment,Excluding Teachers, 2002 158 159 Table A.21. Budget Outtum, 2000/01-2002/03: Deviation of Actual from Targets (%).............................................................................................................. 160 ix Acknowledgments The Public Expenditure Review (PER) 2004 Report is the result ofjoint work carried out by the World Bank and the Government o f Kenyainthe context of the PERprocess that began in October 2003. As part o f the process, 27 Ministerial Public Expenditure Reviews were preparedby line ministries and departments with active donor support to the Ministries o f Health, Education and Agriculture. The PER process was jointly coordinated by Messrs Nelson Muturi and George Anyango through the PER Technical Working Group and supervised by the PER Steering Committee. Both the Working Group and the Steering Committee are chaired by the Permanent Secretary o f the Ministryof Planning and National Development, Mr.David S. 0.Nalo. Representation i s drawn from the Ministries of Finance, Planning and National Development, line Ministries, departments and development partners. The Government team for preparation of the PER Report was led by George Anyango and included Zachariah Mwangi, Edward Gakunju, Dickson Khainga, Jane Kiringai, Henry Honyiego, Samuel Kiiru, MiluMuyanga, G. N.Kirori, Michael Gitau, and Henry Wara. Guy Jenkinson (European Community delegation) also provided inputs to the report. The World Bank team was led by Praveen Kumar and included William James Smith, Andrew Sunil Rajkumar, Sibel Kulaksiz, Andrew Karanja, Michael Mills,Patrick Mamboleo and Marjorie Kingston. The Bank team was guided by Makhtar Diop (Country Director), Kathie Krumm (Sector Manager), Robert Blake (previous Sector Manager) and Frederick Kilby (Lead Economist). David Bevan (Oxford University, U.K.)andJohnMukui(Nairobi) provided consultancysupport. Yaw Ansu, BennoNdulu (World Bank), and Allen Schick (University o f Maryland) acted as peer reviewersfor the report. The Government o f Kenya has separately published the Public Expenditure Review 2004, very much in line with this report, with a view to disseminating the findings widely to stakeholders within Kenya. Details of the report may be accessed on the Ministry o f PlanningandNational Development website: http://www.plannine.g;o.ke/ X EXECUTIVESUMMARY Context 1. This report, Public Expenditure Review 2004, is the final product o f the Public Expenditure Review (PER) work program carried out during the financial year 2003/04. As part o f the work program, 27 Ministerial Public Expenditure Reviews (MPERs) were produced after the government decided to extend regular expenditure reviews to all ministries (compared with eight ministries in the previous year). The MPERs were used toward preparation of Sector Reports during the Medium-Term ExpenditureFramework (MTEF)budgetprocess. 2. The PER report carries an ex-post evaluation o f (a) budget performance in 2002/03 and, where information is available, for 2003/04 against fiscal objectives and spending priorities laid down in the Economic Recovery Strategy for Wealth and Employment Creation (ERS); and (b) the Performance o f budgetary institutions in the context o f government's public expenditure management (PEM) reform program. It also pulls together main features o f spending patterns andpolicy issues inkey line ministries from the MPERs. In each o f these areas, the report makes recommendations for further reform. The report also draws upon the findings o f the PEM-AAP update and the MTEF Review. ' KEY FINDINGS THE PER OF Aggregate Fiscal Performancein 2002103 and 2003104 3. The government's aggregate fiscal policy objective is to reduce the budget deficit through containing increase in overall expenditure, sustain the revenue to gross domestic product (GDP) ratio above 21 percent, and use a higher level of external concessional financing to finance the deficit andnet reduction o f domestic debt. 4. In 2002/03, the overall fiscal deficit after grants was 4.0 percent of GDP. According to provisional data, the fiscal deficit after grants at the end o f 2003/04 is expected to be 2.5 percent o f GDP against a target o f 3.3 percent. The apparently good performance against the target i s a result o f good revenue collection, higher inflow o f grants, and a sharp cut inplanned expenditure. The expenditure cut, inturn, resulted from The Public ExpenditureManagementAssessment andAction Plan(PEM-AAP)updatewas carriedout by development partners in collaboration with the Government. The MTEF Review was organized by the MinistryofPlanningandNationalDevelopment. xi suspension o f procurement activities for governance reasons during the first half o f the year. 5. Revenue collections in 2003/04, though good, are likely to be lower than the targeted 22.2 percent o f GDP. They have been on a falling trend in the recent past. The decline in revenue i s largely a result o f reduction in import duty collections, which i s likely to go further down in 2004/05 with the implementation o f a Common External Tariff (CET) negotiatedinthe Eastern African Community (EAC). 6. Total external support has been going up in the last three years; external grants and loans were K Sh 17 billion in 2001/02, which increased to about K Sh 27 billion during 2002/03. According to provisional figures, 2003/04 is expected to end up with K Sh30billion ofexternal loans and grants. 7. The level o f public expenditure has been on a general upward trend despite the government policy objective o f containing overall expenditure. The ratio of expenditure to GDP rose from 23.5 percent in 1999/2000 to 26.1 in 2002/03. The wage bill inched further up during 2003/04 to about 8.6 percent o f GDP as a result o f salary increases awarded to teachers, police, and university teaching staff. Domestic interest payments, however, went down in 2003/04, partly as a result o f prevailing low interest rates on government securities. 8. Inthe next three years, the conduct offiscal policywill remainhighlyconstrained because o f the limited scope for increasing revenues through higher tax rates and the objective o f limiting domestic borrowing. The objective o f reducing the debt to GDP ratio by about 2-3 percent per year i s unlikely to be achieved at the currently projected levels o f resources and expenditure. There are two main implications o f this tight fiscal position: (1) External concessional assistance will remain the main source for funding capital expenditure inthe near future; (2) hardbudget choices will have to be made to cut unproductive expenditure and channel more resources toward ERS priorities. 9. There are several challenges and risks inthe near future to successful achievement of aggregate fiscal objectives: (a) There i s high unpredictability and variability o f external aid. Within external aid, the program mode has more variability than the project mode. As donors move toward delivering larger volume o f aid in the program mode, the vulnerability o f the budget can be reduced by the government and aid providers working together to harmonize and streamline the practices and prerequisites for aid. (b) Off-budget parastatal liabilities come on-budget without adequate planning.The budget calculations inKenya do not take into account such liabilities o f public enterprises and local authorities for which the central government may be finally accountable and that may suddenly come on- budget with a destabilizing impact. There is a need to systematically collect information on financial operations o f public enterprises and local xii authorities with a view to remain alert to, and adequately plan for, their demands on the central budget. The government has decided to introduce national social health insurance for all Kenyans. However, there i s inadequate information on the design and implementationdetails of the proposed scheme. The fiscal cost of the scheme i s as yet unknown but i s likely to be quite large (about 1.5-2.0 percent o f GDP). It will be difficult to contain the wage billinview o f the fiture increase in teachers' salaries promisedby the government. Implementation o f the CET within the EAC Customs Union is likely to result in a fall in customs revenueby more than 1percent o f GDP in near term. The large volume o f expenditure arrears and large number o f stalled projects have been around for a long time. Whenever they are ultimately resolved, they will demand serious amounts o fresources. 10. The current medium-term fiscal framework i s not credible because it i s predicated upon optimistic increases in external assistance and revenues. It would be prudent to compute financing requirements on the relatively more conservative increase inrevenues. On external assistance, it would be better to explore what might be a plausible range o f outcomes. Given the rather bleak implications for domestic financing, some alternative scenarios-including, not least, a consideration o f how to manage a reduction in spending-is required. Compositionof GovernmentSpending 11. The structure o f government spending has not changed much in the past three years. About one-third o f the total expenditure i s made on wages and salaries o f public servants and about one-sixth on consolidated find services (interest and pension payments). After spending about 8-9 percent on defense and another 10-11 percent on transfers to semiautonomous government agencies (including universities), the residual- about 30 percent of the total expenditure-remains available to be spent on discretionary development and operations and maintenancea2 This i s not enough to support the ambitious growth strategy that hinges on a substantial increase in public spending on creating andmaintainingpublic assets and improvement in service delivery. 12. A successful implementation of the ERS requires increased resource flow to infrastructure, health, less developed regions, and targeted programs benefitingthe poor. The government cannot just depend on additional funds becoming available for meeting Not all development expenditure i s capital. All project receipts, revenue as well as appropriations-in-aid from development partners are recorded as development expenditure. Much o f these project receipts i s recurrent spending, such as o n salaries of project staff, consultancy services, and so forth. xiii these needs. Some hard budget choices will have to be made to restructure the spending pattem. Economic Composition 13. The government spending on operations and maintenance (O&M) has averaged around 18 percent o f the total expenditure in recent years. This is a slight increase from around 15 percent inthe late 1990s. The level o f spendingon operations and maintenance is possibly under-recorded, but it i s still insufficient to maintain assets and provide enough goods and services for effective delivery o f services. There i s a risk that proposed large increases in development expenditure could crowd out O&M expenditure in the coming years. The O&M expenditure should be based on norms related to the cost o f delivering services and maintaining physical assets at targeted levels. 14. Transfers to nonprofit and for-profit (commercial) statutory boards and corporations, including universities, has been going up in the past three years, and it accounted for about 12 percent (projected) o f total expenditure in 2003/04. The increase is mainly accounted for by transfers to for-profit public sector enterprises. The current transfers are not the only cost to the exchequer o f keeping these commercial enterprises. The forgone retum on government equity inthese enterprises i s a huge opportunity cost. Apart from vigorously implementing the strategy to restructure commercial enterprises, there i s a need to revisit the rationale for the existence o f a number o f semi-autonomous government agencies (SAGAS).Many of these can be mergedwith the parent ministries, privatized, or closed down in a hture reorganization. 15. The share o f development expenditure in the total expenditure had been on a long-term declining trend until 2001/02. In 2002/03, development expenditure's share went up to more than 16 percent, from 11percent in 2001/02. Most o f the increase was financed with foreign assistance and an increase in arrears. The share o f GoK-financed development expenditure has been around 5 percent o f the total inrecent years. 16. Kenya continues to have a big stock o f incomplete construction projects (207 stalled projects in 2003) on which more than K Sh 13 billion has been spent, and an equivalent amount i s needed to complete them. The government is now following a policy to drop some o f these projects and complete the rest over time. The 2004/05 budgethas proposedK Sh 1billion for resolution o f some projects. 17. Kenya's public service wage bill is large. Downsizing o f the core civil service during the 1990s has not made a dent in the wage bill because ad hoc salary and allowance adjustments, awarded to different groups o f the public service at different times, appropriated all the fiscal space created by downsizing. 18. Ad hoc salary adjustments have, however, led to a distorted and nontransparent pay structure inpublic service. Salaries o f teachers have increased much faster than those o f civil servants. The pay structure has also become decompressed. There is an overdependence on allowances. The government needs to adopt a systematic and comprehensive approach to pay adjustments that would take into account the xiv employment and wage structure o f the entire public service and also address the issues o f too many allowances and the decompressed salary structure. There i s a need to review the promisedwage enhancements to the teachers. Ministerial Composition 19. Inthe past four years, ministerial expenditure has gone upby almost 70 percent in nominal terms (which comes to about a 30 percent increase in real spending after accounting for inflation). However, the pattem o f spending (interms o f relative shares of different ministries) across various functions o f the government has not changed much, which shows that there have not been any dramatic shifts in public expenditure policy, except for the conscious effort to reduce the role o f the state inagriculture and a tendency toward growth in administrative functions. The recent partitioning o f ministries (Agricultural and Rural Development, Finance and Planning) appears to be partly responsible for the increase inadministrative expenses. 20. In the 2003/04 recurrent budget, the government signaled fairly clear prioritization, significantly increasing allocations to education (by 30 percent) and, to a lesser extent, to health and a range o f administrative functions. Sizable increases were also made ineducation and infrastructure development budgets. Aligning Spending with the ERS and Enhancing the Poverty Focus of Spending 21. The term "poverty focus" refers to equity concems o f public spending, that is, how the benefits of spending are distributed across poverty levels. InKenya, so far the poverty focus of spending has been measured largely in terms o f direct spending on programs seen as more pro-poor. Since 2000/01, the government has identified a list o f budget items as core poverty programs with a view to channel more resources to them and protect them from budget cuts during the year. In 2003-04, about K Sh 45.9 billion were allocated to these items. 22. Identification o f programs benefiting the poor and protecting them through administrative measures can be effective in improving the poverty focus o f public spending.It can also be usedas a vehicle for relating donor resources to specific program expenditures. It should, however, be seen as a transitional measure. Inthe longer run, the objective o f improving the poverty focus should be subsumed within the broader objective o f aligning spending with the ERS objectives, though an attempt could still be made to fully protect the fwnding for identified key priorities o f the ERS from budget cuts. This would complete the move from core poverty programs to core priority programs. PEM 23. An assessmento f the PEM systemcarried out in2003 found that Kenya met 3 of the 15 benchmarks for satisfactory performance. After the assessment, the government prepared an action plan for implementing the recommendations, and action has been xv initiatedinmany areas, yet the progress has been slow. Inan update carried out in 2004, PEM systems were found to meet 4 o f the 16 benchmarks, which indicates that a much more concerted effort i s needed to improve PEM inKenya. There are two key indicators of weak PEM in Kenya: a big gap between the original (printed) budget and outturn, and a huge stock o f expenditure arrears. 24. There i s an extreme disconnect between the printed estimates and actual expenditure, signaling a disconnect between policy intent (as revealed in budget allocations-the printed estimates) and actual priorities (as revealed in actual expenditure). On the whole, the recurrent budget is generally fully spent in any given year, but there i s significant variation across ministries, suggesting that a lot o f "balancing" i s done within the year. The disconnect in the development budget is far more severe and suggests that there are problems with the realism o f development budgets. Infact, the situation appears to be deteriorating over time. Underspending of the Development Budget, 1998/99 to 2002/03 (Current K Sh 1998199 199912000 2000101 2001102 2002103 Printed estimates 11,314 15,474 38,421 41,637 46,469 Actual spending 9,322 7,095 22,998 14,668 19,111 % ofprinted 82.4 45.8 59.9 35.2 41.1 25. There are many reasons for the disconnect between the printed estimates and actual expenditure. The main reasons are an unrealistic budget (inflated resource envelope), policy changes during budget execution, ease o f virements (reallocations within a Ministry's budget), and difficulties incash availability, particularly at the district level. 26. InJune 2003, the stock of expenditure arrears was about K Sh 18 billion-about 1.9 percent o f GDP, o f which development bills accounted for about K Sh 11.7 billion. Most o f the development bills are accounted for by the stalled projects. The bulk o f recurrent bills are for utilities, mainly water, telephone, and electricity. The reasons for pendingbills span the entire budget cycle and include incomplete costing and fundingo f projects, new projects during the year, a tendency to use appropriations allocated to utilities for other spending items, weak financial compliance, particularly with regulations relating to commitment control, and a weak public procurement system, including weak contract design. 27. There are notable weaknesses inPEM in Kenya that must be resolved before the budget can be wielded as an effective instrument for strategic linking o f resources and spending. The weaknesses span the entire budget cycle, but much o f the reform effort in the past few years has focused on the MTEF process. Clearly, most o f the upstream budget planningand prioritization will be o f no avail ifpriorities are changed in a major way duringbudget execution. xvi The MTEFProcess 28. To align broad spending patterns with national priorities, Kenya embarked upon an MTEFprocess in2000/01. Because o f data difficulties, it i s not possible to judge what impact the MTEF process has had as a policy tool. The best we can conclude i s that the MTEF process has only very weakly influenced the change o f direction o f expenditure, and at any giventime, there i s a significant disconnect between the budget and the MTEF ceilings. Aware o f the many weaknesses o f the MTEF process, the government recently carried out a review. 29. The government i s in the process o f devising ways to strengthen the MTEF process. The key issues are: The MTEF and annual budgeting remain two separate processes. The MTEFis seen as an adjunct process andhasnot become the annualbudget formulation process. The Planning-Finance split between the MTEF process and the annual budget is not conducive to effective integration o f spending priorities and resource allocation. There is no single budget strategy document that could provide a reasonably firm reconciliation o f resources and priorities before line-item budgeting and be approved by the political executive and senior management. There is a lack o f political engagement in budget strategy and sector ceilings determination. The budget formulation timetable is too compressed. Ceiling determination is currently a two-step process. Sector ceilings are developed first and then apportioned into ministerial ceilings. The value added by a two-stage process i s suspect. The crucial issue i s that the unit o f accountability, budgeting, and even planning or strategy formulation is a line ministry. Much o f the legitimacy being sought through Cabinet approval o f the budget strategy and ceilings process will be lost if the Cabinet approves only broader sector ceilings and not ministerial ceilings. MinisterialPERs 30. This section covers main findings coming out of education, health and other key ministries'MPERs. xvii Ministry of Education, Science and Technology 31. After the introduction o f the free primary education (FPE) policy and a sharp rise in enrollment rates, the need for education inputs such as teachers, classrooms, and materials and textbooks has risen. At present, the aggregate number o f teachers i s not too low, and the justification for hiring large numbers o f new teachers is weak. There is, however, an issue o f allocation, because there i s significant variation in pupil-teacher ratios across the country. Overall, there are three options: reallocate existing teachers, increase the teaching cadre significantly, or temporarily hire teachers locally in areas where there are shortages. The government is planning on conducting a staffing norms study to look at the various options and their costs, then quantify more precisely the long- term (as opposed to temporary) requirement for increased teacher numbers. 32. As a result of the FPE expansion, overcrowding has taken place in primary schools, but at an aggregate level, there does not appear to be a massive shortage o f classrooms. There i s little information in the ministry, however, on the need for new classrooms, unit cost estimates, or a policy on cost sharing on new construction and rehabilitation. 33. There is an underuse o f teacher training capacity. It appears that there may be scope for reducing the number o f teacher training facilities to generate cost savings. 34. There is great emerging pressure to substantially expand total secondary enrollment. It will be difficult for the government to afford this increase, given current finding sources and levels. Fundingneeds inthe near futurewill depend, inparticular, on issues related to teacher use, bursaries, the government's share o f recurrent funding, and the policy for establishing new schools and expanding existing ones. 35. There is a need for a clear strategy with respect to how secondary expansion will be managed. Apart from the obvious issues o f teachers, bursaries, and new school construction, the more fundamental question o f the "right" exit point for education requires debate. The proposed 70 percent transition rate does not appear to be anchored in a debate about the role o f secondary educationina national development strategy. 36. Technical and vocational education and training (TVET) i s considered a national priority to increase employability o f youth and buildthe country's technical capacity. Yet TVET inKenya faces many of the same difficulties as elsewhere: high unit costs (K Sh 100,000 per student annually), questionable effectiveness, inadequate curricula, lack o f integration with workplace and employer requirements, and trouble in retaining staff. In light ofthese concerns, the proposedincrease infundingfor TVET is questionable. Ministry of Health 37. A mismatch between policy and resource allocation has been a long-standing concern in health spending inKenya. Concerns are related to the high share o f spending on curative services relative to government's stated policy priorities, as well as the high and growing share o f spending consumed by the Kenyatta National and M o i Referral xviii Hospitals. However, data are not nuanced enough to understand what services are really being delivered with the money being spent. For example, tertiary level institutions and provincial and district hospitals may be delivering a substantial share o fbasic health care andprimary services. Also, there is little informationonuse o f facilities and services. 38. There are nonexpenditure factors that constrain the government's ability to spend money effectively, especially at the rural and primaryhealth care levels. These absorptive capacity constraints are related to problems o f recruitment o f staff, leakages or logistical constraints, or bottlenecks inrelease o f funds from the headquarters. A study i s needed to ascertain the extent to which these absorptive and institutional factors constrain the ability to spend effectively and what i s needed to overcome them. 39. This year's MPER has made a good start at estimating what might be requiredto deliver "adequate" health services in the longer term. Although it is difficult to estimate spending requirements because they are potentially almost infinite, it is still possible to arrive at order-of-magnitude requirements using refined unit costs o f different services and marginal cost analysis. The GoK and donors could very usefullybuild on the good work already done by refining these estimates over the coming year. Other Ministries 40. Division o f the Ministry o f Agriculture into three ministries appears to have increased administrative expenditure and made policy coordination difficult without any visible benefits. The issue needs to be analyzed carefully to catalogue the costs and benefits o fthe division. 41. The Ministry o f Roads, Public Works and Housing has a large number o f incomplete projects-the stalled projects. It i s also the ministry with the largest share o f old pendingbills. The ministryneeds to coordinate closely with the Ministry o f Finance inimplementingaphasedprogramto resolve allpendingbills andstalledprojects. The PERProcess 42. The main strengths o f the PER process in 2003/04 were the involvement o f all line ministries-some o f which were quite enthusiastic-in the process, a smooth functioning o f the PER coordination mechanism through the Technical Working Group and the Steering Committee, and the linking o f the PER process with the annual budgetingthrough the Treasury Circular inDecember. The quality o f some MPERs was muchimproved, with attention to deeper expenditure and underlying policy issues. 43. The main weaknesses o f the 2003/04 process were the late start o f the MPERs (end-December), with consequent weak quality o f many MPERs, and the structural disconnect between the planning and budgeting functions in line ministries and at the center. As a result, much o f the analysis o f MPERs did not influence the 2004/05 budget allocations. The timing o f the PER Report (early June) also worked against the process having a major impact on annual budget allocations. xix 44. IfMPERsare seen as merelyinstrument?for annually updating expenditure data (which should be done routinely as part o f budget monitoring), there i s a high risk that they will become routinized and quickly lose their value. They need to be reoriented to focus on several key issues in each sector. Following the analytical work carried out in the past two years, the stage i s now set for advancing and refining the pre-budget analysis, at least inthe eight ministries that have preparedMPERs for two years. KEY RECOMMENDATIONS OFTHE PER Developing and Implementing a CredibleMedium-Term Fiscal Strategy 45. To mitigate the risks o f and meet the challenges to the medium-term fiscal strategy, the following actions are recommended: To counter the unpredictability o f the volume of extemal aid, continuous dialogue with development partners focused on strengthening information sharing and harmonizing and streamlining the practices and prerequisites for aid should be vigorously pursued. Meanwhile, budgeting should be based on conservative estimates o f extemal aid based on past experience. When assigning development ceilings, a planning reserve could be kept unassigned, to be assigned later, close to budget finalization, based on better information about aid availability. The fiscal impact o f the health system change (National Social Health Insurance Scheme [NSHIS])i s likely to be large but has not been carehlly worked out. The decision to implement the NSHIS should be contingent upon a full understanding o f design and realization issues and the public expenditure costs. Debate on the NSHIS could be carried out in the context o f developing a fiscally sustainable health sector strategy. There i s an urgent need to systematically collect information on financial operations o fpublic enterprises and local authorities with a view to remain alert to, and adequately plan for, their demands on the central budget (contingent liabilities). This should also allow the government to monitor the consolidated public sector deficit apart from the fiscal deficit o f the central government, which i s currently monitored. There i s a need to revisit the proposedwage enhancements for teachers. 46. The medium-term fiscal framework should be revisedbased on a hardheaded look at feasible tax reforms and their likely yield. It would be prudent to compute financing requirements on the relatively more conservative increase in revenues. On extemal assistance, it would be better to explore what might be a plausible range of outcomes. Some altemative scenarios, including altemative debt reduction rates as stabilizing debt ratios, shouldbe explored. xx 47. An explicit Cabinet endorsement o f the medium-termfiscal framework andpolicy assumptions behindthe framework (as part o f the Budget Outlook Paper around October and a Budget Strategy Paper inMarch) would be essential for providing credibilityto the overall resource envelope and ministerial ceilings. More civil society engagement in budget strategy would help to build support for difficult decisions. PrioritizingExpenditure toAchieve ERS Objectives 48. The government cannot rely only on additional resources becoming available to restructure spending. Much can be done within the existing resource envelope, but that would require hard budget choices to be made. Specifically, the attempts should be focused on: (a) Curbing unproductive general administration expenditure, which has been rising over the past few years: The ongoing functional reviews should be used to identify such expenditure. These reviews, however, need to be carried out with a clear objective to cut expenditure, rather than an open- ended identification o f "core" functions without an explicit budget constraint. The government should revisit the reorganization o f ministries from an administrative cost andpolicy coordination point o f view. (b) To limit transfers to parastatals, the strategy for restructuring or privatizing for-profit parastatals should be implementedvigorously. There i s also a need to review the rationale for the existence o f many SAGAS that carry out regulatory, regional development, and social service functions. Many of these can be mergedwith parent ministries or closed down in a future reorganization. The review can be conducted as an extension o f the ongoing functional review in ministries or a stand-alone study. 49. Mucho fthe expenditure prioritization withinkey line ministries shouldtake place around core performance indicators being identified as part o f the ERS monitoring and evaluation framework. The process would involve setting targets for these core indicators, then identifying inputs in the form o f programs and interventions-the core priority programs-that are "essential" for achieving the targeted results. The identified core priority programs should then be funded sufficiently andpredictably. This process o f identifyingthe results-to-inputs chain and providing adequate funding is likely to be an iterative one. Once identified and fully accepted by all concerned parties, the core priority programs could take the place of the core poverty programs used at present to prioritize expenditure . 50. For other programs not covered by the core priority programs, the Ministry o f Finance should take the lead, jointly with line ministries, in developing O&M norms linkedto maintenance andservice delivery standards. 51. Stalled projects identified for completion should be given priority over starting new projects. Other than donor-financed projects, no new projects should be started until xxi there i s adequate financial provision for existing projects. The target should be to resolve all stalled projects by 2006/07. 52. Now that a full cycle o f wage enhancements has been completed, the government should develop a medium-term pay policy and set up accompanying institutional mechanisms to replace ad hoc processes for wage adjustments. StrengtheningPEM 53. The government is updating the Enhanced Financial Management Action Plan in the context o f the PEM-AAP update. This updated action plan will provide the overall framework for strengthening PEM. The following paragraphs recount some important areas o freform that should be given urgent attention. 54. The most effective way to integrate the MTEF and annual budget processes would beto bringthem together institutionally inthe Treasury, as well as within line ministries. 55. Key new processes that emerge from the findings of the MTEFReview are: (a) Preparationo f a Budget Outlook Paper inOctober/November (b) Early preparation o f MPERsMTEF submissions during October to December (c) Developing a Budget Strategy PaperMTEF using MPERsMTEF submissions and any other analytical inputs by March. The Budget Strategy Paper should be approved by the Cabinet and include ministerial ceilings. (d) Dissemination o f the Budget Strategy Paper to stakeholders and civil society. 56. The important elements o f the proposed process reforms are engagement o f the political executive and senior management in the following year's budget priorities early inthe budget cycle and development of a credible statement of spending priorities and ministerial ceilings that are seen as bindingby all concerned. 57. The role and functioning o f SWGs should be sharpened. Their terms o f reference should be more narrowly defined, keeping in mind that SWGs are coordination instruments and not self-standing government bodies with regular staff or a secretariat. There i s also a need to ensure systematic interface with stakeholders in SWGs outside the budget cycle. 58. Although settling old expenditure arrears (from before June 2003) should no doubt be a priority, there i s also an urgent needto implement an action planto ensure that new arrears do not accumulate. Such an action plan would include, among other things (a) announcing a transparent policy for clearing accumulated arrears and introducing the rules and regulations to prevent accumulation of fresh arrears, (b) reviewing contract xxii administration procedures, (c) establishing a project implementation and monitoring system, (d) instituting procu.rementlaws and regulations, and (e) establishing a systemfor managing pendingbills. Recommendationsfor SpecificMinistries Ministry of Education, Science and Technology 59. Drawing on the proposed staffing norms study, the options and costs o f both addressing the immediate imbalances in teacher-student ratios, as well as the long-term requirements for increased teacher numbers, should be quantified. Inthe meantime, there should be increased finding for temporary teachers to replace those that are absent because o f long-term illness and in-service teacher training for the current primary teachers. 60. An analysis should also be undertaken to examine the possibility of the decentralization of teacher recruitment, as well as management structures and incentives that facilitate a more efficient use o f teachers. The results o f this analysis are needed to address the shortage o f teachers in an optimal manner. In the meantime, increased funding is probablyjustified for temporary teachers inareas where there are excessively high student-teacher-ratios and in-service training for current teachers to adjust to the new curriculum. 61. The secondary education strategy, including policies on (a) bursaries, (b) the establishment of new schools, and (c) the appropriate proportion o f primary students passingon to secondary education, should be developed. 62. The following should be developed for the university subsector: (a) a strategy for expansion and public financing, (b) expenditure projections that take into account the proposed salary increases for university lecturers, and (c) identification o f effective means o fproviding incentives to universities to respond to national priorities. 63. The plans to significantly increase finding for TVET should be suspendeduntil a thorough study o f effectiveness o f spending i s carried out and options for greater on-the- job, employer-financed, andprivate sector training are explored. Ministry of Health 64. Immediate recommendations for expenditure allocations inthe coming budget are the same as those made in previous years: to increase the allocations for spending on basic health service and contain spending on the two central hospitals. But, according to current evidence, there i s little basis for determiningthe optimal levels o f spending. 65. The PER work o f the past two years has not been sufficient to allow definitive decisions on reallocation of expenditure in the health sector. In the coming year, the following program o fwork shouldbe undertaken: xxiii A more detailed analysis o f actual spending compared with budgets at a disaggregated level by line itemwithin eachprogram A specialized study to determine what services are actually being delivered at each level and type o f facility (health posts, provincial and district hospitals, and the Kenyatta National and M o i Referral Hospitals), including an assessment o f time use and expenditure within the facilities to estimate the real distribution o f current spending by level o fcare An analysis o f use o f facilities and the reasons for utilization andbypassing behavior An assessment o f absorptive capacity constraints related to drugs, staffing, and flow o f funds to identify whether it makes sense to increase financing and the barriers that need to be overcome to do so (this could draw upon existing or ongoing work within the health sector) A true expenditure tracking survey that traces the flow o f specific "buckets" o f funds (and in the case o f drugs and medical supplies, commodities) through the system, to determine where leakage i s taking place and its magnitude Analysis o f the impact o f health fees on exclusion o f the poor and possible reforms to compensate for this A risk analysis o fthe proposedhealthinsurance scheme. 66. It is not realistic to expect this work to be done as part o f a PER exercise conducted over a fairly short time in the runup to the budget. Work should start early as part o f a broader consolidated effort geared toward health sector budget reform, which can lay the basis for both expenditure reform and a common program o f donor support for increased health sector spending. Ifthis i s done well, by the time o f the next PER, the government should be in a position to make informed decisions regarding (a) revised health funding levels for the next budget, (b) proposed reallocations, and (c) complementary measures needed-for example, in institutional reforms, staffing and incentives, systemic or management changes, and so forth. Recommendationsfor the PER Process in2004/05 67. The following recommendations are made to strengthen the PER process and its elements in2004105: (a) The terms of reference for MPERs should have both generic elements and customized elements. Examples o f generic elements include (a) analysis o f updated and reconciled expenditure data at the subvote level, explaining any significant changes in expenditure trends; (b) reporting on xxiv performance indicators; (c) identifyingpolicy changes, new programs and projects, and their expenditure implications in the medium term and for the next budget year; and (d) reporting on the functional reviews and identifyingthe scope for savings over the medium term. The customized elements would be specific to each ministry and would be agreed upon between the PER Secretariat and the ministry at the beginning o f the process in Se~tember.~Ways should be found to merge MTEF submissions with MPERs. Line ministries should not have to make two submissions on expenditure strategy for one budget. (b) The calendar o f the PER process should be completely harmonized with the annual budget and MTEF. The PER Report shouldbe preparedin end- March or early April so that there can be a good discussion o f its findings among all stakeholders. (c) Action on PER recommendations should be monitored through the Steering Committee. Short- and medium-term action plans could be developed. (d) Jointly with line ministries, the Budget Monitoring Division should develop a historical database on expenditures by subvote. This will allow consistent analysis o f expenditures, notwithstanding several reorganizations o f ministries. Line ministries should use the same data for their own analysis inMPERs. (e) Apart from other issues, the next PER could focus generally on the issues o f efficiency o f public service delivery, the performance orientation o f budgeting (how it can be made more results-based), the extent o f contingent liabilities, the costs andbenefits o f reorganization o f ministries, and a systematic review of the main findings o f ongoing functional reviews inkey ministries. Of course some ministries that serve mainly administrative functions may not have any customized elements. xxv 1. INTRODUCTION 1.1 This i s the third Public Expenditure Review (PER) undertaken by the government o f Kenya (GoK) in the last seven years, and it builds on the 1997 and 2003 reviews.' The 1997 PER provided an assessmento f the likely contribution o f the trends in public expenditure management to the stated objectives o f high private sector-led growth and poverty reduction. The report concluded that "present trends in public expenditure management are fundamentally inconsistent with the objectives o f achieving high and sustained growth o f the economy and reducing the levels o f poverty." The composition o f public expenditure was inappropriate and inefficient and could not arrest the continuing erosion o f the public sector asset base (for example, the poor condition o f the infrastructure). The report also decried the growth of informal fiscal instruments such as pending bills and excess issues (unplanned expenditures not related to natural disasters). The 1997 report recommended the production o f annual editions o f the PER to follow up on implementation o f key recommendationsand to study additional topics on a rolling basis. 1.2 When the new government came to power inJanuary 2003, it embarked on the preparation o f new economic blueprint aimed at restoring economic growth, generating employment opportunities, and reducing poverty levels. The blueprint, titled "Economic Recovery Strategy for Wealth and Employment Creation (ERS)," identifiedthree core objectives o f the fiscal strategy over the period from 2003/04 to 2006/07: Fiscal sustainability, expenditure restructuring for growth and poverty reduction, and improving public service delivery. It also elaborated various measures that the GoK will undertaketo strengthen the overall management o fpublic finances.2 1.3 Underthe fiscal sustainabilityobjective, the fiscal policy's aim is to maintain a level o f expenditures that can be funded without either an increase inthe net present value (NPV) o f overall debt to gross domestic product (GDP) or an increase o f external debt growth. This i s consistent with the country's desire to avoid falling into heavily indebtedpoor countries (HIPC) category. 1.4 Regardingthe expenditure restructuring for growth and poverty reduction objective, the ERS calls for increased development expenditures, especially those targeting government investments, core social expenditure (education and health), and core poverty expenditures. Although concessional external assistance i s expected to The 1997PERreportwas not publishedbythe GOK or the World Bank.The 2003 reportwas publishedbythe GOK. '.As an integral component of the ERS, the government planned to undertake public sector reforms whose principal objectives would be to create a smaller, affordable, and effective public service; achievegreater transparency inthe use of publicresources;and create amore productivepublic service that concentrates public finances and human resources on the delivery of core government services. The components of the public sector reform include civil service reform, financial planning and budgeting, parastatal reform and privatization, local authorities reforms, and enhancing integrity and accountability. 1 provide a significant part o f the margin for increased development spending, much o f the increase will have to come from within the GoK's own resources through savings and shifting resources from lower-priority to higher-priority areas 1.5 The objective o f improving public sector service delivery will entail enhancing boththe efficiency andthe effectiveness o fpublic expenditure. 1.6 The government recognizes that achieving the above-stated objectives and improving the overall management of public finances i s going to be a challenging long-term goal. It has therefore decided to carry out PER work regularly as one o f the main instruments to inform public expenditure restructuring and to evaluate budget performance and progress of public expenditure management (PEM) r e f ~ r m s In . ~ a circular dated December 3, 2003, the Treasury extended the preparation o f Ministerial PERs (MPERs) to all ministries (increased from eight key ministries inthe 2003 PER process). 1.7 For benchmarking PEM reform, the GoK i s using the Public Expenditure Management Assessment and Action Plan (PEM-AAP), an exercise carried out in close collaboration with development partners. The first exercise was carried out in 2003, and an update was completed inJune 2004. OBJECTIVES OF THE PERWORK 1.8 Objectives o f the PER work are directly relatedto implementation o f the ERS: (a) To provide regular ex-post analysis o f ministerial expenditures, commenting on the composition, efficiency, and effectiveness o f spending in meeting service delivery targets and other performance indicators. The key idea i s that expenditure analysis will enable policymakers to ask the right set o f questions when making decisions onpublic spending. (b) To help strengthen the medium-term perspective inbudget making by costing medium-term resource implications o f policy choices. (c) To help strengthen PEM by regular monitoring o f and reporting on reform progress. This will ensure that scarce resources are used to achieve priority goals. (d) To help improve the performance focus ,of spending through monitoringindicators o f ERS progress. (e) To promote broad participation in budget-making process. The PER will open up the budget system to public scrutiny by publishing information on budget formulation, budget execution, and public accounts. This i s expected to help improve the quality o f public debate and the accountability o f public agencies. Participation in the budget process can ensure an active role for citizens indecisionmaking. The PERprocess i s evolving inKenya, and at this stage, it is seen as an annual process synchronized with the annual budget-makingcycle. 2 1.9 The process o f annual PER work i s also expected to have several spinoffs. It i s expected to improve capacity in line ministries for analysis o f expenditures and costing o f programs and activities, as well as improve dialogue between the Ministry o f Finance and line ministries on budget-related issues. It i s important that the PER work i s seen as an instrument for regular dialogue among the GoK, donors, and other stakeholders on public expenditure management issues, which would, in turn, allow donors to better orient their support to the needs o f the government. RECOMMENDATIONSOF THE 2003 PER 1.10 The 2003 PER focused on three main areas: (1) to review the recommendations o f the 1997 PER and other works undertaken to improve expenditure management and the extent to which the recommendations were implemented, (2) to take stock o f the trend o f the broad fiscal policy framework for the period from the 1997/98 to 2001/02 fiscal years with emphasis on composition of expenditure and the budget planning and management process, and (3) to develop a better understanding o f budget efficiency and impact on quality o f service delivery. The 2003 PER was based on MPERs covering eight ministries, which were selected on the basis o f their potential role inpoverty reduction and their shares inthe national budget. 1.11 The 2003 PER noted that many o f the problems in the structure o f public expenditure observed in 1997 persisted. These included a relatively high share o f wages and salaries and relatively low operations and maintenance (O&M), relatively low development expenditure, high transfers to organizations outside the main civil service, and weak budget implementation, which resulted ingeneral underspendingon the development budget and overspending on the recurrent budget (especially in ministries primarily engaged in administrationrather than service delivery). 1.12 The 2003 PER recommended improvement in functional classification o f expenditure and rationalization o f ministerial functions with a view to eliminating duplication o f functions and divesting services that can more efficiently be provided by the private sector. The 2003 PER recommended improved analysis of subsidies and transfers to assess their contribution to service delivery. There are also deficiencies in the system o f functional and economic classification o f expenditure and lack o f consistency in the classification system across ministries that limit the usefulness o f the analysis undertaken. ORGANIZATIONOF THE PERWORK IN2003/04 1.13 The PER work comprised two phases: (1) the first phase directly informed budget strategy through MPERs and background studies on important budget policy issues that needed deeper analysis and information collection than i s the routine and (2) the second phase evaluates budget strategy and progress in reforming budgetary institutions inthe past year and makes recommendations for further reform. 1.14 This report, Public Expenditure Review 2004, i s the final product o f the PER work program carried out during the financial year 2003/04. As part o f the work program, 27 MPERs were produced. The MPERs were used for preparation o f Sector Reports duringthe medium-term expenditure framework (MTEF) budget process. The 3 PER report pulls together the evaluation and relevant analytical work carried out in phase one and develops medium-term recommendations 1.15 The report evaluates budget performance in 2002/03 in terms of aggregate fiscal objectives and the composition o f public spending. It then evaluates the performance o f PEM institutions and progress made on reforms. Spending issues in key ministries are discussed, highlighting areas for further analysis to better inform budgeting. The report does not cover performance orientation of the public spending, which i s likely to be the focus of future work. 1.16 The PER process i s likely to evolve over time. Most important, it will need to be harmonized with the process of ERS monitoring and the MTEF process. As the government puts together the institutional ftamework for ERS monitoring, it i s expected that the PER process and ERS monitoring will be organized to avoid any overlap and duplication o f monitoring effort. 4 2. AGGREGATE FISCALPERFORMANCE 2.1 The ERS targets the goal o f economic growth with macroeconomic stability through policies supporting low inflation, low fiscal imbalances, declining net domestic borrowing, and healthy balance o f payments. The major policy objectives include increased domestic savings and investment, improved accountability in the use o f public resources, and restructuring and refocusing public spending toward priority activities. The budget i s one o f the most fundamental instruments o f implementing economic reforms and other policy options outlined in the ERS. The ability to have a stable and sustainable macroeconomic framework i s essential to the budget. 2.2 The ERS proposes to achieve the fiscal sustainability objective by reducing the budget deficit through containing increase in overall expenditures, sustaining the revenue to GDP ratio at 21 percent or more, and securing external concessional financing in place o f the costly and distortionary domestic financing. The ERS also develops a fiscal strategy to achieve these objectives. The strategy is expected to serve as the basis for defining a realistic medium-term government finance framework covering revenues, expenditures, and financing, which would allow for an aggregate expenditure ceiling consistent with the objectives indicated above. The strategy i s also expected to allow for the setting o f expenditure ceilings by both economic category and MTEF sectors. 2.3 This chapter describes recent trends in government revenue, expenditure, budget deficit, debt, and deficit financing. It then identifies the issues and risks relating to fiscal sustainability inthe medium term. Issues o f expenditure restructuring and public expenditure management are discussed inlater chapters. RECENT MACROECONOMIC PERFORMANCE 2.4 The general performance o f key macroeconomic indicators in 2002 and the estimated performance in2003 are providedinstatistical annex table A.1. 2.5 Growthhas beenpicking up very slowly. The economy has grown marginally, with real GDP rising from 1.2 percent in 2002 to 1.8 percent in 2003. Although agriculture retains its dominance in the economy, its recovery remained subdued. Its growth declined from 1.1 percent in2002 to 0.7 percent in2003 because o f depressed world commodity prices, especially for coffee, tea, and horticulture. However, the manufacturing sector recorded an impressive upturn and grew by about 5 percent in 2003, up from 1.6 percent in 2002. This was mainly the result o f the opportunities made available by the African Growth and Opportunity Act (AGOA) and Common Market for Eastern and Southern Africa (COMESA) initiatives. The services sector grew marginally, by 1.8 percent in2003 from 1.4 percent in2002, despite the adverse publicity brought about by the travel advisories on tourism issued by major source countries. Although the building and construction sector was earlier billed as the 5 ultimate trigger source o f the recovery process, most o f the grand construction projects (infrastructure-roads, housing, and energy) were not realized. The earlier projected growth o f 16.7 percent fell to 9.4 percent in2003. 2.6 The pickup in private investment has also been slow. Despite the poor business environment prevailing in 2002 (election-related uncertainties), gross investments as a percentage o f GDP grew from 13.1 percent in 2002 to 16.5 percent in2003. Savings as a percentage of GDP also rose from 13.6 percent to 14.9 percent over the same period. 2.7 Inthe external sector, the volume of exports and imports rose significantly. From 2002 to 2003, exports grew from 5 percent to 15.8 percent while imports grew from 16.7 percent to 13 percent. However, invalue terms, the quantitative growth was undermined by significant declines in both exports and import prices. Export prices declined by 9.3 percent, and import prices declined by 5.4 percent. 2.8 The rate of inflation i s estimated to have risen significantly-to 9.8 percent in 2003 from 2.0 percent in2002-well outside the lower-bound target o f 5 percent. The underlying inflation over the same period was below 5 percent. The exchange rate remained stable. General interest rates were on a declining trend during 2003 because o f declining government borrowing. The 91-day treasury bill rate declined from 8.6 percent in December 2002 to an estimated 3.6 percent by December 2003.' Consequently, commercial lending and deposit rates, which track the treasury bill rate, have followed suit. RECENT FISCAL PERFORMANCE AND TRENDS 2.9 The goal o f a stable fiscal framework has not been achievable inthe past as a result o f both domestic and foreign factors. Domestically, the government has not been able to control recurrent expenditure even though it realized satisfactory revenue collections. Externally, the external resource expectation built into the fiscal stance has not been forthcoming. The GoK remained out o f the International Monetary Fund (IMF) program between 1996 and 2003, and this adversely affected the net financial inflows. However, aggregate fiscal performance during 2003-04 i s expected to be much better than in the previous year because o f higher donor inflows, lower total expenditure, and better revenueperformance. Revenue 2.10 The targeted revenue to GDP ratio o f 22.2 percent i s unlikely to be met during 2003/04. Though income tax collection in 2003/04 i s expected to exceed the target in nominal terms and improve upon last year's collection by more than 10 percent, in GDP terms, it i s likely to stay at 6.8 percent, the same level as in the previous year. The overall revenue performance would be in line with the falling trend inthe recent past: it declined from 23.4 percent in 1999/2000 to 20.7 percent in 2002/03 and i s expected to remain around 21 percent in 2003/04 (table 2.1). The decline in revenue ' The fall of the Treasury bill rate, and other interest rates in general in 2003104, was promptedby a government decision in the 2003/04 budget to lower the cash ratio requirement from 10 percent to 6 percent. Beforethis decision, banks were requiredto holdonaverage 10percent, with a daily minimum o f 8 percent. 6 has been occasioned mainly by the reduction in import duty collection resulting from liberalization, which has gone down from 3.7 percent o f GDP in 1999/2000 to 1.8 percent o f GDP in 2002/03. It i s expected to go down further, by about half a GDP point, in the next fiscal year with the implementation o f the common external tariff under the East African Community (EAC) Customs Union. The share of the valued added tax (VAT) i s expected to go down, largely as a result o f a reduction in tax rates from 18 percentto 16 percent. Table 2.1. Revenue and Its Key Components(% of GDP), 1999/2000 to 2003/04 199912000 2000101 2001102 2002103 2003104 provisional Total revenue 23.4 22.9 21.4 20.7 21.2 Incometax 7.1 6.6 6.4 6.9 6.8 Import duty (net) 3.7 3.4 2.3 1.8 1.9 Excise duty 3.7 3.4 3.5 3.5 3.7 VAT 5.3 6.0 5.5 5.5 5.1 Investmentincome 0.0 0.3 0.1 0.1 0.4 Other 1.4 1.5 1.6 1.2 1.4 Appropriation in aid 2.1 1.7 2.0 1.6 1.9 ( A W Source: Ministry of Finance(MoF). Expenditure 2.11 The level o fpublic expenditure has been on a general upward trend despite the government policy objective o f containing overall expenditure. The ratio o f expenditure to GDP rose from 23.5 percent in 1999/2000 to 26.1 percent in 2002/03 (table 2.2). The fastest growing component o f expenditures was pension expenditures, which grew at an annual average o f 14 percent over this period. The main contributor to the rapidexpansion o f expenditures in2002/03 was the introduction o f free primary education. Overall expenditure during 2003/04 is expected to be about 25.5 percent o f GDP, against a projected 27.8 percent. Suspension o f procurement activities in the first half of the year was one o f the mainreasons for lower than planned expenditure. 2.12 Recurrent expenditure has been declining from a high o f about 24 percent of GDP in 2000/01. It i s expected to be around 21.7 percent o f GDP at the end o f 2003/04. Wages and benefits to civil service account for about 8.6 percent o f GDP and have been inching upward; they went up by about 0.2 percent o f GDP during fiscal 2003/04 because o f salary increases for teachers, police, and university professors. Transfers to semiautonomous government agencies are expected to account for about 1.8 percent o f GDP. 2.13 Interest payments during 2003/04 are expected to be about K Sh 4 billion lower than in2002/03. Lower domestic interest payments partly reflect the success in restructuring government debt toward longer-maturity treasury bonds and also prevailing low interest rates on government securities. 2.14 In GDP terms, development expenditure in 2003/04 is expected to be more than 20 percent lower than targeted and also lower than in 2002/03. This i s largely because of a shortfall o f about K Sh 10 billion in foreign-financed development spending, partly attributable to suspension o f procurement o f goods and services. The 7 GoK's own development spending has been increasing over the past few years and now stands at about 1percent o f GDP, but external financing has remained subdued. The net result has been a rather low volume o f development spending, at a little more than 4 percent o f GDP. The ERS target i s to raise the level o f development expenditure to 6.7 percent by 2006/07. Table 2.2. Expenditureand Its Key Components (% of GDP), 1999/2000- 2003/04 1999/2000 2000/01 2001/02 2002103 2003104 provision a1 Total expenditure 23.5 27.7 24.7 26.1 25.5 Recurrent expenditure 20.3 23.7 21.8 21.7 21.5 Consolidatedfund services 4.5 4.4 4.3 4.5 4.1 Domestic interest 2.7 2.3 2.1 2.2 2.3 Other (commission to CBK) 0.0 0.5 0.5 0.5 0.0 Foreign interestdue 1.1 0.9 0.7 0.8 0.6 Pensions, etc. 0.7 0.7 1.o 0.9 1.3 Net issues 14.1 15.3 15.3 15.5 15.5 Wages and salaries 8.6 8.1 8.4 8.4 8.6 National security 1.6 2.0 2.2 2.1 1.8 Other 4.0 5.2 4.8 4.9 5.0 AIA 1.9 1.8 2.0 1.8 1.9 Changeinpending bills -0.2 0.2 0.2 0.0 0.0 Development and net lending 3.2 4.0 2.7 4.3 4.1 Net issues 0.9 1.1 1.3 1.7 2.0 O fwhich GoK 0.4 0.4 0.8 1.1 1.1 O fwhich loans 0.4 0.5 0.3 0.2 0.3 O fwhich grants 0.1 0.2 0.1 0.4 0.5 AIA 2.1 1.4 1.6 1.8 1.7 O fwhich loans 1.4 0.6 0.8 0.5 0.5 O fwhich grants 0.5 0.5 0.5 1.o 0.9 O fwhich local AIA 0.2 0.3 0.3 0.2 0.3 Guaranteedloans: total 0.2 0.5 0.2 0.2 0.1 Change inpending bills -0.1 0.4 -0.3 0.7 0.0 Other 0.0 0.7 0.0 0.0 0.2 ~~~~~ CBK Central Bank of Kenya. - Source: MoF and M F DeficitFinancingand PublicDebt 2.15 The GoK's fiscal strategy i s to use the higher level o f net external borrowing to finance the deficit and net reduction o f domestic debt by 2005/06. This was considered prudent to allow for increased credit flow to the private sector. The anticipated primary surpluses were to be consistent with the prerequisite debt policy o f ensuring sustainable external debt to GDP ratio while reducing the domestic debt burden. 2.16 The year 2003/04 i s expected to end with a primary budget surplus o f 0.4 percent of GDP and an overall deficit after grants o f about 2.5 percent o f GDP- lower than the targeted 3.3 percent (table 2.3). As discussed before, the main reasons for the lower deficit are lower expenditure, higher revenue collections, and higher inflow o f grants. 8 Table 2.3. Deficit and Financing(K Sh Million), 1999/2000-2003/04 1999/2000 2000/01 2001/02 2002/03 2003104 Drovisional Deficit before grants (commitment basis) -858 -40,308 -31,212 -55,197 -48,987 DeJicit before grants/GDP (%) -0.1 -4.8 -3.4 -5.4 -4.4 Overall deficit incl. grants (commitment basis) 3,557 -16,228 -24,389 -40,255 -28,156 Overall deJicit/GDP(%) 0.5 -1.9 -2.6 -4.0 -2.5 Net foreign financing -14,227 12,489 -1 1,250 -10,340 -1,383 As % of GDP -1.8 1.5 -1.2 -1.0 -0.1 Net domestic financing 11,876 624 39,704 46,922 31,521 As % of GDP 1.5 0.1 4.3 4.6 2.8 Net domesticdebt/GDP (%) 21.2 19.6 21.9 24.3 24.9 Source: MoF and IMF. 2.17 The smaller than projected financing gap, which was helped by higher extemal program disbursements than planned, i s expected to result in lower domestic borrowing, at 2.8 percent of GDP in 2003/04, than the projected 3.3 percent. As a result, the stock of domestic debt, while expected to go up to about 30 percent of GDP, will be lower than projected. 2.18 Table 2.4 shows the details of external concessional assistance, including grants, duringthe 1999/2000-2002/03 fiscal years. Over 1999/2000-2002/03 the GoK received disbursements amounting to K Sh 52.4 billioningrants (including K Sh 12.4 billion o f drought-related grants in the 2000/01 fiscal year) and K Sh 48.4 billion in loans (including K Sh 4.4 billion in drought-related loans and K Sh 4.0 billion in program loans). Donor inflows into the budget totaled K Sh 100.9 billion. However, over the same period, the country repaid extemal principal of K Sh 90.1 billion, implyingthat net inflows over the periodtotaled K Sh 10.9 billionexcluding drought- related inflows that were negativee2The low level of donor support was primarily driven by Kenyabeing "off track" with key development partners. Table 2.4. External Loans and Grants, 1999/2000-2002/03 (K Sh Million) 199912000 2000/01 2001102 2002/03 provisional Program grants - - 4,247 5,955 1,473 458 Project grants (cash) 938 1,521 1,090 3,804 Project grants (AIA) 3,309 4,160 4,260 10,777 Project loans (AIA) 6,020 5,323 7,133 5,276 Program loans 0 4,045 0 4,045 Project cash loans 2,830 4,337 2,898 2,191 Total grants 8,494 11,636 6,823 15,039 Total loans 8,850 13,705 10,031 11,512 Drought-related grants 0 12,444 0 0 Drought-related loans 0 4,382 0 0 Total externalsupport 17,344 42,167 16,854 26,551 Source: MoF. * The level of net external resources accessed by Kenya on a per capita basis was as low as US$1 over the period. 9 Table 2.5. Variability of Budget Resources, 1994/95-2002/03 Source Coefficientof variation Domestic revenues 0.12 Income tax 0.17 Import duty (net) 0.28 Excise duty 0.11 VAT 0.05 Investment income 1.12 Other 0.41 AIA 0.36 Aid 0.33 Project grants 0.47 Program grants 1.21 Project loans 0.3 1 Program loans 1.39 Source: Calculationsbasedondata from MoF 2.19 Donor aid remains considerably more variable than domestic revenues. Table 2.5 shows the variability o f various budget resources over a period of nine years. On average, external aid i s about three times more variable than domestic revenue^.^ Within external aid, the program mode has more variability than the project mode. As donors move toward delivering a greater volume of aid inthe program mode, the vulnerability o f the budget to aid i s likely to go up unless the GoK and aid providers work together to harmonize and streamline the practices and prerequisites for aid. At the same time, the GoK needs to improve absorptive capacity for aid (at this time, about 60 percent o f commitments go undisbursed). 2.20 Table 2.6 shows the composition of public debt and how it has evolved inthe last five years. Over these years, the stock o f domestic debt as a share o f the total debt had been growing as the government depended more and more on domestic borrowing to fund its operations. As part of the fiscal strategy Ithe ERS, the GoK i s partially substituting domestic debt with external concessional loans and grants. The GoK has also innovatively and substantially lengthened the maturity o f government debt over the last three years. For example, between July 2003 and February 2004, the stock o f treasury bills decreased from K Sh 78.7 billion to K Sh 67.7 billionwhile the stock o f treasury bonds increased from K Sh 161.5 billion to K Sh 183.0 billion. Although longer-dated debts incurs higher interest payments than the shorter-dated, a shift inthe maturity structure has tended to pushthe whole structure of interest rates down, so the overall cost o f government finance has fallen. Table 2.6. Public Debt (K Sh Billion), 2000-04 June 2000 June 2001 June 2002 June 2003 Feb. 2004 provisional Domesticdebt 206 212 236 289 301 (as YOof GDP) 27.7 25.2 25.5 28.4 26.9 Treasury bills 131 138 118 106 105 Treasury bonds 37 45 106 162 183 Governmentstocks 3 2 2 1 1 Within the domesticresources, the VAT remains the least, andimport duty the most, variable. 10 June 2000 June 2001 June 2002 June 2003 Feb. 2004 provisional Overdrawadvances 15 13 9 11 12 Non-interest-bearing debt 20 16 1 10 0 Externaldebta 403 394 378 400 407 (as % of GDP) 54.2 46.9 40.8 39.3 36.4 Bilateral 146 132 130 143 152 Multilateral 224 229 223 226 228 Commercial banks 32 29 24 27 26 Exportcredit 1 4 1 4 2 Totalpublic debt 609 609 614 689 708 (as YOof GDP) 82 12.5 66.3 67.1 63.3 a. Includes IMF loans. Source: C B K and MoF 2.21 According to a recent debt sustainability exercise carried out by the IMF, the NPV o f the external debt to exports ratio was about 109 at the end o f 2003. Inview o f higher inflows, it i s expected to go up, but the ratio is still expected to remain relatively moderate. The ratio o f debt service to exports i s expected to decline steadily (from 11percent in2003) inthe mediumterm. FISCAL POLICY OPTIONS INTHE MEDIUM TERM 2.22 The medium-term fiscal objectives stress the need to (1) stabilize falling revenues, possibly raising them somewhat; (2) reduce total expenditure while achieving the radical shift in composition spelled out in the ERS; and (3) reduce the domestic deficit. The overall deficit i s targeted to decline in the medium term to below 3 percent o f GDP, while net domestic borrowing i s forecast for rapidreduction and ultimate elimination. To reinforce the growth targets set in the ERS, expenditure restructuring will seek to increase the shares o f investment and human resource development as priorities. This all amounts to a major transition against the background o f recent performance. 2.23 Several constraints limit the set o f options for achieving these objectives: In the short run, it appears that there is limited scope for raising revenues through increasing tax rates. At 21 percent o f GDP, Kenya's tax burden i s already high, especially on taxable items o f importance to investment incentives (such as the corporate tax at 35 percent). Further, in the short run, implementation o f the common external tariff i s likely to reduce profitability o fmanufacturing enterprises, thereby reducing the possibility o fraising taxes. Borrowing from the domestic market will need to be limited for several reasons. First, reducing domestic govemment debt i s one o f the fiscal targets. It might be argued that given the current absence of symptoms of excess demand in the market for domestic credit, rapid reduction in domestic government debt i s not urgent. However, that does not mean that reduction of this debt should not be a priority at all. By the standards of other low-income countries, Kenya has a high domestic debt level. Ifthe current growth strategy i s successful, it will involve a substantial increase in private investment, and part, at least, o f this increase will need to be domestically financed. Therefore, the government will need to reduce its current call on domestic savings. Second, high levels o f domestic borrowing tend to drive the domestic interest 11 rate up. And third, the total budgetary cost o f domestic borrowing interms o f current and prospective interest burdenwill be higher than the other option of concessional externalborrowing. Itwill be difficult to reduce total expenditure becausemuch ofthe expenditure remains committed to the consolidated fund services (CFS) and wage bill. The government will need to fight the temptation to reduce development expenditure because the growth strategy hinges on a substantial increase in public investment ininfrastructure and humanresource development. 2.24 These constraints imply that external concessional borrowing i s likely to be the main instrument for achieving fiscal objectives. But external concessional borrowing, though appealing, remains subject to many conditionalities and low predictability. As discussed, this i s an area where govemment and development partners needto work together to harmonize and streamline aid practices. At the same time, the GoK needs to look into how absorptive capacity ofaid couldbe improved. 2.25 Inthe absence of sufficient margin to reallocate resources and a rather rigid expenditure structure of line ministries, the objective o f restructuring expenditure with a view to move resources to higher-priority areas will also remain a difficult process. At the macro level, it would require additional resources that become available to be channeled primarily to ERS priority sectors, as well as require line ministries to give sufficient attention to restructuringexpenditure within the ministry and focus it more on priority areas. Debt Reduction Targets and the Deficit: Some Projections 2.26 Annex 1 has a brief technical note that discusses the relationship between debt reduction and deficit. It shows that at the current levels of average4 revenue (20.3 percent), expenditure (24.7 percent), real growth (3.8 percent), seigniorage (0.6 percent), expenditure on interest (2.2 percent), and grants (1.9 percent), the average primary deficit in the next three years would imply a rise in the debt ratio o f about 2 percent per year. This would be an unsatisfactory outcome. The note goes on to show that to achieve the outcome o f a 3 percent per year reduction in the debt ratio, three accommodations (or a combination o f them) could be considered: (1) A rise in grant or other concessional finance-the scale o f increase in grants would have to be equal to 2.6 percent o f GDP (roughly a 150 percent increase on current projections). (2) Raise domestic revenue-domestic revenue would have to be raised to 22.9 percent o f GDP to meet the debt reduction target. (3) Reduction inexpenditure--the expenditure would need to be cut down to 22.1 percent of GDP. All o f these accommodations are severe, particularly because they are the requiredaverage outturns over the next three years. O f course, in practice, the three options may be combined in various proportions, reducing the severity o f each taken separately. Alternatively, the government could consider relaxing the debt reduction target to stabilization o f debt ratios. But even so, there appears to be a major challenge to the design o f fiscal strategy inKenya. The level ofnet externalresources accessedby Kenya on a per capitabasiswas as low as US$1over the period. 12 Challengesand Risks 2.27 External resources: The thrust o f the current fiscal strategy is to optimize use of external resource inflows anticipated from resumption of development assistance under the Poverty Reduction and Growth Facility (PRGF) and pledges made at the November 2003 Donor Consultative Group meeting. Despite the optimism generated by renewed multilateral cooperation, the main challenges to full use o f the new opportunities lies with the GoK's capacity to accelerate the implementation o f requisite reforms (conditionalities). Some o f the key reform measures are not only expensive (such as retrenchment and restructuring o f parastatals) but also have important political dimensions (cost sharing). Although the GoK has made significant breakthroughs in improving governance in a few sectors (the judiciary, land allocations, and publication o f relevant bills), a lot more i s yet to be done. It i s anticipated that the successful conclusion of the constitutional review process will be a milestone inconsolidating the primary elements o f governance and democratization. 2.28 Revenue implications of the EAC Customs Union: According to a study recently carried out by the World Bank, implementation o f the common external tariff (CET) i s likely to result ina loss o f customs revenue to the extent o f K Sh 11billion (more than 1percent o f GDP). 2.29 Contingent liabilities: The budget calculations in Kenya, and especially those involving deficits and debts, are confined to central government operations because information i s typically lacking on both operations o f local authorities and public enterprises. However, the central government i s finally accountable for what happens in these two sectors. When their finances are out of control, there is effectively an additional "quasi-fiscal" deficit runningparallel to the official budget deficit, with an associated buildup o f implicit debt. Incidents-such as the closing of a public enterprise, the impending collapse o f a state-owned bank, or a strike by employees who are not being paid by a local authority-then bring these off-budget items sharply on-budget, sometimes with a devastating destabilizing impact. 2.30 There i s a need for installation o f effective systems o f ongoing reporting and scrutiny o f local authorities and public enterprises. The first requirement i s detailed investigation to ascertain the likely extent o f the problem and its likely profile over time. When reasonably reliable estimates are available, even if only on a "ball-park" basis, then they can be built into the budget and MTEF forward projections as part of the call on government resources. The crucial points are that in many cases, these liabilities must be brought on budget, and when that happens, it either leads to reductions ingovernment spending or raises the measure o f government debt. 2.3 1 National social health insurance: The GoK is actively considering installing a national social health insurance system in Kenya. The system will extend the benefits of health insurance to the entire population. The fiscal impact o f the system change i s likely to be huge, though it has not yet been carefully calculated while design and implementation details are pending. 2.32 Dificulties in reducing the wage bill: With about 80 percent o f the fiscal budget constituting fixed commitments (consolidated fund services, wages and salaries), the options and room for restructuring are likely to remain restricted. 13 Reducing the wage bill is likely to be politically difficult, particularly given the government's commitment to create 500,000 new jobs every year. 2.33 Optimistic economic growth rate: The GoK's recovery agenda targets an average economic growth rate of 4.7 percent over the ERS period. The massive investment resources are premised on Kenya securing substantial external resource inflows in the form o f concessional financing and foreign direct investment (FDI) as well as a substantial increase inprivate investment.Iti s expected that higher levels o f public investment will likely crowd inprivate investment.Given Kenya's low savings potential (the result o f low incomes), the first challenge for effective domestic resource mobilization i s anchored in a successful financial sector restructuring program. 2.34 Pension bill: An emerging fiscal challenge is the size o f the pension bill, currently taking about 4.7 percent o f the total government expenditure and growing at a rate o f 14 percent per year. An immediate partial option proposed for the next budget i s to introduce a contributory pension scheme for all civil servants. However, the transition from the current scheme i s dependant on transfers to a self-sustaining fund, which may take about 30 years, implying protracted dependence on the exchequer. 2.35 The debt burden: Although Kenya does not fall within the HIPC Initiative, it still seeks debt cancellation on bilateral basis. As Kenya successfully implements the strategy to substitute domestic debt with external debt, the stock o f external debt will need careful watching. 2.36 Global and regional economic cooperation: Kenya i s signatory to a number o f obligatory trade agreements, including EAC, COMESA, ACP-EU, the World Trade Organization (WTO), and so forth, that require systematic alignment o f its domestic trade and tariff policy regimes with its commitment to the trading blocs. Most o f these require downward review and harmonization of the country's tariff rates with immediate implications for revenue collection. Embodied within these initiatives i s the extensive liberalization o f trade by member states. Kenyan producers have been adversely affected, especially insectors and industries where the economy. Commentson the medium-termfiscal strategy inthe ERS 2.37 Annex 2 provides details o f the medium-term fiscal strategy outlined in the ERS and its underlying assumptions. The strategy should be seen more as aspirations o f the government in terms o f fiscal objectives; forecasts and projections in the strategy are not entirely credible, largely because o f the uncertainty around external flows that the strategy i s based on. 2.38 The strategy is expected to be revised as part o f the next budget cycle in 2004/05. A few comments are in order here: (a) The strategy sets the target of maintaining revenue above 21 percent o f GDP and also discusses the need for reforms to achieve this. But how the increase in revenues will come i s not clear. There i s a need to take a hard-headed look at feasible reforms and their likely yield. It would be prudent to compute financing requirements on the relatively more conservative increase in revenues. (b) On external assistance, it would be better to explore what 14 might be a plausible range of outcomes. (c) The base-case macroeconomic assumptions do not seem unreasonable inregard to growth, but there must be doubts about the implied speed o f changes envisaged for savings and investment. The dramatic fall inconsumption (on the order o f 20 percent over the period from 2003 to 2007) i s also implausible. (d) Given the rather bleak implications for domestic financing, some alternative scenarios-including, not least, a consideration o f how to manage a reduction in spending-are required. (e) In general, there should be some discussion o f the main risks to implementing the strategy and how these might be managed. 15 3. COMPOSITIONOF GOVERNMENTEXPENDITURE ECONOMIC COMPOSITION 3.1 The government expenditure inKenya is broadly classified into recurrent and development categories: recurrent referring to expenditure o f repeated nature that i s generally less discretionary and i s made on ongoing programs and activities, and development referring to one-off expenditure that i s generally more discretionary and i s made on new programs and activities that have yet to reach completion. The two categories also differ in the source o f funding: recurrent expenditure i s largely domestically funded, and development expenditure i s largely externally funded. The dividingline betweenthe two is, however, not very clean, andrecurrent expenditure is probably undercountedbecause the development budget also funds expenditure that i s essentially recurring in nature. In other words, development expenditure overcounts capital expenditure. The long-term goal in Kenya is, o f course, to move away from this dual classification to a unified, program-based classification inline with the best practice. 3.2 Recurrent expenditure accounted for about 84 percent o f the total in the past two years. Figure 3.1 shows that, historically, the share o f recurrent expenditure had been increasing until 2001/02. The trend has changed since 2002/03, and the share o f development expenditure has increased inthe past two years. Figure3.1. RelativeSharesof Recurrentand DevelopmentExpenditure Y.". Source: Ministryof Planning andNational Development (MPND). RecurrentExpenditure 3.3 Figure 3.2 has line plots o f shares o f different components o f recurrent expenditure in total expenditure between 1994/95 and 2002/03. The noticeable long- term trends are the falling share o f interest payments and rising shares o f defense expenditure and pensions. The share o f the civil service wage bill has started 16 increasing again after a significant fall in 2000/01. The share o f O&M expenditures has been flat lately. Transfers to parastatals have been going up inthe past three years. Figure 3.2. Shares inTotal Expenditure, 1996/97-2002/03 j 40.0 -W- Interest payments K 30,0/ -I+sewicewage Civil bill 20'o- +Pensions, - etc. O&M -+-Transfers xx Defenseand NSIS' *National Security IntelligenceServices. Source: MPND. Transfers to Parastatals 3.4 Transfers to parastatals (excluding the Teachers Service Commission (TSC) accounted for more than K Sh 28 billion and 34 billion-about 10.7 percent and 11.8 percent of the total expenditure-in 2002/03 and 2003/04, respectively.' These transfers include equity and loans to parastatals but not on-lending. Recurrent transfers to meet current expenses o f SAGAs or to compensate for operating losses, pay for the goods or services produced, sold, or imported by the enterprises constituted about two-thirds o f the total transfers in these two years. The balance o f the transfers was made on the development account (though not necessarily for acquisition of assets). 3.5 The level o f transfers has steadily increased over the past three years in terms o f share of total expenditure as well as GDP (from 2.4 percent o f GDP in 2001/02 to 3.O percent of GDP in2003/04, as shown infigure 3-3). 3.6 Statistical annex table A.6 lists the amount o f transfers made to various parastatals between 1999/00 and 2003/04. Over the period 1999/2000-2003/04, the ministries that recorded the highest transfers were Education (29.5 percent), Finance (20.9 percent), Agriculture (14.6 percent), Health (13.1 percent), and the Office o f the President (6.0 percent). For fiscal year 2003/04, the highest transfers were receivedby Education (25.3 percent), Finance (26.1 percent), Agriculture (8.6 percent), Health (13.0 percent), and the Office o f the President (7.0 percent), accounting for 73.7 percent of total transfers. "Parastatals" include SAGAs as well as state-owned enterprises, under the State Corporations Act. The current institutional classification does not differentiate between institutions providing a public service, generally not profit-oriented, and enterprise-like bodies that are, in principle, established on a commercial, primarily profit-oriented, basis. The former group (nonprofit) i s part of general government, while the latter (commercial) are part of the wider public sector, according to GFS principles. From a budget point o f view, transfers to nonprofit government institutions would be termed grunts, while those to state enterprises, should their incomes not cover their operating costs, subsidies. 17 3.7 Withinthe MinistryofEducation, virtually all transfers (excluding transfers to the TSC) are grants to the public universities; the balance i s for the operating costs o f the Kenya National Examination Council, Kenya Institute o f Education, and Kenya Institute of Special Education. The data exclude grants to primary schools for the implementation o f free primary education and subsidies to public secondary schools. Figure 3.3. Transfersto Parastatalsas Share of Total Expenditure(YO) 12.0 - - - - - - -1 8.0 - - - 2.7 4,4 3.9 2.1 3.8 - - 0Development 4.0 - 0Recurrent 7.6 6.4 7.4 6.9 7.5 0.0 Source: MoF. 3.8 The transfers from the Ministry o f Finance are mainly grants to the Kenya Revenue Authority (KRA) and occasional subsidies to profit-making enterprises (for example, the Agricultural Finance Corporation (AFC), Kenya Meat Commission, Kenya National Assurance Company, Kenya Ferry Services Ltd., Kenya Railways Corporation, and Kenya Reinsurance Corporation) and loans/equity in Telkom Kenya, the National Bank o f Kenya, and the Nzoia Sugar Factory. The grants to the KRA mainly constitute wages, salaries, and operating expenses within the agency. The loan and equity transfers are one-off expenditures, for example, to assist a for-profit enterprise that mightbe experiencing financial problems. 3.9 The transfers within the Ministry o f Health are mainly grants to Kenyatta NationalHospital, followed by the Kenya Medical Research Institute, Kenya Medical Training Colleges, and M o i Referral and Teaching Hospital. Within the Ministry of Agriculture, 85 percent o f the transfers are grants to the Kenya Agricultural Research Institute (KARI), with some occasional subsidies to the AFC. Within the Office of the President, grants have beenmade mainly to the Kenya Anti-Corruption Authority and NationalAIDS Control Council. 3.10 The trends in ministerial shares o f subsidies and grants mainly reflect the change in the organization o f the GoK. For example, grants to Kenya Ferry Services were moved from the Ministry o f Finance to the Ministry o f Transport and Communication, and grants to Kenya Meat Commission were made by the Ministry o f Finance rather than the parent Ministry o f Livestock and Fisheries Development. KARI has been reconstituted to include the Kenya Trypanosomiasis Research Institute (KETRI) and Kenya Veterinary Vaccines Production Institute (KEVEVAPI) under the Ministryo f Agriculture. 3.11 Table 3.1 sorts transfers to parastatals by the broad functional categories used by the Ministry o f Finance. Although subsidies to for-profit (commercial) public enterprises appear relatively small compared with grants to SAGAS, the current 18 transfers are not the only cost to the exchequer o f maintaining these commercial enterprises. The forgone return on government equity in these enterprises i s a huge opportunity cost. Also, transfers to for-profits have been increasing in the past three years. 3.12 It is important to conduct an in-depth analysis of the expenditure and cost- recovery profiles o f the main government units that receive grants to ascertain their outputs in relation to the funds transferred to them from the central government. In addition, some o f the SAGAScan be merged with the parent ministries or privatized or closed down ina future reorganization o fthe government. Table 3.1. Transfers to Parastatals by Functional Categories-Share of Total Expenditure (YO) 1999/2000 2000/01 2001/02 2002/03 2003104 Commercial 1.2 2.8 0.6 2.0 2.6 Regulatory 2.4 1.8 2.1 1.8 2.4 Researchinstitutes 1.5 1.5 1.2 1.2 1.2 Educationandtraining 3.6 3.0 3.6 3.5 3.3 Regionaldevelopment 0.4 0.3 0.3 0.2 0.2 Social service 1.2 1.o 1.8 1.9 2.1 Total 10.3 10.3 9.5 10.7 11.8 Source: Basedondata providedby the Budgetary SupplyDepartment(BSD). Operations and Maintenance 3.13 Operations expenditure pertains to the employees' productivity, and maintenance expenditure i s related to the productivity o f existing capital stock. The outcome o f inadequate provision for operations i s reducedproductivity o f government staff and weak delivery o f services. The outcome o f inadequate provision for quality maintenance i s high depreciation and depletion o f public capital stock. Clearly, combined, O&M expenditure is a critical determinant o f the productivity o f the public sector and the quality o f service delivery. 3.14 In recent years, O&M spending has stabilized at about 18 percent of total expenditure; however, for two reasons, this figure probably understates the true extent o f O&M spending. First, a significant portion o f O&M spending occurs through donor funding and thus via the development budget.Second, transfers and grants going from the main budget to autonomous organizations often fund some O&M expenses, but these are not recorded as such.2 3.15 In Kenya, as in most other developing countries, O&M is a "discretionary" expenditure and receives the residual amount o f recurrent funding after meeting personnel costs. It i s also the category that gets squeezed out intimes o f budget cuts. Low provisions for O&M have been a cause for concern in the past. Among the * A roughestimate ofthe true level of O&M spendingcan be arrivedat by addingthe following: (a) the official figure for O&M (18 percent o f total expenditure), derived from the recurrent budget; (b) an estimated level of O&M spending from the development budget (based on previous analysis) of 20 percent of development spending, or 4 percent of total spending; and (c) a rough calculationthat the amount of spending from transfers and grants that goes to O&M spending i s about 3 percent of total expenditure. The latter comes from the working assumption that half of all transfers and grants, excludingtransfers to the education sector (which are used almost exclusively for teacher salaries), are spent on O&M. This roughexerciseprovidesan estimate of 25 percent. 19 systemic reasons for the suboptimal spending on O&M i s separation o f recurrent and development budgets, which results in inadequate provisions for O&M on project completion. The mechanism for estimating and "factoring in" the recurrent cost o f projects has been weak, especially with the elimination of the Public Investment Programme (PIP). Donors and the government have also revealed a preference for new capital infrastructure over better maintenance o f the existing one. Also, ministries fail to maintain a register o f assets, which would permit the implementation o f a schedule o f routine, periodic maintenance. Last, the lines o f responsibility for maintenance are often unclear, thus neither the facility manager nor the public works office carries out the necessary work. 3.16 Analysis o f O&M expenses at the aggregate level does not provide much insight into the adequacy o f funding levels becausethere are significant interministry variations incounting o f operational expenses and how they are funded. For example, the operations budget in the education sector funds boarding expenses, the school feeding program, teacher travel, and schools' utility bills, and so forth. But school equipment and teaching and learning materials are funded through separate budget heads (that is, they are not reflected in the O&M expenditure as counted) and are largely met through donor programs. Clearly, a case for increasing boarding or school feeding program expenditure must be based on a careful policy analysis, and it cannot be made on the general grounds o f increasing productivity in the education sector through higher O&M spending. In the health sector, though, the operations budget funds purchase of drugs and dressings, vaccines, food for patients, linen and clothing, and other similar items, which are all directly related to the effectiveness o f service delivery. Therefore, discussion interms o f aggregate operations expenditure would be more meaningful. 3.17 The appropriate level o f O&M expenditure depends on the service delivery standards that should be developed for specific programs, sectors, and services. These standards would, in turn, determine the level o f physical inputs and their costs, which would be used as norms for future funding. The government should give priority to developing cost norms and service standards for programs and interventions that directly help in achieving identified indicators and targets o f performance. To arrive at realistic cost norms would be an iterative process in which an attempt i s made to match the level o f coverage and services to the aggregate resources. Much o f maintenance expenditure would also be amenable to such norms, for example, for vehicles, buildings, and roads according to their age. The Ministry o f Finance needs to take a lead in calling upon line ministries to develop databases and management information systems, as well as norms and an updated database o f physical assets, to justify their O&M budget estimates. DevelopmentExpenditure 3.18 The share of development expenditure inthe total expenditure was on a long- term declining trend until 2001/02, but it went up in2002/03 to more than 16 percent (from 11 percent in 2001/02). Most of the increase was financed by extemal assistance and accumulation o f arrears. The share o f GoK-financed development expenditure has been about 5 percent o f the total expenditure (and about 30-40 percent o fthe development expenditure) inrecent years (table 3.2). 20 3.19 Development expenditure exhibits a higher degree o f variability across years. This scenario can perhaps be attributed to the development budget's high level of donor dependency and greater government discretion in this area, resulting in expenditure cuts. 3.20 A significant proportion ofthe donor component is actually recurrent innature but is conventionally recorded inthe development vote. Some of the recurrent items in donor projects are procurement of vehicles and equipment, consultancy services, and wages and salaries o f project staff. The donor contribution to capital expenditure i s therefore lower than shown intable 3-2. Table3.2. DevelopmentExpenditure,ExcludingNetLending(K Shbillion) Year Total Donors GoK % GoK 2000/01 33.9 11.6 22.4 66.0 2001/02 24.9 14.5 8.6 34.3 2002103 43.7 17.9 25.8 59.0 2003/04a 45.5 a. Projected. Source: BMD. 3.21 Stalled Projects. During the 1990s, Kenya built up a substantial inventory o f stalled projects as lax fiscal control led to the initiation o f more projects than couldbe sustained by the development budget. As a result, by 1997, the completion rate for projects was as low as 3 percent. Stalled projects (those that were initiated but not completed and are currently not receiving funding) mushroomed. By 1999, the GoK had a total o f 164 stalled projects with an estimated original contract cost o f K Sh 31.4 billion, accrued expenditures o f K Sh 13.3 billion and an estimated completion cost o f K Sh 13.2 billion for those that could be completed and K Sh 2.4 billion for those that required termination. By 2003, the stalled projects were estimated to have increased to 207. The government will review all stalled projects with a view to terminating as many as possible and completing only those in the ERS core priority areas or that have potential for meeting ERS core objectives. It i s estimated that reviewing the stalled projects will cost K Sh 90 million. The government has also decided to give priority to completing stalledprojects over initiatingnew ones. 3.22 The ERS foresees a massive increase in development expenditures. The strategy has an ambitious target o f raising development expenditures from K Sh 43 billion to K Sh 92.7 billion, an annual average o f 23.6 percent. As the government embarks upon new projects, it will be important to carefully work out and provide for the O&M o f new projects and programs and also ensure that a balance i s maintained between creating new assets and O&M provision for the existing one. In this regard, the ERS investment program i s rather ambitious and too focused on creating new public assets. Public Service Wage Bill 3.23 The ERS sets two mainobjectives for the government wage bill: (a) Continue the rightsizing o f the civil service to reduce the wage bill to GDP ratio from its current level o f about 9 percent to 8.5 percent by the end o f 2005/06 and progressively down to 7.2 percent by 2007/08. 21 (b) Reform pay so that the public service provides adequate incentives to attract and retain skilled personnel and achieve a competitive remuneration structure within a sustainable wage bill. 3.24 These objectives were established in recognition o f a long-standing concern that the government's wage bill is too large. The wage bill currently absorbs about 9 percent o f the government's recurrent expenditure. As shown in table 3.3, the wage bill is considered high relative to those of similarly situated countries in the region. The table shows that the wage bill also appears high by regional standards when measured as a share o f domestic revenue (41.0 percent) or recurrent expenditure (37.4 per~ent).~ Table 3.3. Public Service Wage Bills Relative to Other Macroeconomic Indicators:SelectedSub-SaharanAfrican Countries(YO) Country 1999/2000 2000/01 2001/02a 2002/03a 2003/04a Ghana Wage bill to GDP 5.60 5.20 5.20 5.30 Wage billto domestic revenue 34.15 29.38 29.89 27.04 Wage billto recurrentexpenditure 34.15 28.11 26.80 30.11 Kenya Wage bill to GDP 8.6 8.1 8.4 8.7 9.0 Wage bill to domestic revenue 41.7 35.6 36.7 41.0 39.3 Wage bill to recurrent expenditure 37.0 34.2 32.1 37.4 38.3 Malawi Wage bill to GDP 5.20 6.40 6.20 Wage bill to domestic revenue 28.42 38.10 34.44 Wage bill to recurrent expenditure 23.O 1 26.02 26.84 Mozambique Wage bill to GDP 5.80 6.00 6.70 6.60 Wage bill to domestic revenue 48.33 41.96 49.26 52.80 Wage bill to recurrent expenditure 47.54 46.15 42.95 48.18 Rwanda Wage bill to GDP 5.30 5.20 5.20 5.10 Wage bill to domestic revenue 54.10 53.30 45.20 41.30 Wage bill to recurrent expenditure 40.00 41.00 36.30 33.20 Senega1 Wage bill to GDP 5.69 5.70 5.79 5.29 Wage bill to domestic revenue 32.91 31.48 32.19 29.08 Wage bill to recurrent expenditure 47.45 43.16 42.29 44.11 Tanzania Wage billto GDP 4.20 4.00 4.00 4.30 4.40 Wage billto domestic revenue 37.17 33.90 32.79 34.96 35.20 Wage bill to recurrent expenditure 35.59 31.01 31.01 28.67 29.33 Zambia Wage billto GDP 5.30 5.50 6.30 6.20 7.10 Wage bill to domestic revenue 30.29 27.78 34.81 34.25 - Wage billto recurrent expenditure 31.74 31.79 32.98 35.84 - -Datanotavailable. a. Projections. Source: Valentine and Wheeler (2003). 3.25 The total volume o f the wage bill depends on both the number o f employees and the average wage levels in the public service. The paradox of the wage bill in Kenya has been that downsizing o f the core civil service during the 1990s did not make a sustained dent in the wage bill. The fiscal space created by civil service downsizing was appropriated by an increased number o f teachers and ad hoc wage and allowance increases. The next paragraphs argue that previous reform efforts in Kenya's wage bill does not include defense forces; if included the current wage bill will likely be around 11percent of GDP. 22 Kenya focused largely on the number of public servants and not enough on pay reform. A sustainable public service reform would need to focus on the mechanism for salary adjustment and structuralpay reform, along with the size o f the service. 3.26 In 2002, total public service employment in Kenya was about 427,000,4 of which teachers constituted the biggest single group. TSC employees, who are essentially all teachers, constituted 55 percent o f the total public service employment in2002. Uniformedservices staff, medicalpersonnel, and other workers inthe central government accounted for 16 percent, 6 percent, and 23 percent o f the total, respectively (figure 3.4). Figure3.4. Structureof the PublicService Employment in2002 0Other central government 23% 0 Medical personnel Uniformed Services 16% I Source: Department o f Personnel Management. 3.27 The structure o f public service employment in Kenya has seen large changes over the years. As shown in table 3.4, between 1995 and 2003, central government employment, including uniformed services (but excluding defense), went down by about 24 percent; at the same time, the number o f teachers went up by about 10 percent. The reduction in the size o f the core civil service was effected through implementation o f the 1993/98 Civil Service Reform Programme (CSRP) and the post-1998 program that reduced the size o f the core civil service through volunta early retirement, mandatory retrenchment, a freeze on new recruitment, and attrition. r Public service includes the civil service, uniformed services, and teachers. It excludes defense staff, the NationalAssembly, and employees o f SAGASand commercial public sector enterprises. Total employment has gone down as well in parastatal bodies and in commercial public enterprises where the public sector has a majority control. Civil service reductions have been accompanied by significant increases in local government employment. 23 Table3.4. Employmentinthe PublicSector (Thousand), 1995-2003 Public sector component 1995 1998 2001 2003 Change, 1995-2003 (%) Central government (civil service 256.2 219.1 195.7 195.0 -23.9 and uniformed services) TSC 214.2 241.3 231.3 234.8 9.6 Parastatalbodies 109.7 112.8 101.6 97.3 -11.3 Majority control by the government 50.3 52.5 47.5 46.4 -7.8 Local Government 57.9 74.9 82.3 85.6 41.8 Total 688.3 700.6 692.5 659.1 -4.2 Source: Economic Survey 2004. 3.28 The CSRP has had a profound effect on the grade structure o f the civil service, too. Annex 3 provides a short note on how the employment and pay structure o f the civil service has changed as a result o f the downsizing carried out over the 1990s. The main point o f this note i s that any future civil service rightsizing exercise must also track the overall changes in grade structure so that the government's need for a balanced skillmix i s consistently met. 3.29 The civil service downsizing was accompanied by many ad hoc changes in salaries and allowances over the years. The most important was the substantial salary award to teachers before the 1996 election. At that time, the government implemented only the first phase o f the award. However, the National Rainbow Coalition (NARC) government has agreed to implement the unpaid phases o f the 1996 award over a six- year periodn6Apart from teachers, significant increases were given to salaries and allowances o f top civil servants and constitutional officers inJuly 2002, the minimum basic salary o f uniformed service personnel was reviewed in January 2004 and backdated to July 2003, medical doctors were given significant increases in their monetary allowances, a housing allowance benefiting all public servants was introduced in July 2001, and there was an ad hoc upgrading o f selected civil service posts. The core civil service (apart from top posts) has not received any salary adjustment since the last review in 1997. 3.30 The ad hoc salary and allowance adjustments have distorted the wage structure of the public service and made the remuneration structure less transparent. The wages of teachers have increased faster than those o f the rest o f the public servants. As table 3.5 shows, between 1996/97 and 2002/03, real average wages for primary teachers increased from K Sh 97,532 to K Sh 146,630, an increase o f 50 percent. Average wages for secondary school teachers rose from K Sh 100,120 to K Sh 210,084 over the same period, a rise o f 100 percent. Wages o f civil servants have fallen relative to those o f teachers. In fact, there appears to have been a real decline in the average earnings o f civil servants over these years. Teachers received a 9 percent salary adjustment in financial year 2003104, with a commitment to 9 percent salary adjustment duringeach o f the next five fiscal years. 24 Table 3.5. Evolution in Teachers' Real Wages, 1996197-2002lO3 1996/97 1997198 1998199 1999/2000 2000/01 2002/03 Primary Total pay and allowances 17,653 23,988 22,646 22,841 21,942 26,100 (K Shmillion) Total number of teachers 181 189 190 183 179 178 (thousand) Average annual allowance 97,530 126,921 119,189 124,814 122,581 146,629 andpay (K Sh) Secondary Total pay and allowances 6,573 9,551 10,551 9,893 9,369 13,194 (KShmillion) Total number of teachers 66 66 77 68 70 63 (thousand) Average annual allowance 100,120 143,912 136,505 144,887 134,486 210,084 andpay (K Sh) Note: All wages deflated, usingconsumerprices of October 1997. Source: World Bank, 2004a. 3.3 1 Inmost countries, parastatals and local authoritiesuse the central government pay structure as a benchmark for determining their own pay structures. However, in the absence o f a systematic and regular review o f salary and allowances, parastatals and local authorities in Kenya seem to have embarked upon ambitious salary increases. Infact, as shown intable 3.6, average earnings inparastatals appear to have gone up by more than 100percent inreal terms over the 1998/2003 period. Table 3.6. Estimated Average Real Earnings per Employee (K Sh per Year), 1998-2003' YOchange 1998 1999 2000 2001 2002 2003 (1998-2003) Civil service and uniformedforces 107,076 105,669 97,729 103,721 106,893 100,462 -6.18 TSCb 137,126 134,049 125,881 124,942 142,747 136,089 -0.76 Parastatalbodies' 120,103 150,926 183,622 209,779 250,389 263,202 119.15 Majority controlby the 154,865 176,478 242,027 276,049 328,674 346,000 123.42 public sectord Local Government 117,655 136,738 150,414 167,724 191,145 196,211 66.77 Total public sector 124,037 130,741 136,409 147,971 167,670 166,886 34.55 Memorandumitem: Total privatesector 123,113 135,339 141,972 154,279 173,295 181,833 47.70 a. Adjusted for the rise in consumer prices, with baseperiodOctober 1997. b. Refersto position at June 30, but the June figures are annualizedby multiplyingby 12. c. Refersto GoK wholly owned corporations. d. Refers to institutions where the government has a 51 percent or more shareholdingbut does not fully own them. Source: Kenya Economic Survey 2004. 3.32 The lack o f a comprehensive view o f pay reform has left civil service wages excessively decompressed. Although Kenya compensates its topmost personnel far better than civil services inother countries with per capita GDPs that are comparable to its own, its median salary lags behind such countries as Tanzania and Uganda. The 'Numbers in table 3.6 do not control for changes in employment structure in each sector and should therefore be interpreted carefully. Comparing average earnings across subsectors (or across time inthe same subsector) is not instructive in the absence o f more information about differences (or changes across time) inthe underlyingemployment structure. For example, a lower average for the civil service could be the result o f a relatively more bottom-heavy employment structure in the civil service. Similarly, in education, the number of primary teachers decreased by about 4 percent between 1999 and 2003, while the number o f secondary teachers increased by about 14 percent during the same period. This shift in employment structure is partly responsible for the higher average earnings o f teachers during 199912003, 25 ratio o f the top to minimum salary in Kenya i s a staggering 118:1, which is far in excess o f those o f other countries in table 3.3, except for Malawi. The top to median salary ratio (53: 1) is exceeded only by Malawi. 3.33 The compensation structure in Kenya has also come to rely too heavily on allowances. On average, allowances make up about 46 percent of total monetary compensation in the civil service. For some job groups, especially those toward the top o f the compensation structure, allowances far exceed basic salaries as a form o f compensation. Incontrast, allowances in Tanzania make up only about 6.7 percent of the civilianportion o f the central government wage bill. 3.34 The proliferation o f allowances has several negative consequences, includinga reduction inrationality and transparency and a weakening inthe relationship between job performance and remuneration. Allowances also make budgetingmore difficult because they fluctuate significantly from month to month. In addition, they are often providedtax free and thus weaken the government's potential to broadenthe tax base. Lessons GoingForward 3.35 The main lessons from previous efforts at wage bill reform are that the high wage bill issue i s not just a matter o f the number o f employees. It i s very much also an issue o f pay reform. There i s a need for a systematic and comprehensive approach to pay adjustments that would take into account the employment and wage structure o f the entire public service and also address the issues o f too many allowances and too decompressed a salary structure, while linking pay reform with creating adequate incentives for performance. Specifically, and most importantly, there may be a need to revisit the number and wages o f teachers. There can be no solution to the wage bill issue as long as teachers are out o f the equation. 3.36 Concerns have also emerged around the continuation o f a freeze on new hiring. The average age of public sector employees has been rising because o f the freeze, which has been inplace for more than 10 years. In 2002, the average age was about 44 years, excluding the uniformed forces, which implies that average age could rise to 49 years (equivalent to current estimates o f life expectancy) by 2007 unless the freeze on employment i s lifted. The concern i s that a high average age could lead to succession shocks, where a sizable portion o f institutional memory will be lost in a short time. Recent Developments 3.37 The government has recently announced a Targeted Voluntary Early Retirement Scheme (TVERS) that will aim to bring21,388 civil service workers into early retirement by June 2008; this i s in addition to the expected net natural attrition o f 8,000 civil service workers over the same period.' The TVERS i s expected to reduce the size of the civil service by slightly more than 11 percent. It will focus on specific areas where staff i s in excess of requirements. After careful analysis of The expected net attrition comes from expected gross natural attrition of 3,000 per year, along with planned hiringof 1,000 new employees annually. 26 staffing levels, staff found to be in excess o f the required number will be allowed to retire under the scheme or be redeployedaccordingly. Recommendations 3.38 The TVERS i s a large step in the right direction. If it succeeds in reducing civil service strength as planned, it should be complemented with systematic pay reforms on a larger-scale that raise compensation for core professional, technical, and management personnel in public service in the middle to upper-middle salary groups and introduce a pay structure that i s overall more rational. Meanwhile, adjustments to salaries should be based on a more comprehensive review that looks at the pay structure of the entire public service. 3.39 Although not strictly a central government wage bill issue, there i s a need to review transfers being made to SAGAS because the government often funds staff salaries in these agencies, and these salaries appear quite high compared with civil service salaries. The high increase in employment and wages in the local authorities should also be o f concern because the central government supports them through Local Authority Transfer Fund(LATF), their wage bills and employment are growing faster than their revenues at the aggregate, some have loans guaranteed by the central government, and local authorities and the central government play complementary roles inservice provision. 3.40 The GoK should also implement institutional mechanisms to establish a more performance-oriented approach. This would focus on improving accountability, monitoring performance, and rewarding good performers. The latter could be done through attractive annual salary increments and step differentials. EXPENDITURE COMPOSITION BY MINISTRIES Broad Shijts in Composition of SpendingBetween 1999/2000 and 2002/03 3.41 Figure 3.5 shows the composition o f total public expenditure in 1999/2000 and in2002/03.9 Infact, the changeshavenot beenvery large. The share ofeducation has declined slightly (from 37 percent to 35 percent o f total spending), as have the shares o f agriculture and related activities (from 7 percent to 5 percent) and infrastructure (from 9 percent to 6 percent), while the shares o f administrative and security-related functions (including the Office o f the President) have increased from about 27 percent o f spending in 1999/2000 to about 32 percent in 2002/03. Apart from this, there has been remarkably little change in the broad composition o f expenditure. Clearly these very global estimates mask some more pronounced changes inthe shorter run, as well as in the more detailed composition o f spending (both discussed below), but they do show that there has not been any dramatic shift in public expenditure policy accompanying the wider changes that have affected Kenya in the past five years- except, perhaps, for the conscious effort to reduce the role o f the state in agriculture and a tendency toward growth inadministrative functions. Actual public expenditurefor bothyears, recurrentplus development. 27 Figure3.5. ChangeinDistributionof PublicExpenditure(Actual % of Total) - 2002/03 I 1999/00 AgricuitureLan 5".:, \ Economic "\ Economic 0.0% 1 Other 2% Other Agriculture 2% General Office Of \\// I General Admin the 20% Officeofthe 14% /President 37% \Defense Defense 11% 10% Source: BMD. 3.42 This discussion of relative shares tends to conceal the very significant growth that has taken place in absolute levels of spending in almost all sectors (table 3.7). Total spending increased innominal terms by about 70 percent over this period.loFor example, even though the share o f education fell, absolute spending nonetheless increased by K Sh 21.8 billion (US$235 million equivalent, or 37 percent innominal terms), representing the second-largest absolute increase in public spending. The largest increase has been inthe area o f general administration, which increased by K Sh 23 billion, almost tripling innominal terms over the period.'' Note that spending on infrastructure, health, and defense have all about doubled in nominal terms over the past four years, representing real increases in the neighborhood of 50-60 percent, or about 10 percentper year. Table 3.7. Total Ministerial Spending by Broad Functional Categories, 1999/2000-2002103 Actual spending Actual spending Change 199912000 2002103 K ShmillionUS$ million K Sh millionUS$million K Sh million % change General administration 17,818 234 36,356 492 18,538 104 Office of the President 15,579 204 20,717 280 5,138 33 Otherjustice & security 564 7 1,45 8 20 894 159 Defense 12,703 165 20,917 283 8,214 65 Education 46,192 605 63,377 858 17,185 37 Health 9,641 126 15,514 210 5,873 61 Infrastructure 10,969 144 11,742 159 773 7 Agriculture, lands 8,536 112 9,176 124 640 7 Economic 1,917 26 1,917 n.a. Other 2,398 31 3,097 42 699 29 Total 124,400 1,630 184,835 2,501 60,435 149 Source: BMD. loIn an environment where overall inflation was about 35 percent over this period (change in GDP- deflator), this is significant, representing about a 30 percent real increase in spending. The explanations for this very large increase are not clear. 28 Changes in Actual Spendingin thePast Two Years 3.43 Turningto more recent experience, table 3.8 shows the composition of actual recurrent spending in the past two years by ministry. Major increases in recurrent spending took place in health and education last year, with offsetting declines in spending on public works and agriculture. Spending on the Electoral Commission increased (by about K Sh2.5 billion) as a result o f the elections; otherwise, ministerial recurrent expenditures remained similar to those in the previous year. Note that MinistryofFinance expenditure varies significantly from year to year becauseone-off items such as parastatal restructurings, contingencies, and the like are included under this vote.12 Table 3.8. Actual Recurrent Spending, 2001/02 and 2002/03 (Current K Sh Million) 2001102 2002103 % share o ftotal Actual Actual 2002103 spending Office of the President 19,166 17,75 1 10.7 Finance 6,864 8,043 4.9 ForeignAffairs 3,928 3,817 2.3 Home Affairs 4,255 4,534 2.7 Local Government 3,697 3,156 1.9 National Assembly 3,296 3,700 2.2 Electoral Commission 1,306 3,863 2.4 Other administrativea 3,144 2,553 1.6 Judiciary & relatedb 1,625 1,439 0.8 Defense & NSIS 19,048 20,917 12.6 Education 52,589 60,938 36.8 Health 11,985 14,422 8.7 Public Works 9,260 5,473 3.3 Transport 1,657 1,863 1.1 Water - 1,305 0.8 Trade & Industry - 1,756 1.1 Agriculture 4,762 3,663 2.2 Environment & Natural Resources 2,791 1,874 1.1 Other rural-relatedc 1,149 1,437 0.9 Otherd 3,881 3,220 1.9 Total 154,403 165,724 Note: Spendingon Trade and IndustryandWater were classifiedunder different ministry votes in2001/02, makinghirect comparisonsdifficult. a. Includes the State House, Directorate of Personnel Management(DPM), Officeof the Vice Presidentand Ministry ofNational Reconstruction(OW), Public Service, and Auditor General. b. IncludesJustice, thejudiciary, andthe Attomey General. c. Includes cooperativesector and Lands andSettlement. d. Includes Labour, Culture and Sports, Energy, and Information andTourism. Source: BMD DevelopmentSpending 3.44 Analyzing year-to-year changes in development budget spending i s not very instructive, because there can be relatively large sectoral swings inthe short term due to individual projects starting or stopping, being funded, or failing to be funded. Figure 3.6 shows the broad composition of development expenditure averaged over l2The reasons for the apparent large changesinspendingbythe Office ofthe President are not clear. 29 the past five years. The current composition i s roughly comparable, although the share o f education has increased (to 13 percent of development spendingin2002/03, largely due to inclusion of funding for the free primary education initiative under the development budget). The large share o f apparently administrative functions i s because the development budget i s not a capital budget, but includes many recurrent items or items that represent project initiatives. Figure3.6. Composition of DevelopmentSpending(Actual), 1998/99-2002/03 Agriculture, General Lands, etc. Other administration Health Education 19% 7% 7% Source: BMD. Signals of PolicyIntent-Changes inthe Budgetover the LastTwo Years 3.45 Actual spending tells us how public resources are being used; changes in budgets tell us more about what the government's intentions are. Table 3.9 summarizes the main changes made in the recurrent budget in this past year for the largest ministries. 3.46 In the recurrent budget, at least, the government has signaled fairly clear prioritization: increasing allocations significantly on education (by K Sh 16 billion, or 30 percent) and to a lesser extent on health and a range o f administrative functions (some o f which are compensated for by reduced spending on the Electoral Commission). The large increase inthe budget for the Office o f the President (K Sh 4 billion) i s mostly accounted for by the needto fundthe pay increase for police. Table 3.9. Changes in Budgeted Recurrent Expenditure, 2002/03-2003/04 (CurrentK Sh Millions) 2002103 2003104 Major budget budget revisions Change % change Office ofthe President 20,333 20,536 (24,653) +4,320 +2 1 Finance 7,576 15,319 (8,413) +837 +11 ForeignAffairs 3,827 4,405 +578 +15 HomeAffairs 4,473 5,992 +1,519 +34 Local Government 3,804 4,467 +663 +17 NationalAssembly 3,359 4,672 +1,313 +39 ElectoralCommission 4,661 1,213 -3,448 -74 Other administrative 3,151 2,689 -462 -15 30 2002103 2003104 Major budget budget revisions Change % change Judiciary & related 1,670 1,957 +387 +17 Defense & NSIS 17,338 21,755 (17,736) +389 +2 Education 54,709 71,943 (70,920) +-16,211 +30 Health 13,652 16,005 +2,353 +17 Public Works 10,597 10,865 +268 +3 Transport 1,719 1,662 -57 -3 Water - 2,117 n.a. n.a.a Trade & Industry 2,299 2,196 -103 -4 Agriculture 6,976 6,163 n.a. Environment & Natural Resources 3,516 2,440 n.a. n.a. Other rural-related 1,282 2,246 n.a. n.a.a Other 2,307 5,219** ** ** Total 167,249 203,861 +22 n.a. Datanot available. ** Data error-should be closer to K Sh 3.7 billion. a. Functions and ministries were reclassified in agriculture, natural resources, and water supply, thus budget lines are not directly comparable. However, net change in total spending on these functions represented only a slight decline. Source: BMD 3.47 Changes in the development budget also involve big increases for education and infrastructure (especially the Ministry o f Public Works), with offsetting declines in Home Affairs, Local Govemment, and Environment and Natural Resources (see statistical annex table A.15). As in previous years, the large discrete changes in allocations to the Office o f the President and the Ministry o f Finance represent the costs o f specific policy measures. What i s most notable, however, i s the extreme disconnect between budget intentions as revealed in the printed estimates and actual spending under the development budget. This is discussed inthe next section, but it suggests that we should not readtoo much into the budget estimates. The Apparent (and Growing) Disconnect between Budget Intent and Actual Spending 3.48 There are two elements to the divergence between budgets and actual spending: one i s the conformance o f ministries' spending behavior with budget intentions, which i s discussed in the section on expenditure management; the other (perhaps a greater concern from our point o f view in considering expenditure composition) i s what the large divergence between printed estimates and actual spendinginthe recentpast says about realismo f spending plans for the current year. 3.49 The recurrent budget is generally fully spent in any given year (actual spending equaled 99 percent of the printed estimates in2002/03, see statistical annex table A.17). This i s not surprising, given the weight o f salaries and other unavoidable obligations in the total and the tightness o f the overall funding constraint. There i s nonetheless significant variation across ministries, suggesting that line ministries are doing a lot of "balancing" within a year. For example, in 2002/03, labor-intensive ministries, such as Education and Defence were overspent (by 21 percent and 11 percent, respectively), while others, such as Local Government and Public Works, 31 were underspent (spending only 83 percent and 52 percent o f their recurrent budgets, respectively). l3 3.50 The disconnect inthe development budget is far more severe, and it suggests that there are real problems with the realism o f budgets. On average, only 40 percent o f the development budget was spent last year (table 3-10), with significant underspending in all ministries except Home Affairs and very notable underspending in the infrastructure sectors. The Ministries of Public Works, Energy, and Water combined, for example, spent only K Sh 2.4 billion out o f a budget of K Sh 10.7 billionin2002/03. Table 3.10. Underspending of the Development Budget, 1998/99-2002/03 (Current KSh Million) 1998199 199912000 2000101 2001102 2002103 Printed estimates 11,314 15,474 38,421 41,637 46,469 Actual spending 9,322 7,095 22,998 14,668 19,111 %underspent 82.4% 45.8% 59.9% 35.2% 41.1% Source: BMD 3.51 In fact, the situation appears to be deteriorating over time, as illustrated in table 3.11. The emerging picture suggests that the very large increases indevelopment spending that have been programmed since financial year 2000/01 have never materialized. To the extent this i s due to expected donor resources not becoming available, it argues for both more realism and greater follow-through on commitment by the donor community. However, to the extent that it is not due to donor-funding issues, it needs to be recognized that either the capacity i s not there to implement these very high levels o f development spending or the approved plans o f the individual ministries exceed the government's realistic capacity to finance them. 3.52 The reasons for disconnect are covered in depth in the discussion o f public expenditure management in chapter 4. What is clear i s that the degree o f unrealism i s persistingand increasing over time. For example, if we compare the 2003/04 printed estimates with what was actually spent (last column o f table 3.11), not with the previous year's budget, we see that the anticipated increases are massive, with some ministries programming spending o f up to five and six times the level o f actual spending in the previous year, which suggests that this disconnect has not been confronted. Table 3.11. Development Budgets Compared with Actual Spending (Current KShMillion) % actually 2003104 budget 2002103 2002103 2003104 spent in increase over budget actual budget 2002103 2002103 actual Office o f the President 9,644 3,146 5,900 33 88 Finance 4,779 3,807 7,837 80 106 Home Affairs 672 953 442 142 -54 Local Government 2,552 849 1,080 33 27 Other administrative 802 329 2,137 41 550 Education 3,218 2,439 8,434 76 246 l3This may be because their recurrent budgets incorporate substantial allocations for capital spending. 32 % actually 2003104 budget 2002103 2002103 2003104 spent in increaseover budget actual budget 2002103 2002103 actual Health 4,662 1,092 5,116 23 368 Public Works 5,235 1,198 8,663 23 623 Energy 4,992 1,086 6,578 22 506 Water Resources 450 101 579 22 473 Agriculture 3,506 1,384 3,882 39 180 Environment and Natural Resources 3,592 351 980 10 179 Other 2,365 1,726 3,290 73 91 Total 46,469 18,461 54,918 40 197 Source: BMD. THEMTEFAND ALLOCATIVEBEHAVIOR 3.53 The MTEF has been used for the past four years. Three questions are o f interest: I s the MTEF being followed in actual expenditure allocations? I s it being used to redirect expenditure over the medium term? Are the broad directions o f expenditure allocation in the MTEF consistent with stated policy priorities, as expressedinthe ERS and elsewhere? Historical Comparison of MTEF Ceilings and the Budget 3.54 Itis difficult to directly compare the historical allocation of spending with the MTEF ceilings, because different definitions were used for budget shares.l4 However, constructing an index o f the shares o f total spending under the MTEF and under budgets (table 3.12) suggests that over the four years 1999/00-2002/03, the direction of change o f total budget shares did not particularly conform to the MTEF, with substantially larger-than-programmed proportional increases in budgets for public administration and infrastructure and lower-than-programmed increases inthe shares o f human resource development and law and order." Table 3.12. Index of Change in Shares in MTEF Ceilings, Budget Allocations, andActualExDenditureRelativeto 1999/2000a 2002103 index (199912000 = 100) MTEF MTEF sectors ceilings Budgetb Actual Agriculture and rural development 101 104 72 Physicalinfrastructure 81 115 77 Humanresourcedevelopment 120 86 95 Trade, industry, andtourism 168 183 197 Public administration 95 125 150 l4Before 2003104, the MTEF specifiedthe share of recurrent funding excludingwages and other fixed recurrent costs, whereas sector allocations are reportedinclusive of the wage bill. Thus, relativesector shares are not comparable without doing additional computations. Instead, an index was created to compare directions of change. Note, however, that because of the divergence between actual and budgeted expenditure, the overprogramming on infrastructure did not materialize, and actual overspending on public administrationrelativeto the MTEF targetswas evengreater thanthe budgetswould suggest. 33 2002/03 index (1999/2000 = 100) MTEF MTEF sectors ceilings Budgetb Actual Public safety, law and order 125 110 97 Nationalsecurity 120 88 111 a. Index of change inshares calculated as (sector share in2002/03 / sector share in 1999/2000)*100. b. Printed estimates. Source: MTEF Secretariat. 3.55 Again, note the earlier caveat, that because o f definitional differences, this measure o f change i s only very indicative. A somewhat more meaningful indicator i s the changes in expenditureshares under the development budget, for which historical data are roughly comparable. The development budget shows a substantial-and continuing-disconnect between sectoral shares under the MTEF and those reflected in the budget (see statistical annex table A 16). For example, in 2002/03, infrastructure was to be allocated 50.5 percent o f development spending under the MTEF,butitreceivedonly 17.2percent ofthe budget (and only 14.1percent ofactual spending); conversely, public administration was programmed at 3.7 percent o f development spending under the MTEF, but accounted for 22.8 percent of the budget and 27.3 percent o f actual spending. These differentials may be being driven in part by the impact of donor-financing decisions on the developmentbudget, but giventhe problems o f comparability and the lack o f conformity o f budgets with the MTEF, it i s difficult to judge what impact the MTEF has had as a policy tool. The best we can conclude i s that the MTEFprocess has only weakly influenced the change o f direction o f expenditure. Conformity of the 2003/04 Budget with the MTEF Ceilings 3.56 The 2003/04 budget i s more directly comparable with the MTEF than in previous years because o f changes in the definitions used.16 Table 3.13 compares actual budget allocations in 2003/04 with the proposed MTEF ceilings. By and large, allocations conform, with the exception o f the substantial overallocation to public administration; this conformity i s not surprising, given the large weight o f recurrent spending in the totals. Comparing the development budget with the MTEF ceilings (last two columns o f table 3.13) reveals more divergence, with a substantially larger share o f the budget going to public administrationthan programmed inthe MTEF and a much smaller share to infrastructure investments. Table 3.13. Shares of 2003/04 MTEFand Budget(MTEF Sectors, 'YO) MTEF sector MTEF total 2003104 Development Development spending budget MTEF budget 2003104 2003/04 Agriculture and rural development 6.2 6.1 10.1 8.6 Physicalinfrastructure 15.2 15.5 38.9 21.5 Humanresourcedevelopment 37.2 39.2 27.1 24.1 Trade, industry, andtourism 1.7 2.3 2.6 2.5 Public administration 6.0 11.8 5.3 30.4 Public safety. law and order 16.0 13.8 13.3 11.6 Nationalsecurity 9.5 8.5 0 0 Source: MTEF Secretariat, MPND. l6Salaries appear to be distributedacross sectors inthe MTEF from this year. 34 3.57 Given the methodological difficulties, it i s hard to draw very firm conclusions, but at best it would appear that the MTEF is having only limitedinfluence on actual expenditure allocations over the medium-term: at any giventime, there i s a significant disconnect betweenthe budget and the MTEF ceilings. Poverty Focusof Spendingand Core Poverty Programs 3.58 Poverty focus refers to equity concerns o f public spending-that is, how the benefits o f spending are distributed across the poor and the nonpoor. An improvement inpoverty focus would therefore meanpotentially a larger poverty impact. InKenya, as in many other countries, the government has since the 2000/01 budget tried to improve the poverty focus o f spending by identifying a set of programs that are considered to benefit the poor more than the nonpoor. These programs and projects are referred to as core poverty programs. The CPPs were selected according to a set o f criteria agreed upon by the GoK and development partners. These include programs that would directly create employment; provide access to basic education; increase agricultural productivity; ensure access to health services, especially curative health and family planning; reduce gender disparity; provide decent shelter, clean water, and sanitation; and rehabilitate criminals; as well as programs aimed at disasters and emergencies management and environmental protection. 3.59 The criteria for selecting CPPs and relatedprojects were revised in2003/04 to make them more comprehensive and take into account new programs that were identifiedinthe ERS. Specifically, the new criteria sought to cover pro-poor programs that would increase incomes o f the poor and improve their quality o f life, security, and equality. The revision o f the selection criteria resultedin an increase inresources allocated to the CPPs for both recurrent and development expenditures in2003/04, as shown in figure 3.7 and figure 3.8. The total amounts allocated to these programs increased from K Sh28.1 billionin2002/03 to K Sh45.9 billionin2003/04. Figure3.7. RecurrentBudgetAllocation andActualExpenditurefor CPPs 20,000 15,000 -.*." h BlBudget 0 estimates v El 10,000 Q k4 W Actual expenditure 5,000 0 2001102 2002103 2003104 Fiscal year Source: BMD. 35 Figure3.8. Development AllocationandActualExpenditurefor CPPs 25,000 2 20,000 -.-.-E FilBudget 0 estimates 15,000 v 5 HActual y 10,000 expenditure 5,000 0 2001102 2002103 2003104 Financialyear Source: BMD. 3.60 The idea behind identifying CPPs was that these programs should get priority inbudgetallocations and also beprotectedagainst midyear expenditure cuts. Between 2001/02 and 2002/03, recurrent estimates for CPPs increased by 1.7 percent, while that for all ministerial expenditures rose by 6.2 percent. The corresponding numbers for the development budget were 42 percent and 11.6 percent, re~pectively.'~ These numbers show that CPPs perhaps received higher priority than the rest of the programs when it came to development budget. As to expenditure cuts, in 2001/02, the ratio o f actual to estimated development budget for CPPs was 73 percent compared with 60 percent for all ministerial expenditures. The corresponding numbers for 2002/03 were 44 percent and 33 percent, respectively. These data show that there was marginally more protection for the CPPs in 2002/03 than for other programs. 3.61 Much improvement has been made in the past year in developing the monitoring format for the CPPs, but the implementation is still not effective. Getting timely and complete reports from line ministries i s a problem. The CPPs are supposed to be protected from in-year expenditure cuts and receive full disbursements o f the budgeted amounts in a timely manner, but full disbursement rates have not been achieved. In 2002/03, for instance, the disbursement rates were 92 percent and 57 percent of the budgeted recurrent nonwage and development expenditures, respectively. In many cases, amounts disbursed have not been fully used, and there have been instances where monies disbursed for CPPs have been used for other programs. 3.62 To enable the CPPs to achieve their intended objectives, measures are needed to ensure full disbursement o f the budgeted amounts; sensitization o f accounting officers, finance officers, and project coordinators on speedy implementation o f these programs; and proper costing of poverty-related programs. Ministries should be instructed to clearly indicate the types o f program when requesting exchequer issues from the Treasury. The dissemination o f quarterly monitoring reports on CPPs produced by the BMD needs to be strengthened. The reports should be distributed widely to all stakeholders and eventually be posted on the Ministry o f Planning and National Development Website. l7The large increase in development allocations was made to Water and Health Ministries. 36 3.63 Monitoring o f CPPs can be supplemented by an analysis by poverty levels and by regions of actual incidence of spending. Such an analysis should be carried out in the next year's PER. Issues with the CPPs 3.64 Identification o f programs benefiting the poor and protecting them through administrative measures can be effective in improving the poverty focus of public spending. It can also be used as a vehicle for relating donor resources to specific program expenditures. It should, however, be seen as a transitional measure. Experience invarious countries demonstrates that the link between poverty reduction and public spending i s complex. Excessive focus on identified programs and projects has the potential to skew the spending away from the most efficient mix of interventions. Furthermore, it could detract from policy measures that may not cost muchinterms o f public spendingbut can have an indirect impact on poverty through higher incomes. These kinds o f issues have shifted the discussion toward a focus on results and identification o f spending priorities within the ERS framework. Such an approach i s briefly introduced later inthis chapter. LOOKING FORWARD-SPENDING PRIORITIESINTHE MEDIUM TERM PublicExpenditureCost of ReachingMillenniumDevelopmentGoals (MDGs) 3.65 The MDGs provide an internationally agreed set o f benchmarks (to which Kenya i s a signatory) against which to measure progress toward some fundamental development objectives. Annex 4 discusses Kenya's status on various MDGs. It can be seen that Kenya cannot hope to achieve most of the MDGs at the current pace o f progress toward them. 3.66 There i s a natural interest inunderstandingthe additional resources needed for reaching these goals. Because deriving national spending priorities from several documents would be confusing, it i s important that MDGs are properly articulated and embedded inthe next update of the ERS. Such an approach would ensure a consistent treatment o f MDGswithin an overall prioritizing framework. 3.67 However, note that costing the MDGs i s not a trivial exercise, and caution needs to be exercised against simplistic efforts to quantify the public expenditure costs of the goals. In many cases, the determinants o f MDGs-related outcomes may lie largely in other areas. For example, earlier research in other countries has suggested that the greatest single determinant o f maternal mortality was the distance a pregnant woman lived from a paved road. Similarly, progress in halting the spread o f HIV/AIDS depends as much on political leadership for attitude and behavior changes as it does on spending on HIV/AIDSprograms. Further, research inKenya shows that deworming programs have a significant impact on school attendance. Thus, an exercise that attempts to work out the public expenditure cost o f attaining the MDGs mustaccount for intersectoral synergies. 37 PlannedMTEFSector CeilingsandERSPriorities 3.68 Table 3.14 shows how the shares o f the eight MTEF sectors are planned to evolve during the period 2004/05-2006/07. The shares are basically unchanged, with about a 2 percent increase in physical infrastructure (due to a large proportional increase in development spending-see statistical annex table A 18 for the recurrent- development breakdown) and a minor increase in agriculture and rural development. The share o f human resource development i s planned to remain roughly constant, and minor decreases are plannedinpublic administration andnational security. 3.69 The directions o f change are broadly consistent with the ERS's emphasis on infrastructure investments; but as noted before, the large disconnect among MTEF allocations, budgets, and-especially-actual spendingsuggests that we can take only limited comfort from this fact. Table3.14. MTEFSector Ceilings(YOof Total) 2003104 2004105 2005106 2006107 revised ceilings ceilings ceilings Public administration 6.0 5.3 5.5 5.5 Public safety, law and order 16.0 15.0 15.1 15.1 Humanresourcedevelopment 37.2 35.6 36.5 37.0 Tourism, trade, and industries 1.7 1.5 1.4 1.4 Agriculture andrural development 6.2 5.8 6.7 6.7 Physicalinfrastructure 15.2 16.8 17.0 17.0 Informationtechnology 0.3 0.3 0.3 0.3 Nationalsecurity 9.5 8.3 8.2 8.0 Other recurrent expenditures 7.9 11.5 9.3 9.1 Total recurrentand development 100 100 100 100 Source: MTEFSecretariat. 3.70 More generally, the relatively small changes in the recurrent budget programmed under the MTEF over time suggest that budgeting i s still largely taking place incrementally. This may be rational, given that the scope for making radical shifts in budgets in the short term i s limited. It i s also worth bearing in mind that the MTEF instrument generally has two purposes: to redirect expenditure in accordance with broad policy and to assure ministries of predictability o f financing, in exchange for prioritizing within a fixed envelope. It may be that the Kenyan MTEF i s serving the second purpose to some extent in the case of the recurrent budget, because the extent that it i s adhered to it should give ministries some comfort regarding future funding. However, we cannot judge from the available data whether ministries are in fact pursuing national priorities: The limited evidence from the health sector (see chapter 5) suggeststhis may not be taking place. 3.71 Also note that the very broad categories involved could hide expenditure restructuring that could take place within each MTEF sector. For example, changes in the regional composition o f spending or changes inspending across different levels o f service could still be plannedwithin the proposed program. 38 LinkingERSPrioritieswith BudgetAllocations 3.72 In the longer run, the objective of improving the poverty focus should be subsumed within the ERS objectives. In a well-functioning MTEF and budget process, budget allocations would be closely aligned with the ERS priorities, and if the budget i s executed as drafted-that is, outturns correlate closely with the original allocations-the focus o f spending would by design become aligned with the ERS priorities. 3.73 Much o f the expenditure prioritization within key line ministries should take place around core performance indicators being identified as part o f the ERS monitoring and evaluation framework. The process would involve setting targets for these core indicators and then identifylng inputs in the form o f programs and interventions (the "core priority programs") that are essential for achieving the targeted results (figure 3.9). The identified core priority programs should then be funded sufficiently and predictably. This process of identifymg the results-to-inputs chain and providing adequate funding i s likely to be an iterative one. Once identified and fully accepted by all concerned parties, the core priority programs could take the place of CPPs being usedat present to prioritize expenditure. Figure3.9. RelatingCore PriorityProgramsto ERSPriorities ' ERS Performance Core priority programs: priorities indicators relatedto Budget heads andprograms targeted outputs and (inputs) essentialfor outcomes achieving targeted outputs Servicestandards and cost norms: Selected programsand services 39 4. PUBLIC EXPENDITURE MANAGEMENT RECENT DEVELOPMENTS 4.1 PEM-concerning the processes and institutions for MTEF preparation, annual budgeting, and budget execution and monitoring-remains fairly weak in Kenya. The ERS listed five specific weaknesses: (1) significant variations between budgeted and actual expenditures, (2) inadequate recording and tracking o f donor- funded programs, (3) administrative classification rather than an economic one, (4) failure to comply with multiyear MTEF projections, and (5) poor budgetary control leading to pending bills. These weaknesses were confirmed in an assessment o f the PEM system in Kenya in April 2004, which was carried out by the GoK in collaboration with the World Bank, IMF, EuropeanUnion, and the U.K.Department for International Development (DFID). The assessment concluded that the PEM system in Kenya needed significant strengthening: only 4 o f the 16 benchmarks were met.' 4.2 The GoK i s well aware of the weaknesses and has a reform program inplace. Some reformmeasures have beenundertaken recently: An action plan was developed to implement the recommendations contained in the 2001 Country Financial Accountability Assessment (CFAA) and PEM-AAP 2003. An economic classificationofthe budget isbeingdeveloped. Internal audit i s beingextended from preaudit to systems audit. The medium-term aspects o f budgetplanning are being strengthened. The Public Audit Act and the Government Financial Management Bill 2003 were enacted, and the Public Procurement Disposal o f Assets Bill 2003 was placed before Parliament as a means o f strengthening the legislative framework for PEM. Piloting o f the accounting module o f the Integrated Financial Management Information System (IFMIS) was begun in the Ministries o f Finance and Planning and National Development and eight local authorities. Public Expenditure Tracking Surveys (PETS) were introduced in the Ministries o f Education, Health, and Agriculture in 2003. Follow-up work i s inprogress. The PEM-AAP 2004 was an update o f a similar assessment carried out in 2003. PEM-AAP 2003 had found that 3 o f 15 benchmarks were met. 40 (h) To address institutionalized corruption in procurement, procurement officers were suspended en masse. New officers have been recruited after vetting. 4.3 The government has also decided to carry out PER work regularly to inform the budget process. Initially, the PER work each year will have two phases. The first i s preparation o f MPERs in each o f the spending units and background studies on important budget policy issues that need deeper analysis and information collection than i s routine. This phase will inform the budget strategy and be an integral part o f the annual budget timetable. The second phase involves preparation o f a PER Report. The report will provide an evaluation o f budget strategy, budgetary outcomes, and progress in reforming budgetary institutions in the previous fiscal year and make recommendations for further reform. It will also pull together relevant analytical work carried out inphase one to develop medium-term recommendations. 4.4 The government remains engaged in continuous upgrading o f PEM systems through further reforms to enhance service delivery and value for money. This chapter draws upon the findings and recommendations made by PEM-AAP 2004 and an MTEFreview exercise carried out recently to highlightareas for further reform. INDICATORS OF PEMPERFORMANCE 4.5 Weaknesses in PEM manifest themselves largely as a weak link between government policies and spending plans, low budget realism, and weak expenditure control. These drawbacks can be objectively measured using indicators such as persistence o f expenditure patterns even when they are not aligned with the stated priorities and strategy, aggregate fiscal deficit compared with approved budget, composition o f expenditure outtum compared with original budget, stock o f expenditure arrears and their growth, number o f stalled projects, evidence o f budgeted resources reaching spending units in a timely and transparent manner, instances o f misuse o f public funds, efficiency o f delivery o f services, and so on. Kenya's performance on two key indicators, the divergence between original budget and outturn, and stock o f expenditure arrears, i s briefly discussedbelow. Compositionof OutturnExpenditureversus OriginalBudget 4.6 By definition, a budget is a guide to future spending. A large variation between planned and actual spendingpatterns negates the purpose o f budgeting.* The relevant variation discussed here i s not the adjustment that needs to be carried out during the year to address unanticipated change inbudget circumstances, but routine budget adjustments that canbeplannedwithout mucheffort. * Itcan be argued that intheory, spendingunits must be given as much flexibility inuse of resources as possible within an overall budget envelope. Although this is largely true, there are several qualifications to this general statement. First, "flexibility" refers to ex ante flexibility, that is, for planning and budgetingthe use of resources.To the extent that each spending unit has the freedom to devise policies that best meet its mandate and mission, such flexibility obtains in Kenya. Second, flexibility is relevantand importantwhen spending units have clearly definedindicatorsagainst which their performance can be measured. Such indicators do not yet prevail in Kenya. Third, if reorienting expenditure beyond broad sectoral categories is a central fiscal policy (identifying pro-poor budget heads i s one such example), then Planning and Finance Ministries need to ensure that outturn at the level ofbudgetheadsis as close to budget as possible. 41 4.7 Table 4.1 shows the deviation between budget outturns and original budget at the level of selected economic aggregates for the three years, 2000/01 to 2002/03. Over the period, the average annual shortfall in actual revenue collection was 2.6 percent o f target, with a shortfall o f similar magnitude between actual and budgeted recurrent expenditure (2.9 percent). The highest shortfall inactual expenditure relative to the target was in development and net lending, at 20.2 percent. The poor performance o f the development budget i s partly due to low disbursement rates in donor-funded projects and programs. For example, actual project loan cash disbursements had an annual average shortfall of 30.3 percent, 44 percent for project loan AM, and 24 percent for program loans. Grant'revenue recorded an annual average shortfall of 7.5 percent. These comparisons show the vulnerability o f the development budget to fluctuations in external financing and also the extent o f optimistic projection o f external resources and corresponding expenditures. Table4.1. Budget Outturn,2000/01-2002/03: Deviationof Actualfrom Printed (YO) 2000/01 2001/02 2002103 2000101- 2000101-2002103 three-year 2002103 nominal change (%) Average Actual Printed Deviation A. Revenue(1 +2 +3) 2.3 -7.8 -2.3 -2.6 6.4 11.4 (5.0) 1. Ordinaryrevenue 0.5 -5.7 -0.9 -2.1 8.6 10.1 (1.5) 2. A M -8.6 -12.4 0.0 -7.0 46.2 33.6 12.6 3. Grants 31.2 -34.2 -19.4 -7.5 -37.9 1.1 (39.0) Cash -18.8 -6.8 -2.5 -9.4 -43.0 -52.5 9.5 A M 81.5 -44.1 -24.6 4.3 -35.7 54.9 (90.6) B. Total expenditure -3.0 -9.5 -5.0 -5.8 13.4 15.7 (2.3) 1. Recurrent -2.0 -2.3 -4.2 -2.9 10.9 13.5 (2.6) 2. Developmentandnet lending -8.6 -43.2 -8.6 -20.2 28.1 28.0 0.1 Source: BMDQuarterlyBudgetReview(QBR). 4.8 Among the revenue categories, actual income tax collections exceeded targets by an annual average of 3.5 percent. However, the gains from this category are more than offset by the shortfalls inimport duty (7.4 percent) and excise duty (6.4 percent) as shown in statistical annex table A.20. The shortfalls in collections o f import and excise duties relative to targets may be attributed to some inherent weaknesses in revenue forecasting and revenue losses arising from reductions in import tariffs as part of regional trade agreements, mainly the COMESA. 4.9 Domestic interest payments tend to be systematically underestimated; they exceeded budgeted expenditures by an average o f 4 percent during the three years in absolute terms, while all other expenditure categories had shortfalls over the three years. 4.10 Deviations in expenditure by ministerial votes was discussed in chapter 3. Significant variations exist between actual expenditures and the budgeted amounts at the ministry and vote level, with budgeted expenditures higher than the actual expenditures.The variations were more pronouncedin2003/04, when the government undertook major restructuring in the Directorate o f Procurement that resulted in a 42 slowdown in government spending. It is suspected that variation would be much larger at the subvote level, for which data are not available. 4.11 Reasons for deviation between original budget and expenditure outturn touch upon the entire budget cycle, from budget planning and formulation to execution and reporting. The following reasons can be identified: Budgets are unrealistic. This means that resource and expenditure estimates are not realistic. At an aggregate level, an unrealistic budget would generally overestimate resources and, correspondingly, expenditures. As discussed before, in Kenya, revenues are consistently ~verestimated,~as are external resources. Although estimation o f external resources suffers largely from a genuine lack o f information, there may also be a political incentive to overestimate them to artificially increase the size of development spending planned inthe budget4At a disaggregated level, an unrealistic budget might mean that expenditures are underestimated, for example, on individualprograms and projects. The reasons for underestimationcouldbe many. Some expenditures on programs and projects may be underestimatedbecause of incomplete information on costing, and others may be underestimatedbecause of a desire to accommodate more inthe budget. Lack o f sufficient information on the financial position o f SAGAS,public enterprises, and local authorities does not allow advance budget provisions for foreseeable potential liabilities. Incomplete information on arrears or debt overhang could lead to inadequate budget provisions for clearing them. Policy changes take place during budget execution. Examples o f policy changes include unplanned wage and salary increases and new projects and programs (called "parachuted" projects in Kenyan parlance). The instance of initiating unplanned projects appears to have gone down substantially, but projects still are included in the budget before project guidelines and work plans are fully developed. Actual expenditure composition reveals true priorities. Examples would include the Ministry o f Health consistently planning to spend more on rural or primary health care but consistently spendingmore on curative care. Virements are easy. The appropriation structure is highly summarized, not exceeding a few pages, with one spending limit for each ministry. As a result, there is no legal limit for changing o f the appropriations (approved expenditures) within the budget o f a line ministry.This encourages line ministries to apply for frequentreallocation o ftheir budgets (virements) at any timeduringa fiscal year. There are difficulties in cash releases, particularly at the district level. Cash releases are concentrated toward the end o f the fiscal year, and uncertainty about funds becoming available precludes spending in a planned manner by line ministries. Many MPERs have pointed out the issue o f limited cash availability with the district treasuries. Such difficulties would show up more as deviations at the subvote level, for example, the gap between planned and actual spending on school equipment. Reportingon donor-funded operationsis incomplete. Incomplete reporting can create divergence between budgetedand actual (reported) expenditure. This i s true A little overestimationisjustified to create stretchtargetsfor the revenueauthority. Of course, in such a case, final development spendingwill be smaller than inthe original budget, but little public attention devolves around actual development spending compared to the attention at the time ofbudget speech. 43 mainly inthe case of the development budget. However, this part of the deviation has no effect on the reliability of the budget as a guide to future. a Donor-related issues in the development budget also cause deviation. As a result o f conditionalities tied to projects, donors may withhold funds for various reasons in the middle o f the financial year. Some o f the pending bills in the development budget were projects that were abandoned midstream by donors. Also, the government procurement process i s complex and can lead to delays in implementation and, hence, spending. Furthermore, the government procedures have to be harmonized with different donor disbursement and procurement procedures, thus increasing the complexity of management o f development expenditure management. Stock of ExpenditureArrears (PendingBills) 4.12 The existence and continued accumulation o f expenditure arrears i s a long- standing problem in Kenya. Although all budget systems have some debt float at any given point of time, the volume of such arrears i s very large and persistent in Kenya and is a threat to meaningful budgeting. Table 4.2 presents a summary o f the stock o f pendingbills.' The table also shows the proportion o fpendingbills to GDP. Table 4.2. Stock of PendingBills(KSh billion) June 2001 YOof GDP June 2002 % of GDP June 2003 % of GDP Total 13.0 1.5 11.6 1.3 18.3 1.9 Recurrent 5.2 0.6 6.9 0.8 6.5 0.7 Development 7.8 0.9 4.6 0.5 11.8 1.2 Source: QBRs. 4.13 Pending bills constituted between 1.3 percent and 1.9 percent o f GDP during the 2000/01-2002/03 period. Of this, the development component o f the bills constituted 0.5 percent to 1.2 percent o f GDP. Recurrent pendingbills accounted for the balance o f 0.6 percent to 0.8 percent o f GDP. On average, the total pending bills amount to 1.6 percent o f GDP, the recurrent pending bills being 0.7 percent and developmentpending bills, 0.9 percent. 4.14 Between June 2000 and June 2003, the stock o f pending bills increased from K Sh 7.8 billion to K Sh 18.3 billion. The general trend of the stock was upward during the period under review. Except for the year 2002, the stock o f pending bills under development vote has been generally larger than those under the recurrent vote. This can be attributed to the fact that most o f the stalled projects are indevelopment. Interest and penalty charges arising from delayed payments, court awards, and contract variation orders continue to increase the volume o f pending bills. 4.15 The bulk o f recurrent bills are on utilities (mainly water), telephone, and electricity. At the end o f June 2003, the GoK owed about K Sh 6.6 billion to Telkom, Kenya Power and Lighting Company Ltd., the Nairobi City Council, the National Water Conservation and Pipeline Corporation (NWCPC), and the Ministry o f Water 'There are inconsistencies between the stock of pending bills as reported by line ministries, the Treasury, and even contractors. The QBR for Fourth Quarter 2002/03 reports total pending bills at the end o f financial year 2001/02 as K Sh 11.6 billion compared with a figure o f K Sh 21 billion reported by ministriesin October 2002. 44 Resources Management and Development.6InJuly 2003, the Office of the President issued a circular instructing accounting officers to clear the pending bills to utility providers, and to note that it i s within the discretion of the utility providers to disconnect service to customers (including ministries and government institutions). 4.16 There are many reasons for accumulation o f pending bills. One may tend to think of pendingbills as a budget executionproblem, but like the issue of deviations discussed in the previous section, reasons for pending bills span the entire budget cycle, from budget formulation to execution and reporting. The problem reflects unrealistic b~dgeting,~such as not fully costing and hnding the development budget (which, in turn, leads to low completion rates and a stock o f stalled development projects), unplanned policy changes that result innew expenditure duringthe year and expenditure cuts on already budgeted items, a tendency to use appropriations allocated to utilities for other spending items through virements, weak financial compliance (particularly with regulations relating to commitment control) and accountability, weak record-keeping and reporting on arrears, and a weak public procurement system, including weak cyntract design. Pending bills in respect o f donor-funded projects arise as a result o f the expiry or suspension o f the credits before projects are completed, leaving unsettled commitments, delay in releasing the government component o f funding, and variations and claims on contractual aspects that the donors would not ordinarily meet.. The Way Forward on Pending Bills 4.17 The government i s well aware of the extent o f the problem o f pending bills. Several task forces have tried to verify the bills and provided suggestions on how to liquidate them. However, there i s a lack o f a comprehensive and credible plan for their verification and clearing. Settling old expenditure arrears (from before June 2003) should no doubt be a priority, but there i s also an urgent need to implement an actionplan to ensure that new arrears do not accumulate. Among other things, such an action plan would include (a) announcing a transparent and clear policy for clearing accumulated arrears and introducing the rules and regulations to prevent accumulation o f fresh arrears, (b) reviewing contract administration procedures, (c) establishing a project implementation and monitoring system, (d) instituting procurement laws and regulations, and (e) establishing a system for managingpending bills. Systemic IssuesRaisedbythe Discussionon Deviations and PendingBills 4.18 The above discussion provides evidence that there are notable weaknesses in PEM in Kenya that need to be resolved before the budget can be wielded as an effective instrument for strategic linkingof resources with expenditure plans and then spending according to those plans. The weaknesses span the entire budget cycle, and there i s a need to give attention to all three stages o f budget preparation, execution, and monitoring. Inthe past three years, much o f the reform effort has focused on the Some of the utilities also owe tax arrears to the government. As of June 30, 2003, utilities owed the GoK about K Sh 13.3 billion as tax arrears. Apart from tax arrears, there are large overdue amounts of on-lent loans and interest payments. 'Unrealistic budgeting includes confusion in line ministries about full budgeting of accumulated arrears and treating them as an overall government problem rather than a ministerial budgetingissue 45 MTEFprocess, butmost ofthe upstreambudget planning andprioritization will be of no avail ifpriorities are changed ina major way duringbudget execution. 4.19 Fresh technical solutions may not be needed to resolve most o f these weaknesses. It can be appreciated that most o f the reasons cited above also have an element o f "culture" that needs to change. There are two aspects o f culture that need to be addressed. One, the Ministryo f Finance should exercise its role beyond funding line ministries; it should enforce, along with planning, national spending priorities once they have been agreed upon. Two, the budget should be seen as an instrument for making policy choices (which i s the essence of the MTEF approach). Policymaking at present goes on separately from preparing budgets.This culture can change. Such a cultural change can be helped by improving the capacity o f the MinistryofFinance for technical evaluation ofline-item budgetrequests. MPERs can play a significant role in generating necessary information for the Ministryof Finance to probe several selected issues in line ministries. Another important way to support such a cultural change would be to involve political executive and senior management in discussing budget priorities early in the cycle. A budget strategy paper recommended by the MTEF review (discussed later inthis chapter) could be a good instrument to engage political executive and senior management early. BUDGETPLANNINGAND FORMULATION The MTEFObjectives and Process in Kenya 4.20 The GoK adopted the MTEF process in financial year 2000/01 as a new approach inbudget planning inresponse to problems inbudget preparation and public expenditure management. The 2000/0 1 budget speech stated the following four objectives o f the MTEFprocess: 1. Linkingpolicymaking to planning andbudgeting 2. Maintaining sustainable fiscal discipline (that is, aggregate expenditure controls) 3. Facilitating expenditure prioritization acrosspolicies, programs, and projects (that is, "strategic" allocation o f resources) 4. Encouragingbetter use o fresources to achieve desired outcomes at the lowest possible cost. 4.21 Figure4.1 schematically depicts the MTEF and budget formulation process in Kenya as it has developed over the last three years. (The PER process i s also included inthe figure, but it was not a part of the initial MTEF implementation and has been recently overlaid on the existing processes.) The key elements o f the MTEF process as implemented inKenya, are coordinated by the MTEF Secretariat inthe Ministry o f Planning and NationalDevelopment (MPND). 1. All line ministries and other departments are classified into eightbroad sectors along functional and thematic lines. Eight Sector Working Groups (SWGs) have been formed. The terms o f reference o f the SWGs are extensive (see annex 5). 46 D T' 3 I !!!!!!!!!!!!!!!! ! n +! c 0 2. After a Treasury Circular in early February launches the MTEF budget process, line ministries prepare MTEF submissions for the SWGs. The level o f detail in the submissions varies by ministry, but they typically contain financial plans for three years and identifypriority spendingareas. Line ministries are expected to carry out a careful review of the ongoing programs and activities and also cost them before making these submissions. 3. Separately, during February and March, the Macro Working Group in the MPND prepares a Fiscal Strategy Paper (FSP) that includes a three-year, medium-term, macroeconomic and fiscal framework to form the basis o f forecasts o f tax revenues and grants. The FSP proposes an aggregate resource envelope for the next year's budget based on aggregate resource estimates and fiscal objectives. 4. The SWGs prepare Sector Reports based on line ministries' MTEF submissions, taking into account inter- and intrasectoral linkages. Sector Reports include a financial plan and priority programs and activities for the sector. 5. Sector Reports are presented to stakeholders in sector hearings organized in mid-March. Nongovernmental organizations (NGOs), development partners, and members o f Parliament are invited to these open hearings. Hearingsnormally last for two days (about two hours for each sector). 6. Based on the feedback received at the sector hearings, and taking into account the FSP, sector ceilings are developed by the Treasury and conveyedto the SWGs and line ministries. 7. SWGs then organize sector resource bidding(toward the end o f March), in which line ministries lay claims to shares o f resources within the sector ceilings conveyedby the Treasury. 8. Ministerial ceilings are developed based on the sector resource bidding and conveyed by the Treasury to line ministries in early April. (For the financial year 2004/05, there was no formal conveyance o f ministerial ceilings) 4.22 The conveyance o f ministerial expenditure ceilings to line ministries completes the MTEF process and line-item budgetingtakes over. Line ministries submit three-year line-item budget estimates to the Treasury, which consolidates them into the budget duringthe last week ofApril. The budgetgoes to Parliament inthe first week o f June. 4.23 MTEF implementation varies from country to country. In Kenya, the MTEF is seen as both a process and an output. As a process, it i s all the eight steps discussed above-a process linking planning and budgeting within a medium-term resource framework. It i s the function o f the MTEF process to provide planning inputs to the annual budget through developing medium-term priorities and apportioning ceilings to ministries. As an output, it is a three-year, rolling public expenditure plan at the level o f eight MTEF sectors-a tool for financial planning and management. 48 4.24 Kenya didnot adopt an activity-based MTEFthat sets up the policy objectives- outcomes-outputs-activities hierarchy and then develops costs for achieving objectives. This is, however, not a weakness o fthe process. At this stage ofMTEFdevelopment, the GoK perhaps does not need activity-based costing. There are basic deficiencies that must be addressed before advanced costing schemes are introduced. Evidence of Weaknessesof the MTEF Process 4.25 A well-functioning MTEF process would provide a steady guiding handto align budget allocations with ERS priorities. However, as discussed in chapter 3, the current MTEF process has relatively little impact on budget allocations. Other than in the education sector, which has seen large shifts in the last year, there has been a drifting increase inpublic administration spending, which shows that inertia persists inbudgetary allocations. Resource predictability remains poor for line ministries. Ideally, the MTEF would be expected to improve resource predictability in the medium term, but the first sign o f that would be improved predictability in the annual budget. True, much o f the unpredictability arises from an unrealistic resource envelope to beginwith, but gettingthe resource envelope right i s very much an objective o f a well-functioning MTEF. The MTEFprocess is not taken seriously enough by line ministries, perhaps because they do not find it relevant to their budget estimates approved by the Treasury. An indication o f lack o f sufficient seriousness i s thinattendance at sector hearings. There i s no evidence o f the effectiveness o f stakeholder participation duringSWGs andsector hearings. 4.26 The above problems have acquired urgency because expenditure restructuring, in consultation with stakeholders (including development partners), i s a cornerstone o f the government's fiscal strategy. With several sectors developing their new development strategies, it will be important that resources flow where the policy goes and such changes are implementedwithinthe framework of MTEFbudgeting. Findings and Recommendations of the Recent Review 4.27 The GoK has recently completed a review o f the MTEF process in Kenya.* The review makes a number o f recommendations to improve the MTEF process. The government i s in the process o f internalizing these recommendations before their implementation. Some o f the important findings and recommendations are elaborated and analyzed below: (a) The MTEF and annual budgeting remain two separate processes. The MTEF is seen as an adjunct to the existing annual budget preparation and has not replaced the annual budget formulation process. Once the MTEF outputs have been produced, annual budgeting takes over inmuch the same form as it hadbeen before the implementation o f the MTEF. (b) Institutional issues. The Planning-Treasury split between the MTEF process and the annual budget is not conducive to effective integration o f spendingpriorities and resource allocation. This split carries through inline ministries as well; MTEF submissions are made by the planning units o f line 49 ministries, and budget submissions are made by the budget units o f line ministries. (c) Absence of a single budget strategy document. A budget strategy document that translates all core government policy statements into coherent medium-termexpenditure plans, within the aggregate resource constraint, is a key component o f an MTEF process.' Such a document provides a fairly firm reconciliation o fresources and priorities before line-item budgeting.In Kenya, the role o f a budget strategy document i s played by three (sets) o f documents: the FSP, the guidelines accompanying the budget call circular, andthe Sector Reports. However, these documents are not a good substitute for a single budget strategy document because they are produced inparallel without much integration. Moreover, the FSP focuses more on macrofiscal issues and development o f an aggregate resource envelope. It does not include much discussion o f intersectoral priorities or links betweenthe ERS and the proposed expenditure strategy. As discussed before, the budget guidelines have very little strategic content, and Sector Reports, themselves quiteweak, do not formally feed into the budget-making process. The review suggests two documents that would lay out the government's budget strategy. The first, a Budget Outlook Paper to be produced in late September or early October, would detail the macroeconomic and budget outlook and the ensuing year's proposed strategy, with proposed sector ceilings (which could well be the previous budget's outer years' ceilings)." The secondpaper, called by the review an "Enhanced Fiscal Strategy Paper" (referred to henceforth as a Budget Strategy Paper) i s to be produced in early January after the consolidation o f sector proposals (Sector Reports) and based on a reasonably firm resource position. Such a document would need to be sufficiently detailed to capture the key elements o f government's principal policy commitments in the ERS and ministerial strategy papers and have ceilings sufficiently disaggregated to establish a link to objectives o fprograms inline ministries. It would also provide the ideal instrument for discussion with stakeholders on issues, choices, and tradeoffs that need to be undertaken. The Budget Strategy Paper could help harmonize the dialogue between development partners and the government on public expenditure issues by focusing discussion around one key paper that will be binding on the subsequent budget. 9Note that beyond Sector Reports, the MTEF process does not produce a unified document that could lay down intersectoral priorities, or revised priorities within sectors after Sector hearings, that could be approved by senior management or the political executive. In the absence o f such a document, Sector ceilings decidedby the Treasury involve implicit priority setting across sectors by the Treasury lo Note again that presently the MTEF submissions by line ministries-and, consequently, Sector Reports-are prepared without any indicative ceilings. 50 (d) Lack of political engagement in budget strategy and sector ceilings determination. In the existing MTEF and annual budgeting process, the Cabinet gets involved mainly at the stage o f approving the line-item budget before it goes to Parliament." The lack o f upstream political engagement in the budgeting process, including the MTEF process, i s a critical weakness that contributes to derivative problems, such as inadequate commitment o f line ministries (at both political and senior management levels) to the MTEF budget. Inadequate commitment at the budgeting stage inturn leads to poor execution, with a tendency to initiate new policy proposals and reallocate resources across programs during execution. The review recommends that the Cabinet approve the budget strategy document(s) so that they are bindingon lineministries. The review's recommendation of early political engagement in the budget strategy through a Cabinet decision appears sound and is in line with the practice in other countries implementing the MTEF approach to budget making. Cabinet discussion and endorsement o f the budget strategy would be one way to get political engagement in the MTEF and pose hard policy choices ina collective action context. (e) Compressed budget formulation timetable. The review finds that the timetable for the budget cycle i s compressed. As it operates now, it starts too late, and the MTEF process provides ministerial ceilings far too late in the budget cycle, leaving very little time for line ministries to implement strategic choices in the detailed budget submission. As a result, ministries appear to fall back on incremental budgeting within the overall ceilings provided by the MTEF process. The review recommends that the budget cycle should start early in September or October with the Budget Outlook Paper that would also provide indicative sector ceilings. These ceilings will help line ministries prepare their MTEF submissions (or MPERs) within an indicative budget constraint. The Budget Strategy Paper produced in January would then give line ministries enough time to produce detailed budget estimates. (f) Functioning and constitution of SWGs. SWGs could be useful institutional constructs to serve the important functions o f prioritization across ministries and provide a forum for stakeholder inputs. However, the review found that they are not effective in getting stakeholders' inputs. There i s a need to clearly think through the role and functioning o f SWGs. Their current terms ofreference are too broad (annex 5) and needto be more narrowly defined, considering that SWGs are coordination instruments and not self-standing government bodies with regular staff or a secretariat. At Although sectoral ceilings are seen by.acommittee o f the Cabinet before issue o f the itemized budget call circular, the strategic element o f the circular is l o w and does not draw upon the priorities developed in parallelby the Sector Reports. 51 the same time, discussion of ministerial submissions (sector hearings) could be much enhanced. There i s also a need to ensure systematic interface with stakeholders in SWGs, outside the budget cycle. For example, SWGs could hold discussions on ministerial development plans and strategies. (g) Sector ceilingsversus ministerialceilings.Currently, ceilingdetermination is a two-step process. Sector ceilings are developed first, then apportioned into ministerial ceilings. The value added by a two-step process i s suspect. The crucial issue i s that the unit of accountability, budgeting, and even planning or strategy formulation is a line ministry. Therefore, the budget strategy paper proposed to be approved by the Cabinet should contain ministerial ceilings. Much o f the legitimacy being sought through the Cabinet process will be lost if the Cabinet approves only broader sector ceilings and not ministerial ceilings. (h) Ceilings are not binding. The Treasury should enforce the ministerial ceilings. MTEF Budget Formulation in Relation to thePER and Other Processes 4.28 After the GoK adopted the MTEF process, it was overlaid with a PER process. Another process for monitoring and evaluation o f the ERS may be developed soon. There i s a need to harmonize the MTEF budget process, the PER process, and monitoring and evaluation o f the ERS. The objective should be to arrive at a clear, seamless process that i s efficient and sustainable. The contribution o f the PER process to budgeting would be mainly through ex-post review o f spending, analysis o f equity and efficiency or effectiveness o f government policies and programs, and projection o f resource requirements after accounting for policy changes.12 However, some issues need to be resolvedto better harmonize the PERwork with MTEFbudgetprocesses: 0 The timing ofthe PER work needs to be carehlly synchronized with the MTEFandbudget cycle. Inthis regard, itwould beuseful to beginwork on MPERs immediately after the issue o fBudget Outlook Paper inlate September or October. 0 Natural entrypoints need to be identified for PER outputs to inform budget decisions. Inthis regard, MPERs would be naturally suitedto be discussed by SWGs inplace o f Sector Reports. 0 To the extent possible, the PERwork and MTEFbudgetprocesses should use the same institutions (such as SWGs, the Steering Committee, and so on) to reduce duplication. l2 The PER process itself has several objectives other than informing budget formulation through expenditure analysis. The PER 2004 Concept Note stated background analyses o f cross-cutting policy issues such as contingent liabilities and the wage bill, monitoring the progress o f reforms in budgetary processes and institutions through a medium-term reform program, and providing an instrument for continuous dialogue with development partners as other key objectives o f the PER work. Once the Budget Strategy Paper takes strong hold, the role o fthe PERprocess indialogue with donors i s likely to go down. 52 MergingofMTEF submissions andMPERsinto one document shouldbe explored. A Revised MTEF Budget Cycle 4.29 Figure 4.2 provides a schematic of a revised budget cycle incorporating these recommendations. The key changes proposed inthis budget cycle are: Preparation of a Budget Outlook Paper in September or October. (This i s a new step.) Early preparation of MPERsduring October to December. Developing a Budget Strategy Paper using inputs of the PER process (MPERs and any analytical studies) in early January. (This i s a new step.) The Budget Strategy Paper should be approved by the Cabinet and include ministerial ceilings. Dissemination of the Budget Strategy Paper to stakeholders and civil society. OTHER ISSUES INPUBLICEXPENDITURE MANAGEMENT Extrabudgetaiy Sources 4.30 General govemment activities fbnded through inadequately reported extrabudgetary sources (earmarked revenues o f SAGAs and fund accounts) are estimated at about 10 percent of the central govemment budget.13In April 2004, there were 75 SAGAs engaged in activities related to regulatory functions, research, regional development, education, health, and other social services; several o f them include poverty-reducing expenditures. l3Earmarked revenues (mainly fees and charges) collected and spent by line ministries are incorporated into their budget and reporting systems. Therefore, expenditure transactions financed through these reveaues should not be regarded as inadequately reported expenditures funded through extra budgetary or off-budget sources. 53 -- L tt . 1 I F' B E t . .A . _I At the same time, there were there were 36 fund accounts (annex 6 provides a list o f fund account^).'^ A number o f SAGAs are not included in the budget documents, except for the amounts that are transferred to them from the government budget. Since January 2004, some SAGAs have begun reporting their transactions to the Ministry o f Finance. This should be regarded as positive development, but it appears that until now, no mechanism has been worked out to reconcile these useful reports, which inpart contain extrabudgetary financing from earmarked revenues and occasionally external aid, with reports o f the central government budget transactions (table 4.3). Included inbudget document Regular and detailed reporting Transfers Earmarked Expenditures Transfers Earmarked Expenditures revenues revenues SAGAs Yes N o None Started for Startedfor Started for some some some Fund Yes Some None None None None accounts 4.31 Fundaccounts are credited with lump-sumtransfers from the government budget and by earmarked tax and nontax revenues. These funds are not separate organizations, but simply accounts that are operated by the line ministries outside their budget. Expenditures o f these fund accounts that are not financed through transfers from the budget, but rely, at least inpart, on earmarked revenues, are considered to be funded from extrabudgetary sources.15This i s because the earmarked revenues are not included inthe budget. Although the Ministryo f Finance has now made an inventory o f fund accounts, it does not collect reports o f these accounts, and, therefore, i s not able to prepare a consolidated report for all transactions o f the general government. 4.32 SAGAs and fund accounts keep a sizable volume o f expenditure outside- and off- budget. There is a need to conduct a study into the history and necessity o f SAGAs and fund accounts with a view to reducing their number to a minimumrequired by law and accounting needs and then integrate the remaining SAGAs and fund accounts into the budgets o f the line ministries. l4Examples o f SAGAs include the Kenya Airport Authority, KRA, Kenyatta National Hospital, and all universities. Examples o f fund accounts include the Government Press Fund, Veterinary Services Development Fund, Road Maintenance Levy Fund, and Health Care Services Fund. Fund accounts constitute part o f the annual reports o f the Controller and Auditor General (CAG), though reports are not submittedina standard format and are normally delayed. l5Examples include the Health Care Services Fund and Veterinary Services Development Fund, which receive user fees from various SAGAs and other government operations and are not reported inthe central government budget. b 55 Budget ClassiJication 4.33 The current budgetary classifications inuse inKenya date from the 1960s. Efforts to reform the classifications have only recently started, and they are still in their initial stages. The existing economic classification i s inadequate for both budgetary and accounting purposes. Apart from the lack o f sufficient detail to plan, budget, and account for transactions, the use o f coding i s inconsistent across line ministries. These deficiencies are fully and widely recognizedwithin the Ministryo f Finance. 4.34 To date, the Ministry o f Finance also has not adopted a functional classification of government operations. Initialwork on mappingbudgetary operations to the Government Finance Statistics (GFS) 2001 Classification o f Functions o f Government (COFOG) was undertaken by an IMF statistics mission in October 2003. Reporting on the basis o f finctional classification will significantly contribute to the analysis o f government operations. 4.35 The existing organizational classification requires improvements in the area of SAGAs. These institutions are grouped together with state-owned enterprises, under the State Corporations Act. The institutional classification should differentiate between institutions providing a public service, generally not profit-oriented, and enterprise-like bodies that are, inprinciple, established on a commercial, primarilyprofit-oriented basis. The former group is part o f general government, andthe latter are part o fthe wider public sector, according to widely accepted GFS principles. This differentiation i s important for budgetary purposes because nonprofit government institutions receive operating grants from the budget, and state enterprises will receive subsidies from the budget should their incomes not cover their operating costs. 4.36 The PEM-AAP2004 recommended that the government develop and implement for the central government budget a new economic classification structured according to the GFS 2001 framework. It also recommended improving the organizational classification to clearly separate those enterprise-like bodies that belong only to the public sector. In the medium term, the new economic and functional classifications should be extended to local authorities, fbnd accounts, and SAGAs. 4.37 The GoK has been addressing weaknesses in the processes through which the budget is executed, but problems still exist in cash planning, procurement, and commitment control. The weaknesses have been manifested in problems such as large stock of pending bills, stalled projects, weak commitment controls, and wastage within the procurement system. 4.38 The budget continues to be executed through Authority to Incur Expenditure (AIE) based on quarterly ceilings. The Exchequer Committee meets weekly to review exchequer requisitions and determine the cash releases. Line ministries report that a mismatch between exchequer releases and AIEs undermines budget implementation. A closer examination of the issue o fweekly exchequer releases i s required. 56 4.39 Currently, there are inadequate regulations on virements and reallocations. The government should review the guidelines on virements to minimize reallocations that take place after the budget has been presented in Parliament. There i s a tendency for ministries and departments to put more resources into areas considered as core and later request reallocations from the same areas. Although some adjustments to the budget may bejustified, the excessive use o f supplementarybudgets and reallocations undermines the credibility o fthe budget process. 4.40 Commitments are controlled through the "Vote Book" system, which provides running balances as well as outstanding commitments. Weaknesses exist mainly in recording and reporting. One major challenge duringthe budget execution stage has been time lags between commitments and associated cash payments. This to some extent accounts for the backlog in payments (pending bills problem). The Treasury has noted this problem and has instructed ministries and departments (through Treasury Circular 15/2003, dated July 31, 2003) to ensure that expenditure commitments are guided by the laid-down financial regulations and are kept within the quarterly ceilings. Public Procurement 4.41 A sound and effective procurement system plays a critical role in enhancing the effectiveness and efficiency o f public expenditure through maximizing the value o f money spent and the quality o f goods and services rendered. However, problems such as overpricing, lack o f competitive bidding and tendering, and weak compliance with procurement regulations have underminedprocurement. 4.42 A new procurement regulatory framework was gazetted in2001 to govern public sector procurement. These regulations were designed with the objectives o f achieving economy, efficiency, transparency, and accountability in the procurement system. The regulations were based on the Finance Act, with no specific legislation to underpin them. A new Public Procurement and Disposal of Assets Bill has been published, aimed at providing the legal framework to underpinthe regulations. 4.43 Notwithstanding the new regulations, amidst serious concerns about impropriety and lack o f transparency and accountability, the government took the bold step o f suspending all supplies staff, pending a special vetting process, inMay 2003. The mass suspension nearly paralyzed the procurement system, but it was followed by a careful vettingbefore reinstatement and new recruitment processes that ended in2004. 4.44 The government is committed to further improving public procurement. A Country Procurement Assessment Review is being considered for 2004/05. Budget Monitoring 4.45 Donor-financed operations are not comprehensively reported on because o f incomplete information at the time o f budget preparation. Therefore, difficulties arise in tracking disbursements, particularly in the case o f direct payments made under grant funding. 57 4.46 The GoK is in the process o f institutionalizing a national monitoring and evaluation framework for the ERS. With the introduction o f the Constituency Development Funds (CDFs), there i s needto design and implement a participatorypublic expenditure monitoring and evaluation system that will allow communities to participate in expenditure tracking and performance monitoring. The Constituency Development Committee has been giventhe responsibility o f implementingprojects under the CDF Act 2003. The District Documentation Centres need to be strengthened and equipped with relevant information on social and economic indicators for purposes o f informing their communities and creating awareness to enable them demand services and question the outcomes o f any expenditures meant for them. Expenditure Control and Auditing 4.47 The Internal Audit Department has already been restructured, and a deliberate policy o f phasing out preaudit activities and focusing on system-based audit has been adopted. However, there i s resistance to the phasing out o fpreaudit activities from a wide variety o f stakeholders, including extemal auditors. Problems in enforcing follow-up of recommendations at the ministrylevel, including independence o f audit committees, also exist. 4.48 A number o f countries neighboring Kenya have initiated PETS to estimate the amount o f public money that i s actually delivered by the budget system to frontline service delivery units and, through them, to final users. These surveys provide a double check on the quality o f internal control systems and are more sharply focused than regular extemal audits. 4.49 The GoK has also initiated PETS. The Kenya Institute o f Public Policy Research and Analysis (KIPPRA) has conducted PETS on a pilot basis in agriculture, education, and health ministries. The KIPPRA exercise has found anecdotal evidence o f leakages in the flow o f public resources. For instance, within the Ministryo f Health, some drug kits were found to have never reached rural centers; in the Ministry of Education, "ghost workers" were identified inprimary schools. The surveys also found that inadequacy o f resources has seriously hampered the operation o f extension services in rural areas. The government has decided to build on the KIPPRA exercise and conduct PETS in a more systematic way. A PETS i s also being plannedby the Ministry of Education. 4.50 As shown in Table 4.4, significant efforts have been made in recent years to eliminate the backlog in issuing audit reports on appropriation accounts. However, audit reports for the central government budget are still not being issued within the statutory period. 58 ITable 4.4. Status ofAudit Reportsfor AppropriationAccounts andOther Funds Financial year Status Comments 1998199 Public Accounts Committee PAC is currently reviewing the (PAC) hearings and reports for both 1998199and discussions are taking place 199912000.Reports are expected now. shortly because the current PAC i s due to bereplaced inJune 2004. 199912000 PAC hearings and discussions See above. are taking place now. 2000101 Tabled inParliamenta Not yet reviewedby PAC. 2001102 Signedby the CAG. 2002103 Scheduled for completion. a. Once the report i s tabledinParliament, it becomes a public document. 4.51 The majority o f local authorities have also not prepared accounts for audit purposes for many years. Since 2000, local authorities have been submitting abstracts o f their accounts, including statements o f income and expenditure, as well as assets and liabilities, to meet the requirements for the provision o f central government budget transfers. Untilrecently, however, some local authorities had not prepared their accounts since the mid-1960s. Therefore, the C A G i s not able to make a comprehensive auditing o f the accounts because o f the problem o f verifying the opening balances. Moreover, according to the CAG, the majority has not been preparing accounts as set out in the relevant local government legislation. Although some have employed local accounting firms to prepare backdated accounts, others have not taken action.16 Against this background, it i s clear that a pragmatic solution i s required, and the C A G has proposed placing a special report before Parliament to advise them o f the situation. 4.52 The audit by the C A G o f SAGAs has been delayed by both nonsubmission o f accounts and internal operational difficulties. The merger between the CAG's office and the office o f the Auditor General for State Corporations, which also covers SAGAs, caused a number o f operational difficulties. 4.53 Lack o f resources i s cited as one o f the main reasons for the backlog o f audits. However, in addition to the CAG's work force o f more than 790 personnel, the new Public Audit Act allows the use o fprivate sector firms to carry out audits on behalf o f the CAG; the only constraining factor presumably being adequate financial resources. The existence o f external auditors conducting continuous audits undermines the effectiveness o f internal control systems. In addition, it diverts limited resources, and the adoption o f l6According to information provided by the CAG, 1,183 accounts have not been produced by the local authorities. Two hundred fifty accounts have been prepared by local authorities but cannot be audited becauseof lack o f supporting documentation and schedules, and 272 accounts are currently being audited. 59 more modem auditing techniques could significantly reduce the time and cost o f individual audits. 4.54 The real effectiveness o f the oversight hnction carried out by the CAG depends not only on timeliness but also on the extent to which its recommendations are implemented, as well as the independence of the CAG in carrying out his or her functions. The 2003 Public Audit Act establishes the Audit Services Commission and the Kenya National Audit Office, providing improved autonomy in terms o f work force and financial resources. However, Kenya does not meet a number o f key indicators in terms o f independence, as set out by the International Organization o f Supreme Audit Institutions (INTOSAI). SUMMARY OFRECOMMENDATIONSAND ACTION PLANFOR PEMREFORM 4.55 After the update of the CFAA and PEM-AAP assessments in 2003, the GoK developed the Enhanced Financial Management Action Plan (EFMAP). This action plan focuses on improvements inbudget formulation, execution, and reporting, in addition to considering institutional and humanresource development issues and the reform o f state corporations. The progress inimplementingthe reform program was recently reviewed in the context o fthe PEM-AAP update inApril 2004. 4.56 The progress on implementation o f EFMAP has been slow. The PEM-AAP update noticed little progress except initiation o f action in some areas. The update has produced a revised action plan that proposes a gradual, time-bound implementation schedule for actions consistent with existing capacity. However, stronger progress in deliveringEFMAPmight requiremore attentionto the institutional coordination issues in delivering key actions. In this context, a dedicated coordinator for EFMAP should be considered. Inaddition, more details on the activities falling outside the immediate scope o f the Ministry o f Finance and MPND (for example, external audit, payroll management, and so on) should be incorporatedinEFMAP. 60 5. MINISTERIAL EXPENDITURE REVIEWS' MINISTRYEDUCATION, OF SCIENCEAND TECHNOLOGY (MOEST) Aggregate Expenditure 5.1 Current education spending as shown in table 5.1 is highboth as a proportion o f GDP and as a share of recurrentpublic spending, rangingbetween 6 and 7 percent and 35 and 40 percent, respectively, during the period under review. Kenya's spending in this area i s high by world standards and especially by regional standards (see table 5.2); for example, Tanzania and Uganda spend less than 3 percent o f their GDP on public education. 5.2 From another perspective, education subsectoral shares o f recurrent spending translate to about 51 percent, 28 percent, and 11 percent o f spending on primary, secondary, and university levels, respectively. Recurrent spending accounts for the bulk o f total expenditure in this sector; the development budget has remained insignificant by comparison. Recent donor assistance for free primary education (FPE) in the development budget has, however, tended to reverse the trend. Table 5.1. Total Actual MoEST Expenditure, 2000/01-2003/04 2000101 2001l02 2002103 2003/04a MoEST expenditure (K Sh billion) Recurrent 48 53 61 72 Development 3 1 3 8 Total 51 54 64 80 Shares intotal (%) Recurrent 94.1 98.1 95.3 90.0 Development 5.9 1.9 4.7 10.0 Total MoEST exp. as % o f total GoK exp. 16.5 17.2 18.7 20.1 Total expenditure as % o f GDP 6.1 5.8 6.3 7.1 a. Estimates. Source: GoK annual estimatesand appropriationaccounts; EconomicSurveys, various years. 'This chapter i s largely based on information contained inMPERs. Inmany instances, however, an attempt was made to fill gaps indata and analysis from other sources. Data reportedby ministries in MPERs do not always match the BMD data inQBRs. This canhappenbecauseAIA are reported with a lag and may not get included inthe reports by line ministries to the BMD. Also, BMDmay at times adjust line ministries; reports if they are too much out o f line with exchequer releases. For the above reasons, expenditure data inthis chapter may not always matchthe data inthe statistical annex (which i s based on BMDreports). 61 Table5.2. PublicEducationSpendingaroundthe World, 2000 Country Public spending on education as % o f GDP Botswana 8.6 Ghana 4.1 Kenya 6.1 Malaysia 6.2 Republic ofKorea 3.8 SouthAfrica 5.7 Tanzania 2.2 Uganda 2.5 Source; World Development Indicators,World Bank. 5.3 Public spending per student at the primary, secondary, and tertiary levels, as reported in the MPER, are K Sh 4,400, K Sh 21,800, and K Sh 135,800, respectively. There are some questions regarding the data (for example, revenues retained by the schools do not appear to be included); nonetheless, the ratios o f secondary to primary spending per student (4.95) and o f tertiary to primary per student spending (30.9) appear relatively high. PrimaryEducation 5.4 Performance. With the introduction o f FPE in 2003, the subsector recorded an overall high gross enrollment rate o f about 103 percent for public, private, and community schools and a significant improvement in girls' participation, at nearly 50 percent. Inaddition, there were increased completion rates, reaching 60 percent. 5.5 However, enrollment rates show great disparity across regions and gender, being particularly low among girls in arid and semiarid lands (ASAL). Inaddition, output and quality assessment studies reflect problemso f quality inteaching and learning. 5.6 Handlingthe increasedenrollment. After introduction o f the FPEpolicy and a sharp rise in enrollment rates, the need for teachers, classrooms, and educational materials and textbooks has risen. This has created difficulties as Kenya has tried to meet these increased demands. 5.7 Teachers. At present, the aggregate number of teachers is not too low; the average national pupil to teacher ratio inprimary education i s manageable at about 40:l. Thus, thejustification for hiringlarge numbers o f new teachers is weak, given their high salary levels that cannot be afforded. There is, however, an issue o f allocation, because there i s significant variation in pupil-teacher ratios across the country; district averages range from 11:1to 100:1, and there are vast differences between schools within the same district. Note, however, that the solution is not trivial: Some very low pupilto teacher ratios occur inareas where the denominator is low because o f sparse population, low enrollments, or both, thus the scope for freeing 62 5.8 Overall, there are three options: reallocate existingteachers, increase the teaching cadre significantly, or temporarily hire teachers locally inareaswhere there are shortages. It is recommended that a quantitative analysis be undertaken in the 2005 PER that will draw on the forthcoming staffing norms study, look at the various options and their costs, and quantify more precisely the long-term (as opposed to temporary) requirement for increased teacher numbers. Inmany cases, the needs will likely have to be assessed on a school-by-school bask3 5.9 The analysis should include examining the possibility o f the decentralization o f teacher recruitment, as well as management structures and incentives that facilitate a more efficient use o f teachers. The results o f this analysis are needed to optimally address the shortage o f teachers. In the meantime, increased fbnding is probably justified for temporary teachers in areas where there are excessively high student-teacher ratios, as well as for in-service training for existing teachers to adjust to the new curriculum. 5.10 Teacher salaries in Kenya are high by international standards (5 times per capita GDP, compared with an average of 3.5 times for all developing countries). Direct comparisons can be problematic because o f differences in the cost o f living and labor markets, and higher salaries do help with teacher retention and motivation; this nonetheless translates into average teacher salaries o f about US$2,000 equivalent annually, limitingthe number o f teachers that can be employed. The large salary bill also limits the flexibility of MoEST to use resources for other purposes. It is likely impossible to reduce salaries in the short term, but a long-term plan for adjusting teacher numbers and allocation probably also needs to involve some element of capping nominal salaries and allowing their real value to adjust over timea4 5.11 The impact of HIV/AIDS on teachers. Comprehensive data are not available, but it is widely acknowledged that many teachers are absent because they have a long- term illness. For a school with only four or five teachers, an absence o f one or two because o f HIV/AIDS (or any other illness o f disability) creates very real difficulties, but there are no specific provisions to help such schools. This is an issue that has been repeatedly raised inthe government's MPERwork, suggestingthat provision be made to give head teachers funds to hire temporary teachers when they are short o f staff because o f long-term sickness. 5.12 Classrooms.There is no doubt that overcrowding has occurred as a result o f the FPE expansion. At an aggregate level, there does not appear to be a massive shortage o f classrooms; assuming 40-45 students per class would imply a need for about 170,000 upteacher fromthese areas is limited. Conversely, the relativeunattractivenesso f some amras may make it difficult to physically reallocate existing teachers. These factors need to be critically assed in the 2005 PER. The TSC i s reportedly preparing an analysis of what has been done to reallocate teachers, but it i s not reflected inthe MPER. Other elements of the rationalization o f salaries likely include re-examining pay scales, so that graduates working as primary teachers are paid only the same amount as other qualified primary teachers, or ensuring that heads o f departments at the secondary level have few additional responsibilities, enabling them to handle larger teaching loads while receiving smaller allowances. 63 classrooms, and there are currently 191,000, so the needfor new classrooms i s localized. Also, the gross numberstell us nothing about the condition of facilities, many o f which- about half, by some estimates-are reportedly inpoor repair and need rehabilitation. 5.13 Unfortunately the data do not allow a meaningful quantification o f expenditure requirements. According to the 2004 MPER for MoEST, the ministry (a) does not know where the need for classrooms is, nor its scale; (b) does not have a firm unit cost estimate for classroom con~truction;~and (c) lacks clarity on policy for cost sharing on new construction and rehabilitation. In view o f this lack o f information, there i s an urgent need to process the 2003 school census data6complete school mapping to identify needs, and then come up with prioritized and costed list by the time o f the next MPER and annual budget. In the meantime, efforts should focus on urgent needs and facilities in communities that are too poor to contribute from their own resources. 5.14 The Capitation Grant and Equity. With the shift to FPE, schools are given a grant to finance materials and textbooks, as well as for other spending (minor maintenance, nonteaching staff). The experience so far has been quite successful; and because expenditures o f this type were formerly largely covered by households through the payment o f school fees, there i s a positive equity effect. The total allocation for such capitation grants amounted to K Sh 9 billion in the 2003/04 budget, representing 12.5 percent o f the recurrent education budgets7 5.15 The amount that needs to be spent on capitation grants in the long term is not entirely clear. Although there i s a needfor continued relatively high levels o f spendingin the medium term to catch up with current deficits intextbooks, and to accommodate the costs o f shifting to a new curriculum, the MPER does not suggest that the current level o f the capitation grant is inadequate, and the equilibrium level might be about K Sh 6-7 billion per year. However, in the initial years, a large part o f the capitation grants has been funded by the DFID and World Bank, and these clearly need to be worked into regular recurrent budgets to ensure the predictability o f funding over time. There is also some concern with the additional need for monitoring and reporting on the use o f these funds (see discussion of expenditure management issues below). 5.16 Teacher Training Facilities. Several documents suggest underuse o f capacity; the current teacher training college (TTC) capacity i s 9,000, but output is just 7,500. The MPER and other documents report that output exceeds the hiringneeds o f the ministry, and there is a pool o f unemployed teachers. Although there is some inconsistency in the This has occurred inpart because there is little experience outside donor projects, which have varying costs and are often executed outside the ministry. For the last few years, the school census questionnaires that have been returned have not been processed. Thus, the information could have been available relatively easily, but MoEST has not had the resources, nor assigned priority, to do the work needed to produce it. The same questionnaires will also provide up-to- date repetitionand dropout data. 'There are major inconsistencies in the data: The MPER reports K Sh 3.5 billion in 2002103 as a supplementary budget and K Sh 9 billion annually from 2003/04 onward. This appears to exceed the primary education non-wage recurrent budget (table 2.2 inthe MPER). Elsewhere, the MPER implies the need for approximately K Sh 6 billionper year, which is imputed o n the basis o f K Sh 1,020 per capita and approximately 6 million students. 64 data presented, it appears that depending on how the recruitment and reallocation are handled, Kenya should not need muchmore than replacement o f attrition for some years. 5.17 Inthis context, too, much TTC capacity, and plans to expand, mean too much is spent on recurrent costs o f pre-service teacher training and proposed investment costs o f projects to improve TTCs. Ideally, it would be better to have only just enough teachers, rather than capacity to produce 50 percent more than needed. There may even be scope for reducing the number o f TTC facilities to generate cost savings. Calculation o f optimal TTC capacity and spending should be included inthe next MPER. SecondaryEducation 5.18 There i s considerable emerging pressure to substantially expand total secondary school enrollment, which currently stands at about 840,000 (with about 5-7 percent attributed to private schools). This pressure stems from the larger number o f studentsthat will be passing through primary school following the introduction o fFPE, combinedwith the government's stated policy intent o f raising the share o f students passing from primary to secondary to 70 percent (from a current level o f about 50 percent).* 5.19 It will be difficult for the government to afford this increase in enrollment, given current funding sources and levels, Funding needs in the near future will depend, in particular, on issues related to teacher use, bursaries, the government's share o f recurrent funding, andthe policy for establishing new schools and expanding existing ones. 5.20 Teachers. The current secondary school student-teacher ratio is 17:l. Itprobably needs to be increased to about 25:l or higher over time. The current ration is not affordable today, and it will be less so inthe face o fthe pressure for expansion. 5.21 The MPER estimates that an additional 12,000-20,000 secondary teachers will be needed over the coming three years; and this i s regarded as a minim~m.~ There i s some dispute about the numbers and the assumptions that should be made regarding reallocation, and the situation i s complicated by the genuine needs for certain kinds o f teachers-in mathematics, english, and science, for example-that could not be met by reallocation. Part of the solution i s a shift away from the current practice o f curriculum- based teaching staff strength. As with primary education, the forthcoming staffing norms study i s critical. * As enrollment in standard 8 moves up !?om the present rat, say, 50 percent of gross enrollment rate toward, say, 80 percent, as expected-a move from 50 percent transition to 70 percent implies an increase in gross enrollment rate in form Ifrom 25 percent to between 55 percent and 60 percent, more than doubling secondary enrolment, particularly as population growth continues. According to the MPER, this i s based o n 180 streams and 8.3 teachers per stream. 65 5.22 Bursaries and cost sharing. The level o fcost sharingi s highinKenya compared with other low-income countries (with average annual secondary fees o f K Sh 15,000), implyingthat as much as two-thirds of the costs are being recovered." However, these highfee levels present problems of access andaffordability (K Sh 15,000 is equivalent to US$197, when average per capita incomes are only about US$400). To compensate, the government operates a bursary scheme for "poor but bright" students. However, the scheme i s able to cover only a minority o f students, and the criteria and methodologyfor applying it are not well-defined. 5.23 The total allocation for bursaries in 2003/04 was K Sh 770 million; the MPER estimates this was sufficient to cover only 70,000 students (about 10 percent o f the total). Although there is no objective indication o f absolute need, it i s the view o f the ministry and those in the education sector in Kenya that the proportion currently covered i s insufficient. The MPER i s unclear as to whether the intent i s to significantly increase funding or not; policy documents talk o f doing so, but the expenditure projections do not accommodate a major increase. 5.24 Three options are identified in the MPER: to continue full bursaries for a small number o f students; to cover more students, but at a lower level; or to significantly increase the budget for bursaries. Analyzing the options and costing them for the next PER needs to be part o f the overall secondary financing strategy proposed below. It is acknowledged, however, that these are very difficult issues to solve. Nobody really knows to what extent the cause o f low enrollment in forms I-IV are mostly supply or mostly demand factors. Yet optimal spending decisions depend very much on the balance between these two sets o f factors. 5.25 Secondaryschool construction. A large number o funplanned secondary schools are opening as a result o f local initiatives. Under the current practice, communities can establish schools, and the government i s obliged to provide teachers once the school is "accepted" for funding. Clearly, this i s not sustainable in the absence o f a more coherent strategy and a mechanism for building costs and affordability into the ministry's budget projections. It also results in a proliferation o f small schools that do not necessarily fit into an optimal coverage plan. Above all, this policy i s likely to increase the concentration o f secondary schools in comparatively richer communities that can afford to construct school buildings. The MPER proposes a limit on the licensing o f new schools untilthis issue i s resolved, which would be a step inthe right direction. 5.26 Official policy is also that expansion in secondary capacity should first take place by making all existing schools three-stream before building new ones-but this is not loThis is based on unit costs o f K Sh 21,800 per student, as cited inthe MPER. Note, however, that these are unit costs to government. I t appears that the MPER does not capture a large share o f the costs of secondary education. Schools spend out o f their fee income for many things that are not recorded in any analysis. MPER cost projections are limited to salaries o f teachers who are on the public payroll (some schools complement the numbers o f public teachers with direct h u e paid out o f fees) and bursaries. Furthermore, the data raise questions: at 800,000 students, the fees o f K Sh 15,000 each would imply revenues o f about K Sh 12 billion (15 percent o f total education budget); AIA under the education budget (table 2.4 o fthe MPER), however, shows only K Sh 67 million inMoEST's recurrent budget. 66 happening. A further cost issue has to do with boarding schools, which are expensive to build and operate. Officially, to reduce costs to students, all new and expanded schools are to be day schools (see the Educations Sector Strategy Paper [ESSP]), but this i s politically unpopular because o f the perception that boarding schools are o f better quality than day schools. 5.27 The need for a clearer secondary strategy. More broadly, there is a need for a clearer strategy with respect to how secondary expansion will be managed. Apart from the obvious issues identified above, o f teachers, bursaries, and new school construction, the more fundamental questiono fthe "right" exit point for education requires debate. The current level o f transition o f 50 percent o f primary graduates i s still relatively high for low-income countries, and given the likely evolution o f the labor market, the appropriate number ofyears o f schooling for various partso fthe population needs to be investigated. There is no easy answer to this, but the basis for the proposed 70 percent transition rate does not appear to be anchored in a debate about the role o f secondary education in a national development strategy. Given the cost and financing constraints, such a debate needs to be incorporated indevelopment o f the proposed secondary education strategy. Tertiary Education 5.28 As the MPER recognizes, the government faces a dilemma: there is demand for universityexpansion (to increase the current enrollment level of 72,000), but there is also a limit on what the state can afford to fund. The strategic approach adopted appears sound: Universities in Kenya recover a significant share o f costs inthe form o f fees (for example, the MPER reports that some universities cover half of costs), and the government appears to have limited its commitment to financing the sector (for example, the MPER projections propose only a marginal increase in grants to universities, from K Sh6.3 billionin2003 to K Sh6.8 billion in2006). 5.29 However, like the secondary sector, the tertiary sector suffers from problems of ambiguity with respect to expansion strategy and policy for public financing; it also appears that the expenditure projections have not taken account o f the proposed salary increases for university lecturers. An additional public expenditure concern i s that a large share o f the stalledprojects andmuch o f the unpaid bills are inthe universities; iffunding for the universities i s to be rationalized, it i s important to accept that the stalled projects have to be cut back. 5.30 The MPER recommends a university sub-sector study and reports that there is a commission now at work on this. The findings should be fed into a refined set of decisions regarding the future level o f budgetary fbnding for the universities. Among other things, the study should establish the most effective means o f providing incentives to universitiesto respond to national priorities. Technical and Vocational Education and Training (TVET) 5.31 The overall annual intake is high, standing at 100,000 (compared with 72,000 for tertiary education). As the MPER acknowledges, TVET inKenya faces many o f the same 67 difficulties as elsewhere: high unit costs (K Sh 100,000 per student annually, five times the estimated per student cost o f secondary education), questionable effectiveness, inadequate curricula, lack o f integration with workplace and employer requirements, poor facilities, and trouble inretaining staff. In light o f these concerns, the proposed increase in funding for TVET made in some quarters is questionable: The MPER projections propose more than doubling spending, from about K Sh 1billion inthe 2003/04 budget to KSh2.6 billionin2005/06. 5.32 Additional funding for the subsector should be limited until there i s a thorough review o f effectiveness (which needs to include tracer studies to determine the extent to which trainees are actually employed in the fields they are trained in and the impact on their incomes), as well as o f the appropriate role o f on-the-job and employer-based and - financed, and private sector training as opposed to TVET deliveredinpublic institutions. The MPER mentions an "ongoing consultative process to develop a national TVET strategy (that) is at an advanced stage." As part o f this, a participatory study on TVET is under way; this will draw on past studies recently undertaken by MoEST and by the Ministry of Labor and Human Resource Development. The outcomes of all this, sufficiently informed by considerations o f affordability and cost-effectiveness, should feed into the next PER. 5.33 It is appreciated that TVET is considered a national priority-to increase the employability of youth and build the country's technical capacity-but there i s little benefit in spending large amounts on training that i s not effective, and there i s a risk that inrushingto expand capacity, the government will make commitments that are difficult to get out o f inthe longer run, and which are o f little benefit. Other Types of Education 5.34 Earlychildhood(preprimary)education.Since 1997, this area o f education has been developed under the Early Childhood Development Programme. Its current gross enrollment rate-at more than 40 percent-is comparatively high relative to other countries at a similar level o f development. There are various stakeholders involved, including parents, communities and NGOs, who contribute to funding, and MoEST, which provides teacher training. 5.35 The GoK is, however, aware that provision o f free preprimary education would be extremely expensive, at about K Sh 9 billion or nearly 1 percent o f GDP. Hence, this requires further scrutiny, preferably under the overall education sector reform. 5.36 Specialeducation.It is estimated that about 10percent (approximately 3 million Kenyans) are disabled in one way or other. Again, about 25 percent o f these 3 million are school-age children. An estimated 90,000 have been identified and assessed, but only about 15,000 are currently enrolled in education programs, with an equivalent number integrated inregular schools. At the primary level, there shouldbe a greater focus on an all-inclusive education system that accommodates the needs o f the physically deprived within the FPE structures. At the tertiary level, enrollments have remained dismally low for the physically disabled. 68 Overview of Needsinthe Sector 5.37 The broad policy framework is sound, education is relatively well financed (at least compared with many other countries in the region), and although there are areas requiring additional funding, much of the need is for rationalization, especially with respect to use o f teachers, and for clearer policies to deal with public funding for the increasingdemands for spending on secondary education, TVET, and universities. 5.38 Ifthe governmentmoves aheadwith the analytical work that it already appears to have committed to or considering-including the teacher staffing norms study; processing o f school census data and school mapping; studies o f fbture strategy and financing at the secondary, TVET, and university levels; and strengthening monitoring and evaluation-it should go a long way toward being able to address the outstanding questions by the time o f the next budget.A central need will be to start work early to pull together the expenditure implications of these other pieces o fwork inthe 2005 MPER. ExpenditureProjections 5.39 The MPER team did a good job preparing broad projections of spending requirements for the coming four years on the basis of generally reasonable assumptions. However, in the time available, they could not undertake critical analysis o f the implications o f the projections nor more nuanced costing o f policy options. 5.40 The projections preparedpropose a 50 percent increase in the education budget over the next three years, broken down broadly as shown intable 5.3. Table 5.3. Projected Education Expenditure Targets Modeled in the MPER (Current KSh billion) 2003 2004 2005 2006 Early childhood Education 0.6 2.0 3.0 3.6 Primary 33.7 45.1 48.3 50.2 Secondary 21.3 30.4 34.5 40.6 TVET 1.o 2.5 2.7 2.7 Teacher training upgradinga .05 .06 - - Specialeducation 0.2 0.9 0.9 0.6 Strengtheningadministration 0.4 1.8 1.9 1.7 Grants to universities, student 7.7 8.0 8.0 8.3 bursaries, etc. Otherb 12.1 14.2 15.3 17.4 Total' 82.2 105.6 115.8 126.1 a. Appearsto includeonly incrementalcosts. b. Appearsto cover all other recurrentoperating costs ofMoEST. c. Total includes additionalprojectedincrementaldevelopment costs (see MPER Table 9.1). Source: MOESTMPER 2003. 5.41 There are a number of observations on the projections. (i) It is not clear that the projections realistically reflect the likely increase inwage costs (as a result o f either staffing or wage increases). Of the total increase o f K 69 Sh 44 billion annually proposed by 2006, only K Sh 8 billion i s accounted for by increased wages for primary and secondary teachers. As a result, the amount by which it will be possible to increase spending on other items is probably overestimated. (ii)Itisnotclearthat administrationandgeneralexpenditurerequirementshave been fully taken into account.' * (iii)Theprojectionsrepresentagoodfirstattemptatbottom-upestimationofneeds based on unit costs; inthe next phase, they need to be reconciled more explicitly with the actual budgets12and likely resource availability. OtherAreas for IncreasedSpendingandFutureFundingSources 5.42 Inadditionto areasthat have already beenmentioned, there are some areas where additional spending i s probably warranted in the near term, most o f which are not particularly expensive: Support for community/NGO schools in slum areas (More than 300,000 children were enrolled in these schools, according to data for the 2003 primary school capitation grants, with about 50 percent o f this accounted for by Nairobi slums. Increased funding here i s vital for the government's efforts to target the "hard to reach" including street children.) Increased financing for the inspectorate function (This probably needs to involve increasing salaries for the inspector cadre to attract good, seasoned teacher^.'^) Clearingpendingbills and cleaning up the portfolio o f stalledprojects Training school heads in how to run a school and giving small incentive payments to school principals (At present, no one wants to be a school head because there i s little extra pay and much additional work.) Getting the education management information system working properly and processing census data promptly, which should have a specific budget line for statistics andplanning. 5.43 Somewhat more costly medium-term issues include addressing the shortage o f teachers in overcrowded schools with temporary teachers on contract, hired at school level; providing for temporary teachers to replace those that are absent for the long term ~ MPER table 9.1 appears to account only for incremental costs o f strengthening administration and upgrading teacher training, not recurrent operating costs. The model apparently had a catch-all for administration, an "other spending" category, because o f the problems o f capturing these areas. This will be resolved ifthe projections are reconciled withthe actual current budget. l2There are differences between the model estimates o f costs for the base year and the budget for that year o f about 25-30 percent of the total, and there are things in the projections that are not in the budget and MTEF.Also, the model assumed increasedefficiency invarious areas, which may not have materialized. A thorough comparison of the model projections with the budget-MTEF figures is included in the terms o f reference for the 2005 MTEF l3Inspectors are paid much less than teachers with similar qualifications, so the best ones decide to teach, instead. A significant share of inspector posts is unfilled. This i s a policy issue as much as a spending one. 70 because o f illness; and the costs of continuing feeding and deworming programs in the schools. 5.44 Overall, the costs for attaining these goals are high, and continued donor and private sector support will be crucial. Indeed, the National Conference on Education and Training, held in November 2003, recognized that donor and private sector support can substantially boost efficiency and effectiveness in education financing and hence delivery. BudgetAnalysis,Execution,andExpenditureManagement 5.45 There i s some concern regarding the opaqueness of budget estimates and other issues relatedto data reporting and expenditure management: Teacher salary costs are all included under the general administration heading in the MPER, making it difficult to meaningfully analyze the costs o f service delivery. A move to budgeting by programs would take care o f these splits. Similarly, the partitioning of a significant share o f the costs o f primary education into an FPE operation under the development budget makes it difficult to assess true costs. (This might be a necessary consequence of donor-including World Bard-funding, but there is no reason not to classify the expenditures more transparently for expenditure review and planning purposes.) As in other sectors, the coverage o f aid hding is incomplete. At a minimum, better estimates o f donor funding and its composition are essential-even if not included formally in the budget-to allow the government to assess realistically how much o f its own resources it should allocate to items such as textbooks, training, or infrastructure rehabilitation. Ifthis is not already being done, by the time of the next PER, the ministry should mount an exercise to consolidate total donor education funding by line item and subsector. There i s significant spending by secondary schools out o f their own- revenues. These represent a form o f public expenditure, and they should ideally be captured in the budget. But if not, it i s important that they be included in expenditure planning, so the authorities can have a complete picture o f the costs o f secondary education. Finally, the youth polytechnics (run by the Ministry o f Labor) represent a major item o f expenditure on education and need to be included in the analysis o f secondary and TVET education. 5.46 In addition, the district-level expenditure reporting systems are currently inadequate. The framework for reporting i s slow, constraining service delivery at the local level, and would be improved if the ministry had access to the more immediate reporting on actual expenditures available through the Treasury system. 71 5.47 The MPER does not provide a goodbasis forjudging expenditure execution. This may be a data problem: For example, table 2.2 of the MPER shows almost no variation betweenbudgeted and actual expenditure, whereas the overall expenditure analysis inthis report suggests overspending o f 11percent on the education recurrent budget in 2002/03 (statistical annex table A 17) andunderspending o f 24 percent on the development budget (table 3.6 o f the MPER). There may not be large variations inrecurrent spending largely because so much is for salaries; the development budget is, o f course, quite different. Future MPERs need to more explicitly address actual spending relative to printed budgets. 5.48 With the increasing amounts of the budget passed to schools as grants, there is concern about expenditure control and tracking. The recent preliminary expenditure tracking study (KIPPRA 2003a) reports funds flowing to schools, but this does not tell us anything about how effectively those funds are being used, nor about the extent o f leakage within schools (for example, through management procurement, success of oversight, and so on). Strengthening the inspectorate function would appear to be an important, and not very expensive, endeavor. More generally, strengthening monitoring and evaluation within the ministryshouldbe a highpriority. The PER Process 5.49 To allow time for more rigorous analysis o f the outstanding issues, the ministry has determined that it should start work almost immediately on the next MPER, with the aimo fproducing a first draftby September 2004. 5.50 The work proposed above should provide a sound basis for more refined projections and more informed decisions on expenditure allocations by the time o f next year's PER. Summary of Recommendations 0 Drawing on the staffing norms study quantify the options and costs o f both addressing the immediate imbalances in teacher-student ratios and the long-term requirements for increased teacher numbers. In the meantime, there should be increased funding for temporary teachers to replace those that are absent for the long term because o f illness and in-service teacher training for current primary teachers. 0 Teacher salaries should be contained inthe longer term 0 Process the 2003 school census data, complete school mapping to identify needs, and then come up with a prioritized and costed list for the construction and rehabilitation o f classrooms. In the meantime, concentrate on meeting urgent needs, as well as on facilities in communities that are too poor to contribute from their ownresources. 0 Implement an analysis determining optimal teacher training capacity. In the meantime,do not increase capacity. 72 0 Clarify secondary education strategy, including policies regarding bursaries, the establishment o f new schools, and the appropriate proportion o f primary students passing to secondary education. 0 For the university subsector, (a) develop and cost a strategy for expansion and public financing; (b) prepare expenditure projections that take into account the proposed salary increases for university lecturers; (c) take steps to reduce the large share o f projects that are stalled, as well as unpaid bills; and (d) identify effective means o f providing incentives to universities to respond to national priorities; 0 Suspend plans to significantly increase funding for TVET until a thorough study o f effectiveness o f spending i s carried out and options explored for greater on-the- job, employer-financed or -based, andprivate sector training. 0 In the short to medium term, provide increased funding for (a) support for communityiNG0 schools in slum areas, (bythe inspectorate function, (c) pending bills and cleaning up o fthe portfolio o f stalled projects, (d) training andproviding small incentive payments to school heads, and (e) getting the education management information system working properly and processing census data promptly . 0 Produce improved expenditure projections, taking into account the salary increases that have been committed to, as well as administrative and general expenditure requirements. 0 Strengthen budgeting and expenditure management, including (a) a shift to budgetingby programs; (b) more inclusive expenditure planning, incorporating better estimates of donor funding and its composition, capturing spending out o f own revenues (especially for secondary schools), and inclusion o f the youth polytechnics; and (c) improvement o f district expenditure reporting systems, including mechanisms to share Treasury expenditure returns with the ministry. 0 Strengthen monitoring and evaluation, for example, through a well-conducted expenditure tracking survey and tighteningo fthe inspectorate function. MINISTRYHEALTH OF Sector Performance 5.51 The task o f the Ministry o f Health is to create an enabling environment for the provision o f sustainable quality health care that is acceptable, affordable, and accessible to all Kenyans. In addition, Kenya is a signatory to the United Nations Millennium Summit Declaration, committing the country to achieving the MDGsby 2015. The 2003 Kenya Demographic Health Survey (KDHS) preliminary report indicates that Kenya i s unlikely to attain the MDGsby that time. 5.52 The first PER o f 1997 revealed a mismatch between policy and resource allocations. The subsequent PERs (2003 and 2004) reveal a similar trend. 5.53 Health outcome and access indicators in Kenya have been worsening, as shown bythese data: 73 0 Infant mortality has risen from 62 per 1,000 in 1985 to 74 per 1,000 in 1998. e Maternal mortality is high. At 590 per 100,000 in 2001, it i s one o f the worst matemal mortality rates worldwide. e Life expectancy has fallen from age 57 in 1985 to 46 in 2002 (with malaria and upper respiratory tract infections accounting for about 50 percent o f outpatient morbidity and 25 percent o freporteddeaths). 0 There i s inequityin access. Twenty-five percent o f sick Kenyans do not seek care because o f financial barriers; the majority o f people live more than 5 kilometers from the nearest health facility. 5.54 Aside from inadequate funding and poor distribution o f resources, there are a number o f inefficiencies inthe sector that needto be addressed. These include: 0 Leakage and wastage o f drugs and medical supplies: This occurs in a number o f ways. For example, the recent KlPPRA PETS found that 30 percent o f all drug supplies released from the district headquarters supply depots do not reach the intendedhealthcenters anddispensaries. 0 Shortage and unequal distribution o f personnel: This i s caused by the "brain drain" in the 1980s and 1990s (as well as underfunding). Because o f unequal distribution, rural facilities are especially badly affected, and several have closed down. 0 The referral system does not function well: Patients bypass health centers and go directly to hospitals for basic care because o f inadequate equipment and staffing at the health centers. 0 The flow o f funding to health facilities-especially at the primary care level-is poor, and there i s some leakage o f user-fee revenues: Most health centers and dispensaries rely almost entirely on user-fee revenues for O&M expenditures because there i s little flow o f funds to them from the Ministry's headquarters. Yet the user-fee revenues are not adequate to ensure good service quality, andthe fees deter access for many poor Kenyans. In addition, the KIPPRA survey found that there i s a leakage o f about 22 percent o f user-fee revenues at the primary health care facility level. 0 Other sources o f inefficiency include a reliance on input-based rather than output- based financing and a lack o f efficient management o f resources at the facility level. 5.55 All o f this comes in the midst o f an HIV infection prevalence rate o f about 6.7 percent o f the population. This has strained resources-for example, 50 percent o f all hospital beds inKenya are occupied by AIDS patients. SectorBackgroundandBroadPolicy 5.56 The government's broad strategy, with an emphasis on primary health services and a package o f care geared toward the health problems that most affect the poor, i s basically sound. However, expenditure allocations do not seem to conform to the stated policy priorities, and there appear to be significant nonfinancial constraints that affect 74 where resources go and what impact they have. The data presented inthe MPER do not allow a full analysis o f these issues, and the balance o f this section focuses on identifying a program o f work that would allow next year's PER to come to more concrete proposals for restructuring health expenditure. AggregateExpenditure 5.57 Table 5.4 shows total actual public spendingon health inthe recent past. Table 5.4. Total Public Spending on Health, 1999/2000-2002/03 (Actual, K Sh Billion) 1999100 2000/01 2001102 2002103 Recurrent (gross) 9.4 11.0 12.7 14.4 Development 0.6a 0.9 0.9 0.6 Total 10.0a 11.9 13.6 15.0 Share of GoK spending (%) 8.4 7.7 8.9 ? Share of GDP (%) 1.3 1.4 1.5 ? Spending per capita 332 384 442 470 US$ per capita $4.72 $5.04 $5.61 $6.43 a. This table draws from the Ministry of Health's MPER table 1, which almost certainly understates development spending. MPER annex table 2 shows higher levels of recurrent spending, and MPER annex table 1 shows a much lower share oftotal GoK spendingfor 2002103 and2003/04. 5.58 Health spending has grown fairly substantially in recent years, by about 50 percent in nominal terms since financial year 2000/01; this growth has been somewhat faster than increases in total public expenditure. The share of total public spending, at about 9 percent, i s not far out o f line with that in other African countries (Tanzania, for example, spends 10 percent, and Ethiopia, about 6 percent). Per capita spending, at US$6.40 equivalent, i s relatively low for a country at Kenya's level o f income. 5.59 The international norms, such as the Abuja Declaration o f 15 percent o f total spendingon health and the WHO target o fUS$34 per capita, are not likely to be realistic inthe case ofKenyainthe foreseeable future. However, there is little doubt that the level o f health spending could usefully be raised, provided the incremental resources are allocated to high-return interventions and provided the absorptive capacity constraints discussed inthis section with respect to such issues as staffing and drugs management are addressed. Compositionof Expenditure 5.60 Table 5.5 shows the distribution o f expenditure by level o f care. There are concerns about the high .share o f spending on curative services relative to the government's stated policy priorities, as well as about the high and growing share o f spending consumed by the Kenyatta National and M o i Referral Hospitals. The broad data, however, give only a partial picture o f what is, in fact, happening in the health sector, and they are not nuanced enough to allow us to determine to what uses public spending is really being put, nor how well it conforms with stated policy. There are two elements to this issue. 7 5 Table 5.5. Recurrent Public Spending on Health, by Level of Service, 2000/01- 2002/03 (Actual K Sh Million) 2000101 2001/02 2002103 Share of 2002103 spending (%)a General administration 717 1,780 1,147 7.5 Kenyatta National Hospital 1,349 1,865 2,327 15.2 MoiReferralHospital 0 352 421 2.7 Curative services 6,178 7,395 7,798 50.8 Provincialhospitals District hospitals RuralHealth Services 1,370 1,775 1,634 10.6 Health centers, dispensaries Preventative& promotive 1,260 799 815 5.3 Otherb 1,195 1,265 1,206 7.9 Total 12,069 15,231 15,348 100 Note: This table draws from the MPER's table 3 and annex table 2. Totals shown elsewhere in MPER differ. a. Shares for 2002103 are based on the total of these estimates o f K Sh 15.3 billion. b."Other" includes training, research, andmedicalsupplies. 5.61 First, there is a needto better understand what services are really being delivered with the money being spent. For example, although there is concern about the 68 percent o f recurrent spending going to curative services and the two large hospitals, it i s not known what mix o f services they are providing. These tertiary-level institutions and provincial and district hospitals may be delivering a substantial share o f basic health care and primary services (perhaps because o f bypassing, distribution o f demand, the failure of lower-level facilities to function, or a combination o f these factors). To assess this requires specialized analysis o f a sample o f hospitals at different levels-measuring budgets, activities, time use, and the distribution o f services actually delivered-to determine how much o f the spending that appears in the budget as tertiary care i s really providing basic services, and at what cost. 5.62 The second element would be to more explicitly measure use o f facilities and services, to get a better idea o f the distribution o f the benefits o f public spending. This is not just, or primarily, a question o f income-based benefit-incidence analysis, but also involves an assessment o f the distribution o f expenditure geographically and by type o f service. This should be made possible by drawing on existing studies and the data that will be producedbythe forthcoming IntegratedHouseholdBudget Survey. 5.63 The object o f both o f these exercises is to help assess whether the current spending pattern i s a rational use o f resources. For example, spending on lower-tier facilities, though generally desirable, i s not an end in itself, but makes sense only if it i s more cost effective than delivering the same services through tertiary facilities. Clearly, however, one reason for the low usage o f primary facilities is the absence o f drugs, staff, 76 and resources, and part o f the solution needs to be higher levels o f finding. But finding the "right" balance depends on answering the questions posed above, as well as addressing the absorptive capacity issues discussed elsewhere inthis section. 5.64 The distribution o f recurrent spending between wage and nonwage spending in Kenya (figure 5.1) appears broadly rea~onab1e.l~ share o f 55 percent for salaries is not A excessive in the health sector, although the share of spending on drugs has been declining, and reported actual spending on drugs fell from about K Sh 1.6 billion in 2000/01 to K Sh 1.3 billion in 2002/03. Note, however, that the data are incomplete (for example, there i s likely substantial donor financing o f drugs that is not captured). Furthermore, the classification o f each o f the two large hospitals into single expenditure lines does not allow assessment o f how much o f the total budget i s really going to such items as salaries and drugs. Figure5.1. Distributionof RecurrentSpendinginMinistryof Health Transfers, Moi referral subsidies, & grants hospital ^^I 8% Salaries 55% I Source: Ministryof HealthMPER2003. 5.65 There i s a need to better unpack actual expenditure data, to understand where finds are really going and what uses they are being put to at a more disaggregated level-for example, to determine where, within the budgets o f the various subprograms, over- andunderspending i s taking place and where real shortages or surpluses infbnding occur.) To do this requires a more detailed analysis o f actual versus budgeted expenditure by line item(salaries, O&M, drugs, and so on) within each level of care (for example, at central hospitals, provincial hospitals, health posts, and so on). This, along with an analysis of where aid money is being spent, i s recommended as part o f the preparatory work for the next PER. l4Note, however, that there is uncertainty regarding the compositionof public expenditure on the M o i and Kenyatta Hospitals shown inthe MPERdata, which may be all salaries. 77 5.66 The development budget i s small relative to the recurrent budget, representing only 16 percent o f total public health spending in 2001/02.'5 Although the data are incomplete, there i s little doubt that the development budget i s overextendedrelative to available financing and i s consistently underspent. The 2004 MPER, for example, lists 90 stalled projects at a total value o f about K Sh 3 billion. Ministryofficials say that capital plans consistently exceed the money available during the budget year, resulting in actual releases to projects that are lower than the estimates, and implementation i s not possible. 5.67 Broadly, the government i s following a rational strategy in first financing essential recurrent functions in a sector where the most important interventions relate to routine service delivery, and in a country where the health infrastructure i s fairly well established. But there are essential capital needs for both expansion and rehabilitation, and the lack o f realism in the development budget undermines meaningful expenditure planning. This i s exacerbated by incomplete coverage. The 2003/04 budget reports aid funding o f about K Sh 3.7 billion in2002/03, whereas the UnitedNations Development Programme's Development Cooperation Report for Kenya shows donor funding o f about K Sh6.8 billion (US$87 million equivalent) in2002 inthe health sector, suggesting that almost half o f aid is not being captured in the budget. Furthermore, there are many new nontraditional donors in the health sector, which are not covered by either o f these sources, including the Global Fund, Global Alliance for Vaccines and Immunizations, and numerous HIV/AIDS programs, which together are injecting enormous amounts o f new hnding. Although the amounts are not certain, estimates suggest they could be as much as US$80 million equivalent per year, approximately equal to about half o f current government spending. This makes rational expenditure planning almost impossible, because the government does not know how much of its own resources it should allocate to, for example, drug procurement or training in the absence o f full information about what others are financing. It i s critically important that all public health funding be brought under an umbrella, at least at the planning stage. A first step inthe next PER will be to assemble consolidateddata on health fundingby all donors. BudgetExecution 5.68 The recurrent health budget has been overspent in recent years, with actual spending amounting to 110-120 percent o f the printed estimates (see annex table A 17). The development budget conversely has reportedly been underspent by 75 percent in 2000/01 and by 30 percent in 2001/02, although this i s likely due in part to reporting problems, especially for donor funding. 5.69 It is not clear whether the extent to which the large variation from budgetsis due to (a) mismatches between budgets and real spending requirements (that is, the allocations arrived at during the budget deliberations may not have realistically reflected the spending obligations o f existing programs), (b) bottlenecks in fund flows, (c) intentional misallocation or reallocation o f funds to uses other than those specified inthe budget, or (d) data and classification problems. These are discussed below, but clearly '*This information comes from the 2002 MPER, and as inother areas, the data are inconsistent. A range of estimates are presented at different points inthe MPERs. 78 there i s a need to track specific "buckets" o f money through the systemto assess what is actually happening. 5.70 There are issues of misallocationand actual compliance o f expenditure allocations with stated policy. Particularly striking is the variation between planned and actual expenditure for tertiary as opposed to primary health care. As shown in figure 5.2, starting three years ago, the government made an effort to increase the budgets for basic health care, more than doubling the allocation to rural health services over two years, and cut back spending on curative services to finance these increases. The planned cuts were also very large, amounting to about 30 percent o f curative spending. In reality, actual spending went in the opposite direction. Spending on hospitals went up, exceeding the proposed new budget allocations by K Sh 3.5 billion in 2001/02 (167 percent o f the printed budget), whereas spending on rural services was K Sh 3.3 billion below the budgetedlevel (only 35 percent o fthe new, higher budget level). Figure5.2. Ministryof Health-Actual versus BudgetedExpenditure I Actualvs. BudgetedExpenditure .+-Rural health senices -- - budget 8,000 4-Rural healthser\/ices 7,000 - actual 6,000 Curatiw? budget - - 5,000 4,000 -) Curatiw?- actual 3,000 2,000 ,+Kenyatta + Moi 1,000 budget 0 -8- Kenyatta + Moi 2000/01 2001/02 2002/03 actual Source: Ministry o f Health MPER 2003 5.71 The reasons for this are not well documented, but understanding them is central to making more realistic plans for budget reform in the health sector. There are three possible explanations. (1) The reallocations were unrealistic: it was never going to be feasible to reallocate expenditure between fbnctions so rapidly. (2) Despite good intentions at the planning level, rigidities in the system and demand pressures from the population (including the relatively powerfbl urban-dwellers and curative health care providers) prevented implementation o f the reforms. (3) The reallocationnever really had the support o f senior policymakers at the ministry and political levels. A deeper analysis o fwhat happened i s essential to lay the basis for future expenditure adjustments. 79 Effectivenessof SpendingIssues 5.72 Expenditure transformed into outputs and services. The PER offers some implied observations on effectiveness (for example, the decline in immunization coverage), but more analysis i s needed to inform expenditure decisions. It i s always difficult to measure the impact o f health spending on outcomes (in part because health status is mostly determinedby factors other than spendingon public health services), but one can assess the extent to which spendingon various programs results inservices being delivered. One possibility i s to pick a sample o f specific interventions (for example, the immunization program, malaria, family planning) and do a more detailed analysis of trends inspendingandwhat measurable inputsand outputs have been inthe recent past.16 5.73 Absorptive capacity questions. There are clearly nonexpenditure factors that constrain the government's ability to spend money effectively, especially at the rural and primary health care levels. For example, health posts may not be staffed because o f problems o f recruitment, posting, or incentives; facilities may be being bypassed because o f location, transport, or preference factors; and there may be leakages (discussed below) or logistical constraints that make it impossible to effectively increase the flow o f drugs or funds to lower-level facilities inthe short term. Ininterviews duringthe PER mission, health facility managers also cited bottlenecks in release o f funds from the center and inadequate cash at the local treasury level to fund approved budgets. Evidence from the recent expenditure tracking survey and anecdotal reports suggests that these are problems, but the evidence i s not solid enough to draw firm conclusions. Before concluding that more money should be spent on a given good (for example, drugs) or service (for example, rural or urban health posts), there i s a need to more scientifically ascertain the extent to which these absorptive and institutional factors constrain the abilityto spend effectively, andwhat is neededto overcome them. 5.74 A program o fwork for the coming year could include (a) an assessment o f facility staffing and absenteeism (based on a facility survey) and identification o f measures to overcome constraints; (b) an assessment o f determinants o f utilization behavior, that is, why people use, or do not use, different facilities, along with better quantification o f use and bypass o f services; and (c) a review o f drug supply issues and bottlenecks. These may draw on existing health sector work (for example, with respect to the reform o f Kenya Medical Supplies Agency and drug logistics) and link it more closely with expenditure absorption implications. 5.75 Leakage and misallocationissues. There i s evidence of leakage and wastage o f funds and supplies at all levels, but it is incomplete and not consistently measured. For example, the recent expenditure tracking survey17reported that only 69 percent o f drugs released by districts reached health facilities, although it did not define how leakage was measured, nor attempt to explain what might have happened to these drugs. The same l6 There may already be detailed work going on in some of these areas that could be drawn upon. Deeper analysis could attempt to marry this with some assessment o f impact on outcomes. Such a programo f work i s ambitious. It might make sense to commit to a sequence o f studies over a number o f years as part o f the health sector reformprogram. KIPPRA (2003a). 80 survey reported 22 percent leakage o f user-fee revenues at health facilities, and 29 percent "total funds leakage," although again the terms and methodology were not clearly defined. In general, the data are not reliable enough to be conclusive, but they clearly suggest that there are problems. If confirmed, this raises questions o f whether it makes sense to allocate additional funds until these problems are fixed. It i s recommended that specialized surveys and analysis be undertaken in the coming year to more explicitly quantify leakages. As suggested above, a more nuanced expenditure tracking survey, which traces the flow o f specific releases o f money and resources through the system, would help. FutureFundingRequirements 5.76 This year's MPER has made a good start at estimating what might be required to deliver "adequate" health services in the longer term." But there are a number of limitations: Inthe time available, it was not possible to refine the estimates properly; the projections lacked realism in some respects (for example, they were based on financing "ideal" staffing and supply levels for a health center with 14 staff to treat only about 8 new curative cases a day); and at the macro level, they were unrealistically unconstrained (for example, total costs amounted to K Sh 43.8 billion by 2006/07-about three times the current health spending, representing an increase to 3.25 percent o f GDP from about 1.5 percent today). 5.77 In the case of the health sector, it is always difficult to estimate spending requirements because they are potentially almost infinite. Arriving at projected costs needs to involve a degree o f judgment as to what is realistically affordable, combined with some assessmento fthe cost-effectiveness ofinterventions andthe type o fabsorptive capacity concerns described before. Still, it i s possible to arrive at order-of-magnitude requirements using more refined unit costs for different services and marginal cost analysis. The govemment and donors could very usefullybuildon the good work already done by refiningthese estimates over the coming year. Public-PrivateRoles in Financing 5.78 The level of private financing is quite high. One estimate puts private spendingon health at 5.4 percent o f GDP," implying spending o f almost US$25 per capita, which seems high. Other estimates put the share lower, but in general, the relatively high proportion i s consistent with Kenya being at the upper end o f low-income countries, and it is somewhat different from the situation inmany other sub-Saharan countries. 5.79 Fees are charged at all levels in the public system. The MPER reports that about K Sh 1.5 billionwere recovered last year inthe form of user fees, representingabout 11 percent o f recurrent spending.20 The use o f fees without a well-functioning system o f waivers, however, can result in exclusion o f the poor. One study cites a finding that 25 l8See section of 5 o f the 2004 HealthMPER. l9KIPPRA (2003b). 2oAgain, the data are inconsistent: Other sources report cost recovery of only about 5 percent. 81 percent o f those interviewed didnot seek treatment because o f cost concerns. At the same time, Kenya is almost unique in not compensating facilities for lost revenues due to waivers. Additional survey and analytical work is being undertaken to better understand the situation and identify possible reforms. This should include drawing on household survey data and modeling ability to pay relative to householdbudgets. The NationalSocialHealth Insurance Scheme (NSHIS) 5.80 Given the very large public financing needs projected above, the government i s proposing a universal health insurance scheme, based on expanding the existing National Health Insurance Fund(NHIF).21 The rationale i s understandable, and such schemes have been an important part o f the health-financing solution in other countries (especially in Latin America), but there are three related concerns. One i s whether the scheme can be funded at the level proposed inthe near term. The total annual cost i s estimated at K Sh 40 billion annually (more than double the total current public spending on health), o f which at least K Sh 11billion is expected to need to be funded fiom tax revenues.22The second concern i s that the f h d will requirecontinuing large infusions o fpublic money to cover the poor. The third i s the practical problem o f how the large number o f Kenyans in the informal sector, rural areas, or both andwho are uneducated will be able to participate ina formal insurance scheme. All ofthese concerns potentially canbe overcome, but all o f them present challenges, and there i s a risk o f significant fiscal liabilities if the government goes into this expansion without adequate financial planning. It i s recommended that a comprehensive risk analysis be undertaken, concluding with an actionplanto address any shortcomings or risks identified. Conclusions 5.8 1 Immediate recommendations for expenditure allocations inthe coming budget are the same as those made in previous years: to increase the allocations for spending on lower-level service delivery facilities and staff, as well as spending on drugs, and to contain spending on the two central hospitals. But, considering current evidence, there i s little basis for determining optimal levels o f spending. 5.82 The PER work o f the past two years has not been sufficient to allow definitive decisions on reallocation o f expenditure in the health sector. In the coming year, the following program o fwork should be undertaken: 0 A retroactive assessment o f why the proposed major reallocations o f two years ago did not work The NHIF has existed since 1966. It currently includes some 1.8 millionmembers andprovides about 10 millionpeople with varying degrees o f medical coverage. 22 According to the Ministry o f Health (2003), the estimated total outlays are K Sh 40 billion, with financing expected to be composed as follows: payroll deductions fiom civil servants, K Sh 7 billion; earmarked VAT, K Sh 11 billion; contributions by the self-employed, K Sh 10 billion; employee and employer contributions, K Sh 12 billion; other (for example, donations), K Sh 1billion. 82 A more detailed analysis o f actual spending compared with budgets at a disaggregated level, by line itemwithin each program A specialized study to determine what services are actually beingdelivered at each level and type o f facility (health posts, provincial and district hospitals, and the Kenyatta National and M o i Referral Hospitals), including an assessment o f time use and expenditure within the facilities, to estimate the real distribution o f current spendingby level o f care An analysis ofuse offacilities andthe reasons for use andbypassingbehavior An assessment of absorptive capacity constraints, related to drugs, staffing, and flow o f funds, to identify whether it makes sense to increase financing, as well as an assessment o f the barriers that need to be overcome to do so (This could draw upon completed or ongoing work within the health sector.) A true expenditure tracking survey that traces the flow o f specific "buckets" o f funds (and inthe case o f drugs andmedical supplies, commodities) through the system, to determine where leakage i s taking place and its magnitude A catalogue o f donor inflows to the health sector (including those from new, nontraditional donors) andwhat they are financing Analysis o f the impact o f health care fees on exclusion o f the poor and possiblereforms to compensate A risk analysis o ftheproposedhealthinsurance scheme. 5.83 It is not realistic to expect this work to be done as part o f a PER exercise conducted over a fairly short time frame in the runup to the budget. Work should start now as part o f a broader consolidated effort geared toward health sector budget reform, which can lay the basis for both expenditure reform and a common program o f donor support for increased health sector spending. 5.84 Ifthis is done well, by the time of the next PER, the government should be ina position to make informed decisions regarding (a) revised health funding levels for the next budget, (b) proposed reallocations, and (c) complementary measures needed-for example, in institutional reforms, staffing and incentives, systemic or management changes, and so forth MINISTRIESAGRICULTURE,LIVESTOCKFISHERIES OF AND DEVELOPMENT, AND COOPERATIVE DEVELOPMENT AND MARKETING Sector Background and Broad Policy 5.85 Agriculture i s the major sector inKenya, and although its contribution to GDP has declined from 35 percent in 1964 to about 24 percent in 2000, its contribution to development i s still significant. Agriculture employs about 75 percent o f the labor force, provides raw materials for the agro-based manufacturing industries (which constitute 70 percent o f all industries), and accounts for about 45 percent o f government revenue. Besides, the sector i s the growth engine for the nonagricultural sector. Thus, agriculture i s the mainstay of the Kenyan economy and i s expected to maintain its role as a primary engine o f growth for the economy in the foreseeable future. Since July 2003, three 83 ministries-all formerly under one umbrella-handle agricultural activities inKenya: the Ministries o f Agriculture (MOA),Livestock and Fisheries Development (MoLFD), and Cooperative Development and Marketing (MoCDM. Together, these are key to providing a conducive environment and appropriate services for the agricultural sector to develop. 5.86 The livestock subsector accounts for about 42 percent o f Kenya's agricultural GDP and 10 percent o f the entire GDP. It employs almost 50 percent o f the labor force and supplies the domestic requirements of meat, milk and dairy products, and other livestock products. In addition, livestock products account for 20 percent o f the total marketed agricultural products and about 5 percent o f the total agricultural exports. The fisheries subsector contributes about 3 percent o f GDP and 3 percent o f total export earnings. It employs about 58,000 people directly and 500,000 people indirectly through fish processing and trade. These two subsectors fall under the MoLFD, which has the mandate o f promoting sustainable development o f the livestock and fisheries sector and ultimately contributes to the achievement o f food security. 5.87 The cooperative movement plays an integral role in the procurement o f agricultural and livestock inputs and marketing of outputs. The movement also plays a major role infacilitating the buildingup o f revolving funds for cooperative movements in various organizations. The creation o f the M o C D M i s therefore expected to spearheadthe growth and development o f an economically viable cooperative movement through formulation, development, and implementation of policy guidelines, programs, and legal frameworks that meet the aspirations of cooperative members. AggregateExpenditure 5.88 Available statistics indicate that on average, Kenya used to spend more than 10 percent o f its total government budget on agriculture (including livestock, fisheries, and cooperatives) in the first decade afler independence. This declined to an average o f 7.5 percent inthe period 1980-89 and to just 3 percent inthe 1990-2000 period. In2002/03, this figure was just 3.1 percent, about 0.7 percent of GDP. Agricultural recurrent expenditure has been consistently higher than development expenditure over the period under review. 5.89 The recurrent expenditure on agriculture (excluding livestock, fisheries, and cooperative development activities-that is, now handled by the MOA)accounted for more than 70 percent o f total expenditure (see table 5.6), which is dominated by salaries for employees, including the extension officers. However, less than 30 percent i s spent on agricultural development, which includes agricultural research and market information, animal health services, crop protection, seed inspection, mechanizationservices, and farm planning services. Government expenditure on agriculture over the period under review has fallen slightly as a share of total expenditure (fi-om 2.4 percent to 2 percent) and also as a percentage o f GDP (from 0.44 percent to 0.4 percent). 84 Table 5.6. TotalPublicSpendingon Agriculture, 2000/01-2003/04 2000101 2001102 2002103 2003104 actual actual actual printed Recurrent (K Shbillion) 2.6 3.2 3.1 3.3 Development (K Shbillion) 1.o 0.9 0.9 1.2 Total (K Sh billion) 3.7 4.1 4.1 5.0 Share of G o K expenditure 2.4 2.4 2.0 1.9 Share of GDP 0.44 0.44 0.40 0.45 Agriculture recurrent as % of total agriculture 71.8 78.1 77.1 64.9 expenditure Agriculture development as % o f total 28.2 21.9 22.9 35.1 agriculture expenditure Source: BMD. 5.90 The share o f total government expenditure devoted to the MoLFD was 1.7 percent in 2000/01, which declined to 1.1percent in 2002/2003 (table 5.7). As a proportion o f GDP, the expenditures have ranged from 0.33 percent in 2000/01 to 0.25 percent in 2002/03. Table 5.7. TotalPublicSpendingon LivestockandFisheries,2000/01-2003/04 2000101 2001102 2002103 2003104 actual actual actual printed Total (MoLFD) expenditure (K Sh billion) 2.8 2.3 2.9 1.1 Share o f G o K expenditure 1.67, 1.37 1.06 - Share of GDP 0.33 0.25 0.25 0.1 -Not available. Source: MoALD. 5.91 On average, recurrent expenditure accounted for more than 80 percent o f the funds allocated to the ministry over the period under review; this consists mainly o f salaries and transfers, with little provision for O&M. Development expenditure accounts for the difference and funds such core poverty programs as livestock extension services, fisheries development, development o f veterinary farms, and disease and pest control. Other development initiatives include research and extension, inspection and quality assurance, infrastructure. and monitoring and surveillance. For the sector to grow, more funds should be allocated, especially to fundingthe development projects. 5.92 The total M o C D M expenditure as a proportion o f total government expenditure was 0.17 percent in 2002/03, while as a proportion o f GDP it was 0.045 percent (table 5.8). Recurrent expenditure accounted for 89.5 percent in 2002/03, and this share increased to 97.5 percent in 2003/04, while the share o f development expenditure declined from 10.5 percent to 2.5 percent over the same period. Table 5.8. Total PublicSpendingon CooperativeDevelopment,2000/01-2003/04 2000101 2001102 2002103 2003104 actual actual actual printed Recurrent (K Shbillion) 0.2 0.3 0.3 0.2 Development (K Sh billion) 0 0 0.1 0.1 Total (K Shbillion) 0.2 0.3 0.4 0.3 Share of GoK expenditure 0.10 0.114 0.165 0.092 Share of GDP 0.029 0.029 0.045 0.024 Source: BMD. 85 Compositionof Expenditure 5.93 Actual agricultural expenditures (including spending on livestock, fisheries, and cooperative development activities) continued to deviate from budgeted expenditures over the period under review, with major deviations occurring on the recurrent account (table 5.9). However, actual expenditures relative to budgeted amounts have declined over time, perhaps as a result o f increased financial discipline attributable to budgeting based on the MTEF or decreased spending as a result o f a freeze on procurement o f goods and services inthe later part o f 2002/03. Table 5.9. Deviation between Budgeted and Actual Expenditures for Agriculture, IncludingLivestock,Fisheries,andCooperativeDevelopmentActivities 1999100 2000101 2001102 2002103 Recurrent 116.7 96.7 80.2 52.5 Development 12.7 17.9 34.8 39.5 Total 62.7 92.5 84.6 48.1 Note: These deviations are actual expenditures as a percentage ofbudgeted amounts. Source: BMD 5.94 Expenditures on the MOA'Score functions o f extension and research have generally consumed the lion's share o f the annual recurrent budgets, averaging about 80 percent during the period under review. Activities under these hnctions (largely handled by parastatals) show low deviations because the parastatals used all their allocations. Duringthe period under review, large amounts of MOAresources have been transferred to parastatals. More than 50 percent of MOAexpenditure i s spent on salaries and wages, and less than 1 percent is spent on plant and equipment. O&M accounted for only 18 percent over the period under review (figure 5.3). Transfers to parastatals remained well more than 30 percent o fthe total recurrent budget of the ministry. Figure5.3. DistributionofMOARecurrentSpending(Actual), 2002/03 Operation and Plant & maintanance equipment nnI Salaries & subsidies 34% Source: MoALD. 86 5.95 The bulk o f the recurrent budget allocation to the MoLFD has been going to payment o f salaries and wages: 46 percent in 2000/01, 61 percent in 2001/02, and 55 percent in 2002/03 (figure 5.4). This trend appears to be maintained for the current financial year 2003/04. A significant amount o f finds is transferred to parastatals: about 14 percent in 2000/01, which declined to 9 percent in 2003/04 because o f the transfer o f the KEVEVAPIand KETRI to the MOA.O&M, however, account for about a third o f the recurrent allocations to the ministry. Figure5.4. Distributionof MoLFD TotalExpenditure(Actual), 2002/03 Plant &equipment 3% Operations Salaries & wages 60% 10% Source: MoLFD. 5.96 The trend indicates that much o f the expenditure was incurred in the control o f livestock diseases and pests under subvote 106 that accounted for 30 percent o f the total expenditure in 2000/01, 35 percent in 2001/02, and 32.6 percent in 2002/03. The least expenditure was under subvote 107 on the protection o f the natural resource base for livestock. 5.97 Expenditure on public enterprises for the MoCDM took up 29.9 percent, transfers accounted for 6.2 percent, and O&M took 63.9 percent o f the ministry's allocation in 2002/03. The ministry generally has three subvotes or functional units classified as general administration and planning services (vote 220), cooperative management services (vote 22l), and cooperative training and development (vote 223). Expenditure on cooperative management services took the lion's share o f the allocations in 2002/03, at 72 percent, while cooperative training at the cooperative college was allocated the least, about 11percent (see figure 5.5). By January 2004, the ministryhad spent 56.3 percent o f its allocation for the year, with an allocation o f 7 percent to the cooperative training college. 87 Figure5.5. Expenditureby Subvote for the MoCDM(Actual), 2002/03 I Vote 223 Vote 220 11% 17% I 72% Source: MoCDM. 5.98 O&M accounted for the bulk o f public spendingunder the ministry, while wages and salaries took up to a thirdo ftotal expenditure in2002/03 (figure 5.6). However, plant and equipment typically accounted for less than 1 percent o f the allocation, despite the need for improvement o f humancapacity via provision o fnecessary efficiency-enhancing equipment such as information and communications technology. Figure5.6. Distributionof MoCDMTotalExpenditure Plant & Salaries & equipment - wages 30% Source: MoCDM. Transfersto Parastatals 5.99 There are 26 parastatals under the MOA,classified into farming and allied (5), regulatory bodies (9), research and training institutes (4), and sugar factories (8). Transfers to parastatals still constitute a large proportion o f the ministry's development (65 percent) and recurrent (37 percent) budgets. The bulk o f the transfers go to research under the KARI, including its recently acquired departments, KETRI and KEVEVAPI. The allocation to KARI in 2000/01 was about 92 percent o f total recurrent transfers, 88 which declined to 84 percent in 2003/4 after the government's decision to provide the Horticultural Crops Development Authority (HCDA) with grants. Table 5.10. Transfers to Parastatalsas Percentage of the MOA'SRecurrentBudget, 2OOOlO1-2003/04 2000/01 2001102 2002103 2003104 actual actual actual printed KAIU 26.6 24.7 25.4 26.3 KETRI 6.0 6.0 6.2 6.7 KEVEVAF'I 0.4 0.3 0.3 0.3 Kenya Plant Health 2.4 2.2 2.2 2.6 Inspectorate Service (KEPHIS) CottonBoard 0.4 0.3 0.3 0.3 HCDA 100.0(AIA) 100.0 (AIA) 3.5 2.5 Total parastatals 35.7 33.4 37.8 39.8 Total MOA(K Sh million) 2,637 3,239 3,315 3,27 1 Source: MOA. 5.100 As shown in table 5.10, transfers to parastatals in the period under review averaged 36 percent, with KARItaking the lion's share. These trends leave the MOAwith little room for saving from the large transfers. The large amount o f these transfers indicates how crucial it i s to ensure efficient service delivery by these parastatals. The ministry's development transfers have mainly been going to the National Cereals and Produce Board (NCPB)-57 percent in2002/03 (figure 5.7). DevelopmentExpenditure 5.101 The net actual development expenditure for the MOAincreased from K Sh 1.038 billion in 2000/01 to K Sh 1.765 billion in 2003/4 as a result o f the inclusion o f new activities such as the Lake Victoria Environment Management Programme, loans to Cooperative Bank for onward lending to farmers, and grants to special food security initiatives. However, actual development expenditures as a share o f approved budget levels fell from 82.5 percent in 2000/01 to 65.1 percent in 2003/04 because o f the freezing o f government procurement in the later part o f 2003/04. A unique feature o f development expenditure i s the distortion created by the large allocation channeled to the NCPB for its commercialization activities under subvote 105. KARI and the NCPB together accounted for 99 percent o f development transfers in 2003/04. Most important, perhaps, i s that most o f the development expenditure i s hnded by donors. The problem with donor finding is that it is usually unstable because o f the donors' changing policies and, hence, is not a sustainable long-term strategy for agriculturaldevelopment. 5.102 The MoLFD has three parastatals and two statutory bodies that receive grants from its vote. These include Kenya Marine and Fisheries Research Institute (KEMFRI), Kenya Dairy Board (KDB), Kenya Meat Commission (KMC), Kenya Veterinary Board (KVB) and Central Artificial Insemination Station (CAIS). Grants to these parastatals accounted for 14 percent o f the total budget in 2000/01 and 2001/02, increasing to 19 percent in 2002/03 but declining to 9 percent in 2003/04 with the transfer o f KETRI and KEVEVAPIto the MOA(figure 5.8). The implication o f this policy is that only limited finds are left for O&M. Recurrent expenditure allocation accounted for 95 percent of the transfers to parastatals from the ministry in 2002/03, while the development allocation 89 took up only 5 percent. KEMFFU takes the lion's share o f the allocations at 87 percent, and the KVB takes less than 1percent. Figure5.7. DistributionofMOADevelopmentSpending(Actual),2002/03 Coffee HCDA 1 Research no/n _ I " Foundation NCPB 57% 1% Source: MOA. Figure5.8. RecurrentTransfersto Parastatals(Actual), 2002/03 KVB CAIS KDB 0% 3% 10% Source: MOA 5.103 In the MoCDM, transfers accounted for 6.2 percent in 2002/03 alone and were mainly to the cooperative college of Kenya. However, in 2003/04, this allocation increased as a result o f the government's allocation o f funds for the revival o f Kenya Cooperative College (KCC) and Kenya Farmers' Association (KFA). BudgetExecution 5.104 The recurrent MOAbudget was underspentby about 6 percent in2003/04 because o f the govenzment's freeze on procurement. Other reasons for the overall decline in use o f budgeted funds include nonadherence to set work plans, liquidity problems at the 90 district treasury caused by reimbursement delays and late or nonrelease o f funds by the Treasury. Similarly, there were no overexpenditures inthe development vote. 5.105 The variance between budgeted and actual expenditure for the MoLFD has widened over the review period. Overall, actual expenditure was below the budgeted estimates, with significant underexpenditure o f about 38 percent in 2002/03. Possible explanations for this include reduced capacity to manage budgets within the MoLFD because of the recent retrenchment program as well as the GoK freeze on procurement in 2003. 5.106 Overall, actual expenditure for the MoCDM appears generally low because o f public expenditure problems and freezes on procurement early inthe fiscal year 2003/04. The ministry faced problems in using development funds because o f cumbersome reimbursement problems. Similarly, major deviations are common between budgetedand actual expenditure-especially on recurrent expenditure, where the divergence (ratio o f actual to budgeted amount) was up to 92 percent in 1998/99 but declined to 20 percent in 2002/03 (table 5.11). Table 5.11. Percentage Deviationsbetween Budgeted and Actual Expendituresfor the MoCDM 1998199 2002103 Recurrentexpenditure 92.1 19.5 Developmentexpenditure 74.0 0 Total expenditure 89.7 17.9 Note; These deviations are actual expenditures as a percentage of budgetedamounts. 5.107 Budget outruns and divergences are common because o f 0 Unpredictable budgetary cuts by the Treasury as a result o f low revenue collections. Often, approved budgets are lower than the initial printed estimates, and priorities do change during revision o f estimates leading to realignment o f resource requirements. 0 Poor public expenditure management due to incremental nature o f budgeting. Resource requirements are not appropriately matched with realistic departmental requirements, which lead to diversion o f resources to noncore activities. It i s also common knowledge that work plans are not strictly adhered to once funds are secured through budgets, and this leads to ad hoc spending. 0 There i s lack o f effective monitoring and evaluation and tracking of public expenditures by finance and planning. There i s overall less supervision o f the field byheadquarters. 0 Some AIA or revenues from own sources generated by the ministry are used to support cooperative development and are therefore not remitted to the government as revenue. There has been generally low realization o f AIA revenue collection because o f a backlog of audit work. 91 DevelopmentBudgetandDonorFunding 5.108 A significant proportion o f the ministry's development budget has been from AIA. AIA shares have increased from 59 percent in2001/02 to 68.8 percent in2002/03. Revenue under AIA covers donor assistance, self-generated income, and user fees (as opposed to income from state sources). Inthe past, there was a highreliance on AIA from various development partners, which did not always materialize. In2003/04, this share i s expected to be 83.3 percent, signifying the growing importance o f extrabudgetary resources in meeting development expenditure. Recurrent vote AIA receipts come from such sources as sale o f nonplant, plant, and equipment,'as well as sale and fees for various services renderedby various sections inthe ministry. 5.109 During the period under review, actual receipts as a share o f total budgeted receipts dropped from 29.6 percent in 2000/01 to 18.3 percent in 2002/03. This could be attributed to gross underestimation o f the expected AIA as well as the centralized sale o f various old and excess govemment plant and equipment. A general review o f the AIA accounts indicates that there are wide imbalances between budgeted and actual receipts. There is, therefore, a need to look critically at the issue o f setting annual targets for AIA across the budget heads to come up with achievable levels o f AIA. In addition, there i s a need to look for ways o f capturing data on any revenue collected but currently not reflected, inaddition to information on directly funded project activities. 5.110 The major foreign donor to the MoLFD was the U.S. Agency for International Development (USAID) under the smallholder dairy project. AIA collected, however, showed a general increase between2001/02 to 2002/03. However, there was a divergence between actual collections and targeted collections, particularly for the veterinary department, which can be attributed to management problems. Other reasons for the divergence included unpredictable budgetary cuts from the Treasury, late release o f AIEs for districts, cumbersome bureaucratic requirements (especially among donors), liquidity o f the govemment, aid freeze, ambitious budgeting (especially in AIA), disease outbreaks, and mismanagement. Although the veterinary department leads in revenue collections, there was a notable increase inAIA collections from the fisheries department as a result of increased collections on licenses arising from efforts to establish a monitoring, control, and surveillance system within the Exclusive Economic Zone . Duringthe first six months o f financial year 2003/04, the fisheries department collected 78 percent o f its targeted collections for the whole year. 5.111 There is little donor fbnding to the MoCDM. However, the ministry generates AIA from four main sources: registration fees for cooperatives, registration o f charges with the cooperatives tribunal, audit fees, and miscellaneous receipts to the cooperative college o fKenya. Compliancewith ExpenditureAllocations 5.112 As noted before, variations between planned and actual expenditures for the MOA are not striking. Perhaps the main issues that need to be looked at here are the low levels of budgetary allocation to the MOA;the high rates o f transfers to parastatals, especially 92 the allocation to KART. and NCPB; and the effects o f freezing o f government procurement o f goods and services to the ministry. 5.113 Lack o f capacity in terms of both numbers and skills at all levels o f the MoLFD, especially in technical departments, has implied low transparency and accountability o f funds. This is also reflectedinthe form o flow spendingrelative to budgeted amounts and divergences between actual and targetedrevenue collections, especially for the veterinary services department. This lack o f transparency may require drastic surgery, which can only be achieved by way o f a thorough management review and would also indicate enhancement o f training requirements. 5.114 Overall in the MoCDM, expenditures appear generally low because o f public expenditure problems and freezes on procurement early in 2003/04. The ministry also faced problems in using development fbnds because o f cumbersome reimbursement procedures. The revival o f the KCC and KFA i s also likely to stretch the ministries' budgetary allocation. Further, mismanagement, low accountability, and lack o f transparency are rampant. These problems are worsened by the lack o f an effective monitoring system. Effectiveness Issues 5.115 Budgeting based on the MTEF has improved the efficiency o f use o f resources, especially with regard to core poverty programs within the MOA.But it has also resulted, inparticular, in a reduction inthe allocation to provision of extension, crop protection, and food security monitoring. The critical issue is whether reduced funding and strict discipline have reduced farmers' access to ministries' services. There are indications that there i s potential to cut costs and increase efficiency in service delivery in parastatals. Other issues, such as the freezing o f procurement, nonadherence to set work plans, and late release or nonrelease o f funds by the Treasury, also affect efficiency significantly. Other factors that weaken the ministries' absorptive capacity include delayed starts, difficulty in changing attitudes to embrace transparency and accountability, failure to appreciate work plans, and inadequate staffing. 5.116 Poorly planned withdrawal from some key services under the MoLFD, such as AI (artificial insemination), veterinary clinical, and dipping services, created a gap that continues to negatively affect the livestock subsectors. For example, privatization o f AI services was hurriedly done before assessing the capacity o f private providers to give the service, hence limiting coverage to only 24 districts. Also, dipping services almost collapsed throughout the entire country. These trends negatively affect the delivery o f services. Allocations to disease control have been on the decline, which does not augur well for this vital service. 5.117 Other constraints to efficiency include the retrenchment program that has led to lack o f capacity in terms o f both staff numbers and skills at all ministry levels. The delivery of services i s also constrained by inadequate budgetary allocations, inadequate technical personnel (as a result o f freezing employment of new staff by the government), and inadequate complementary services (such as infrastructure) to enhance service 93 delivery in the ASAL. Further, underspending relative to budget allocations and divergences between actual and targeted revenue collections indicate a lack o f accountability and transparency in public funds use that negatively affects the effectiveness o f the resources. Development o f appropriate exit plans for the remaining noncore services shouldbe undertaken ifefficiency is to be improved. 5.118 Over the years, the M o C D M has been underspending its budgetary allocation because o f public expenditure problems and the recent freeze on government procurement. Furthermore, the poor reporting mechanism and the lack o f a monitoring and evaluation system weaken effectiveness. Service delivery was also constrained by corruption in the cooperative movement, high levels o f illiteracy among members, poor record-keeping, inadequate mechanisms o f communication and conflict resolution, inferior equipment, and lack o f office space in some parts o fthe field. Absorptive Capacity 5.119 One o f the major hindrances to effective use o f the MOA'S allocation was a freeze on GoKprocurement. However, given the nature of the ministry's inability to fully use its allocation and frequent underspending, its absorptive capacity needs to be improved. 5.120 The main factors that constrain the MoLFD's ability to spend resources effectively include inadequate technical personnel, equipment, and complementary services. These are compounded by the govemment's freeze on procurement o f goods and services, which increased the percentage o f the budget allocation that was not spent. The veterinary services department has demonstrated a lack o f capacity to collect revenues and account for their use, which calls for an improvement o f its management team. The ministry also lacks an effective monitoring and evaluation system within its functional units, and this hinders evaluation o f accountability o f resources. There i s also a weak understanding o f the budget process among officers, especially o f the PER and MTEF exercises. The absorptive capacity is also hindered by low morale due to poor terms o f service, meager flow o f information, and inadequate office and laboratory accommodation. Leakages and Misallocation of Resources 5.121 Misallocation o f resources within the MOAhas been cited, especially because over the last few years, the projects portfolio has been contracting as a result o f a decline in extemal funding. Donors froze their commitments, citing lack of accountability and transparency in many projects' implementation and lack o f prudent management. However, more concrete data on funds use need to be produced to justify such statements. Perhaps a more important contribution would be the devising of an effective monitoring and implementation system that tracks allocated expenditures. Moreover, impact assessmentso f allocated expenditures can shed some light on leakages. 5.122 The only apparent leakage that can be demonstrated in the MoLFD is that occurring in the veterinary services department, because it has shown significant divergence between actual and targeted collections, as well as management problems. 94 However, efforts to capture any misallocations and leakages are again constrained by the lack o f a monitoring and evaluationunit. 5.123 Though evidence on leakages and misallocation under the M o C D M i s lacking, the existence o f divergences between budgeted and actual expenditures, with frequent budget outruns, i s a clear pointer to the existence o f misallocation. Further, with the liberalization of cooperatives, the ministryhas little control over its auditing systems and, as a result, the cooperative movement has beenbedeviledby mismanagement. FutureFundingRequirements 5.124 Given the heavy transfers to parastatals under the MOA,proposals made for the NCPB to draw its funding from outside the sector resource envelope should be pursued. Perhaps causes o f variations in use o f funds need to be understood and a number of budgetheads need to be rationalized before future funding opportunities are identified. It i s clear, however, that there is need to increase funding to core poverty programs in the ministry. 5.125 One apparent observationi s that the MoLFD i s allocated a very low share o f GoK expenditure-about 0.355 percent, not withstanding the sector's contribution to GDP of about 13 percent. There i s a need to identify future funding requirements, although it i s not possible at this stage to estimate the needed requirements. Future hnding requirements need to take into account the impact on staff retention, rationalization, and motivation. Although they are noncore to the ministry, other areas that require urgent funding are AI and disease and pest control for livestock, and increased monitoring, control, and surveillance are necessary for the fisheries department. 5.126 The M o C D M has to look for funds to hlly rehabilitate the cooperative movement. Specifically, there i s a need to fund the ongoing rehabilitation o f the K C C and KFA. There is also an urgent need to fund the department o f cooperative marketingand research to enable it to undertake market research and development and disseminate marketing information, thus strengthening the marketing function o f the ministry. Public-Private Roles in Financing 5.127 The level o f private financing in agriculture is low. The private sector has also failed to fill the gap left by the public sector's withdrawal from provision o f some services, and as a result the sector has suffered. There is a need to create a conducive environment to entice the private sector to fund agricultural initiatives. A good example in this regard is the provision of financial services to agriculture. Privately funded agricultural research and extension have also been hardto come by, and the public sector is forced to take up the whole burden inthe provision o f these services. 5.128 The hurried withdrawal o f the government from the support to animal health service delivery in AI, disease control, and dipping negatively affected the performance o f the sector. The private sector should have been prepared and motivated to take up these roles. In areas where private sector providers are present, the service i s expensive, 95 which culminates in farmers opting for traditional means. Functions appropriate to the private sector should be discontinued from government funding, though development o f appropriate exit plans for these services should be undertaken if efficiency i s to be improved. These functions include provision o f AI services (which are already partly privatized) and animal disease control, where private veterinarians are already practicing profitably in some areas. However, regulatory roles and critical services such as meat inspection should remain priority areas for the government. 5.129 The private sector has not been active in fimding initiatives under the MoCDM. Even in cases where the private sector is active, such as the auditing o f cooperative societies, little collaboration exists between the private and public sectors. There i s a need to initiate mechanisms for the private and public sectors to collaborate in funding initiatives under the ministry, such as inmarket research and information generation and dissemination. Conclusions 5.130 Agriculture still offers the best prospect for economic growth, and as such, more resources need to be directed towards this subsector if it i s to spearhead economic recovery. The ministries' allocations also need to be rationalized to enable a higher proportion o f the total to go to development expenditure. O f particular importance i s the need to increase the allocation to O&M and improve the efficiency o f use o f funds. It has also been established that large amounts o f funds are transferred to parastatals under the MOAwithout sufficient controls or monitoring, which at times weakens efficiency. Inthe future, parastatals need to prove that they add value to the agricultural sector by providing data on their outputs for inclusion inthe PER. Factors that hinder effective use o f funds, such as nonadherence to work plans, delayed release o f funds from the Treasury, accountability, and transparency, need to be addressed. Duplication o f efforts and fimding o fnoncore functions o fthe ministry shouldbe done away with. Some o f the recommendations that need to be implementedinclude: There i s a need for institutionalization and strengthening o f an effective monitoring and evaluation unitwithin the ministry. Budget monitoring needs to be strengthened. Fragmentationo fministries should be avoided. Duplication o f efforts should be eliminated. Spending units shouldjustify their allocations. Funds allocated to the NCPB should be considered outside the ministries' envelope duringresource bidding. Variations between budgeted and actual receipts o f AIA should be narrowed for effective implementation o fprograms. Accountability and transparency should be improved. There i s a need to fund the proposed legislative reform for the sector, KEPHIS, crop protection, and PCPB. 96 0 There i s a need to stop future funding o f noncore functions such as agricultural mechanization. 5.132 The allocation to the MoLFD i s dismal when compared with its contribution to the economy. However, the creation o f the ministry has not helpedimprove the quality o f information and data needed to assess the effectiveness o f resource use. Various weaknesses inthe ministry can be identified that stem from a lack o f capacity interms o f both numbers and skills, attributable to the retrenchment program. Similarly, the sector was negatively affected by a hurried withdrawal o f the government from provision o f some critical animal health services. As such, development o f appropriate exit plans for the remaining noncore services should be undertaken ifefficiency is to be improved. 5.133 Some recommendations for these sectors include: 0 There is a need for the development o f an effective monitoring and evaluation system. 0 The management o f the veterinary services department needs to be improved, to enhance its accountability and transparency. 0 An analysis o f the effectiveness o f the recent expenditure on marine surveillance should be carried out. 0 There is a need for development o f exit strategies from noncore activities. 0 More resources need to be allocated to this ministry. 0 Fundingto the KDB shouldbe reduced andreallocatedelsewhere. 0 The responsibility o f constructing access roads to fishing bays should be devolved to the Ministry o f Local Government. 5.134 The M o C D M has just gained full ministerial status from the MOA,and as such, there i s a need to finalize the ongoing legal reforms under the new ministry. Other specific recommendationswould include: There i s a need for sound financial management practices to cut out ad hoc public spending. There i s a need to adhere to set work plans and budgets. The Poverty Reduction Strategy Paper (PRSP) and MTEF process should be internalized at the field level for officers to emphasize the importance o f priority setting as a basis for public expenditure management. A monitoring andevaluation system needsto be set up. Data for public expenditure management ought to be improved to facilitate timely data capture for expenditure analysis. The MPND should train line ministries on MPERwork. Public expenditure management needs to be improved. 97 MINISTRYROADS, PUBLICWORKSANDHOUSING OF MissionandAdministrativeStructure 5.135 The mission o f the Ministry o f Roads, Public Works and Housing i s to facilitate provision and maintenance o f quality infrastructure, mainly in roads, buildings, housing, and other services related to public works,23 to promote and sustain socioeconomic development. Although the ministry takes charge o f three subsectors-roads, housing (including public buildings) and other public works-the roads subsector i s the dominant one interms o f spending. 5.136 As o f January 2004, the ministry had two state corporations under its administrative control, the Kenya Roads Board (KRB) and the National Housing Corporation (NHC). The KRB coordinates road maintenance and rehabilitation and development o f the road network in Kenya. It also advises the government on matters related to roads. 5.137 The KRB i s financed solely from the Road Maintenance Fuel Levy Fund (RMFL), which is o f the order o f K Sh 7-8 billion every year. The KRB is responsible for allocating RMFL finds to various agencies, including the ministry's roads department, based on a formula set by the KRB Act. More details on this allocation are given below. 5.138 The N H C i s responsible for providing housing services through mortgage, tenant purchase, and rental housing schemes for the public. The N H C does not receive any government transfers; however, the government recently converted debt owed by local authorities into equity. AggregateExpenditureTrends 5.139 In 2002/03, the total expenditure by the ministry fell substantially from the 2001/02 levels-from about K Sh 10 billion to K Sh 6.7 billion-because of a sharp decline inrecurrent expenditure inthat year (table 5.12). The fall was reflected mainly in the recurrent expenditure, which fell from K Sh 9.3 billion in2001/02 to K Sh 7.9 billion in2002/03. Unlike recurrent expenditure, development expenditure for this ministryhas fluctuated fairly significantly over the last four years, without showing any trend. 23"Other services" include the Kenya Building Research Centre, government housing, the Kenya Institute ofHighways andBuilding Technologies, and mechanical and transport services. 98 Table 5.12. Actual Expenditureof the Ministry,1999/2000-2003/04 19991 2000101 2001102 2002103 2003104 2000 (revised) Recurrent (K Shbillion) 8.4 7.8 9.3 7.9 10.8 Development (KSh 0.6 1.7 0.8 1,532 9.8 billion) Total (K Shbillion) 9.0 9.5 10.12 9.4 20.65 Total as % of GoK 5.7 6.0 3.6 5.0 8.0 expenditure Total as % of GDP 1.2 1.1 1.1 1.o 1.8 Source: BMD. 5.140 Table 5.13 shows expenditure by subvotes in the ministry. The roads subsector dominates the ministry's spending, with more than three-fourths o f the total. Recurrent finding for this subsector, which accounts for the bulk of total spendinginthe subsector every year, comes almost solely from the earmarked RMFL. The expenditure was maintained at the 2001/02 level o f 6.5 billion but declined as a share o f total expenditure. Table 5.13. Ministry of Roads, Public Works and Housing: Breakdown of Total ActualExpenditurebyFunctionKSh Million), 1999/2000-2002/03 Vote 199912000 2000/01 Rec. Dev. Total Rec. Dev. Total Rec. Dev. Total Rec. Dev. Total I 2001/02 2002103 Vote 130 General 522 administrativeplanning 619 619 620 620 522 (as % of total expenditure) 7 6 6 Vote 132Buildings andworks 162 162 548 242 790 149 25 174 (as % of total expenditure) 2 8 2 Vote 133 Other services 1683 68 1751 995 140 1135 1498 152 1650 s7O 38 608 (as % oftotal expenditure) 20 16 6 Vote 134 Housingevelopment 59 59 88 88 77 77 (as % oftotal expenditure) 1 I 1 I 1 I 6,353 1,469 8,002 Vote 136 Roads 5,896 491 6,387 5,983 1,564 7,547 /6,506 428 6,934 (as % o f total expenditure) 71 79 69 Total 8,419 559 8,978 b,842 1,704 9,546 19,260 822 10,082b,851 1,532 9,383 I Source: Ministry of Public Works andHousingMPER 2003 5.141 Table 5.14 shows the breakdown o f expenditure from another perspective-by economic classification. O&M accounts for about 85 percent o f total expenditure. Table 5.14. Ministry of Roads, Public Works and Housing Shares: Total Expenditureby EconomicClassification,1999/00-2002/03 (KSh million) 199912000 2000101 2001102 2002103 Personnel 1,346 1,166 1,277 1,245 Operations 993 1,236 676 250 Purchases 26 106 191 56 Maintenance 6,014 5,285 7,07 1 6,259 Transfers 40 49 45 41 Total 8,419 7,842 9,260 7,851 99 199912000 2000101 2001102 2002103 As % of total expenditure Personnel 16.0 14.9 13.8 15.9 Operations 11.8 15.8 7.3 3.2 Purchases 0.3 1.4 2.1 0.7 Maintenance 71.4 67.4 76.4 79.7 Transfers 0.5 0.6 0.5 0.5 Source: Ministry of Roads, PublicWorks andHousingMPER2003. Roads 5.142 Seen from a longer-term perspective, recurrent spending on roads has increased substantially over the last decade. As shown in table 5.15, recurrent expenditure on the classified roads rose from US$41.3 million in 1991/92 to US$89.9 million in2001/02 (in constant U.S. dollars at 2001 prices). The main reason for this substantial increase in recurrent road expenditure is the introduction o f the RMFL in 1994. This earmarked fund is impressivelyhandled and has been a steady source o frevenues. Table 5.15. Expenditure on the Classified Roads Sector, 1991/92-2001/02 (US$ Million, 2001 Prices) Recurrent Development Total expenditure expenditure expenditure 1991192 41.3 115.5 156.8 1992193 38.9 104.1 143.0 1993194 29.1 68.3 97.4 1994195 34.2 29.5 63.6 1995196 60.2 65.5 125.7 1996197 81.0 100.3 181.3 1997198 96.5 58.1 154.6 1998199 85.8 51.3 137.1 19992000 89.2 46.7 135.9 2000101 87.7 37.3 125.0 2001102 89.9 47.7 137.6 Source: World Bank 2004b 5,143 However, development expenditure in the roads sector has fallen substantially, due largely to a sharp reduction in donor funding. As table 5.15 shows, development expenditure on roads fell sharply in real terms, from US$115.5 million in 1991 to US$47.7 million in 2001. The net result i s that overall expenditure on roads fell marginally, from US$156.8 million in 1991to US$137.6 million in2001. TheKRB 5.144 The KRB, which began full-fledged operations only in 2002/03, is in charge o f coordination of all maintenance activities on Kenya's roads. It allocates RMFL revenues according to this formula: (a) 57 percent to the roads department o f the ministry for highways and the largest roads (of classes A, B and C); (b) at least 24 percent to the district road committees (DRCs), which are central government agencies, for roads o f 100 classes D and below; (c) 16 percent to the Parliamentary constituencies, also for roads o f classes D andbelow; and (d) 3 percent or less for administrative expenses o f the KRB. 5.145 A substantial portion o f allocation meant for DRCs goes to other agencies that are seen as handlingthese allocations on behalf o f the DRCs. These are the roads department o f the Ministryo f Roads, Public Works and Housing, which receives allocations to finish maintenance works that had already been started under its command; local government authorities; and the Kenya Wildlife Service (KWS), which receives allocations for roads inparks andwildlife areas. All ofthese allocations are meant for roads of classes D and below, that is, for rural roads and urban roads other than highways andmajor roads. Only funding that goes through the roads department is reflected in expenditure by the Ministry o f Roads, Public Works and Housing. 5.146 All recipients o f RMFL revenues are only allowed to implement work plans that have been approved by the KRB, which has the right to delay or withhold funding for agencies that do not perform as expected or that do not follow KRB directives. 5.147 Table 5.16 shows the details o f RMFL disbursements inthe last three years. The year 2002/03 shows a sharp fall indisbursements, due partly to low revenue collection in that year, which stemmed, in turn, from fuel supply disruptions. In addition, disbursements in the last quarter o f that year were partially held up, with the additional revenues disbursed in2003/04, instead. Table 5.16. Disbursementsof RMFLRevenues,2001/02-2003/04 (Ksh Million) 2001102 2002103 2003104 (untilJune 9) Roads department for roads o f classes A, B, and C 4,208 3,816 4,103 Roads department for roads o f classes D and below (from 1,344 710 1,034 allocation meant for DRCs; part o f this to be passed on to DRCs andpart to be handledbythe roads Department) Local governments (from allocation meant for DRCs) 894 514 924 KWS (from allocation meant for DRCs) 30 69 Parliamentary constituencies 1,175 840 1,470 Administrative expenses of the IWB 235 228 202 Total 7,858 6,139 7,752 Source: Ministryof Roads, Public Works andHousingMPER 2003 5.148 DRCs tend to focus on rural roads and do not see urban roads as their responsibility. Local governments take charge o f maintenance o f urban roads networks (other than major roads), which they h d from RMFL revenues, LATF revenues, and their own sources. The overall level o f funding for the urban roads o f classes D and below (those handled by the local governments) appears low, as discussed below. I s Spendingin the RoadsSubsector Sufficient,and How EffectiveI s It? 5.149 Kenyahas a large road network about 195,000 kilometers long, approximately 6 percent o f which i s paved (table 5.17). A significant part o f the network has fallen into disrepair during the past decade o f neglect. The government i s currently carrying out an 101 inventory and survey o f conditions o f the classified roads. Preliminary data show that about 43 percent of the classified roads network i s inpoor condition. Table 5.17. The State of Roadsin Kenya(Kilometers) Paved Gravel Earth Total Classified roads 8,937 27,181 27,172 63,290 Urbanroads 2,491 12,937 0 15,428 Countylcouncil 7 116,254 0 116,261 Total 11,435 156,372 27,172 194,979 Source: World Bank 2004b. 5.150 Maintenanceneeds. According to a recent World Bank an estimated K Sh 9.4 billion are ideally needed per year for road maintenance. Current spendingon road maintenance, funded from the RMFL, appears to be nearly sufficient. Despite the near- sufficient funding for roadmaintenance, road quality remains poor and heavily criticized, with much evidence o f misallocation or misuse o f resources. There is evidence that Kenya can do much better with its current allocation o f resources to road maintenance. To supplement the RMFL resources, the idea o f setting aside revenue from vehicle licenses for road maintenance could also be explored. 5.151 Capitalinvestmentneeds. The ERS estimates that about K Sh 19 billion will be requiredto reduce the share o f classified roads inpoor condition from 43 percent to 20 percent. It lists other projects worth about K Sh 62 billion as priority projects to be carried out duringthe 2003-07 period (table 5.18). Additional fundingwould be desirable to upgrade heavily trafficked unpaved roads. It i s very unlikelythat the GoK can provide fhding of this magnitude within the time required. Apart from a significant increase in donor funding, there will be a needto explore other sources. Table 5.18. PriorityCapitalRequirementsinthe RoadsSector, 2003-07 Improvement (KShbillion) Construction andrehabilitation of the roadnetwork 19.2 Rehabilitation and upgrading of key road links 11.0 Rehabilitation of rural access roads (Roads 2000 Programme) 10.8 Concessioning o f the Mombassa-Malibi Highway 15.0 Development of roads under East African Roads Network Programme 11.5 Rehabilitation of roads innational parks 5.5 Improvingroads inthe ASAL 3.2 Construction o f northernand southernbypasses inNairobi and Mombassa 4.8 Total 81.0 Source: Kenya2004. 5.152 If government and external funding are not adequate for capital expenditures, an alternative i s to use private sector finance, serviced either directly through tolls or through an increased fuel levy over and above what is required to cover maintenance expenditure. Long-term performance-based rehabilitation and maintenance contracts can be hndedthrough an agreed schedule o f annual payments or on the basis o f shadow tolls. 24World Bank (2003b). 102 Private sector capital funding may appear expensive, but it can be viable on heavily trafficked roads. 5.153 Given the shortfalls in overall funding, Kenya will have to carefully prioritize in its use o f resources and make better use o f existing funds. A carefully thought-out road strategy is needed; this should include proposals to better leverage and attract funding from donors. Inaddition, ways to involve the private sector to a greater degree-and ina manner that enhances incentives for good performance-are needed. If done appropriately, this would enhance efficiency, as well as increase funding for the sector. How Can Maintenance Funds Be UsedMore Effectively 5.154 There i s a need for a more systematic and coherent approach toward road maintenance and a better system o f allocation o f road contracts, including steps taken to stop overcommitting available resources. There does not appear to be a coherent road maintenance program; instead, the system appears to consist o f emergency responses on failed roads. When the public sector does undertake maintenance work, it i s largely through the employment o f day labor supervised by ministry foremen and technicians. Quality control and cost-effectiveness are often poor, and there i s a tendency to award contracts far inexcess o f available funding, leading to higher operational costs because o f delays, court claims, interest penalties, and so on. Inaddition, there does not seem to be a transparent and competitive process for the awarding o f contracts. 5.155 One way o f improving the allocation o f funds for road maintenance i s to allocate a larger portion o f the RMFL revenues to paved urban roads other than highways and major roads (that is, to paved urban roads o f classes D and below). Among roads o f classes D and below, there i s too little funding for paved urban roads relative to rural roads and unpaved urban roads; the former are relatively heavily trafficked and should receive an amount o f funding that i s more in line with their heavy level o f usage. As noted above, the main agencies that handled paved urban roads o f classes D and below are the local governments; the DRCs tend not to see urban roads as their responsibility. Thus, one solution to the current state o f affairs would be to increase the portion of RMFLfunds going to the local governments. 5.156 One reason for the lack o f coordination i s that there are a number o f sources o f funding for road maintenance other then RMFL revenues, such as revenues from the LATF.Yet works funded bythese other sources are done independently o fthe KRB.This worsens the problem o f a lack o f coordination; ideally, the KRE3 should be in charge o f coordinating and approving the work plans o f all road maintenance efforts, including those not funded from RMFLrevenues. 5.157 The private sector could be more involved in maintenance, and a change in the types o f contracts awarded would give contractors the incentive to improve performance substantially. A separate regulatory highways agency may need to be created. With the current system, which focuses on relatively short-term contracts, contractors are not responsible for road performance beyond the initial construction, reconstruction, or rehabilitation stage; they are not involved much in road maintenance. Thus, they do not 103 have an incentive to ensure that their work i s o f high quality in the longer run. It would be better to rely on long-term performance contracting, where contractors are also put in charge o f road maintenance and are paid not according to the work they perform (perhaps beyond the initial capital phase), but by the results they achieve over time. Altematively, they could be paid to achieve specified road standards and penalized for not doing so. Additional regulatory capacity would be needed, and hence a separate highways agency may needto be created. 5.158 Rural road spending needs to be focused on partial rehabilitation and spot improvements to bring the network to a maintainable standard and improve the level o f access during the rainy season, rather than on full rehabilitation or upgrading. This is in recognition o f the fact that, currently, most o f the funding goes to the rehabilitation or upgrading o f a very few rural roads, with most o f the rest o f the network neglected. This i s partly a result o f the political appeal o f such a strategy. Yet, from an economic efficiency and equity viewpoint, a more broad-based and less intensive approach would be preferable. Budget Execution 5.159 Pending bills continue to be a problem, although they were recently reduced from K Sh 8.1 billion to K Sh 1.8 billion with special help from the Treasury. The pending bills arise largely from overcommitment o f available resources, mainly for additional works or increases in scope in existing works that tend to be approved by the ministry without beingsanctioned by the Treasury. Problems with Structure of Salary and Wagesfor Employees 5.160 The amount spent on maintenance o f staff housing in the Ministry of Roads, Public Works and Housing i s too high. This should be reduced or ceased, with civil servants possibly asked to obtain mortgages with government assistance, instead. 5.161 There i s a need to change the formula for payment made by the ministry to architects, engineers, and other design consultants. Currently, this appears to be a percentage o f the estimated cost o f the proposed projects; this i s expensive and needs to be rethought. Conclusions 0 The allocation o f RMFLrevenues needs to be modified so that more finding goes to paved urban roads that are not classified as major roads (that is, paved urban roads o f classes D and below). One way o f doing this i s to increase the portion o f the revenues going to localgovemments. 0 The private sector needs to be involved to a greater degree inroad maintenance, and contracts awarded need to be better designed, with contractors involved not just in the initial construction, reconstruction, or rehabilitation stage, but also in subsequent road maintenance. In addition, the contracts should be performance 104 based. All this would maximize their incentive to ensure high-quality work inthe longer run. A separate regulatory highways agency may need to be created to oversee these contracts. Rural road spending needs to be focused on partial rehabilitation and spot improvements rather than on full rehabilitation or upgrading. All pending bills arisingfrom development projects should be settled ina phased three-year program to be financed within the development ceilings. Meanwhile, the ministryshould introduce line itemsinits budgetsfor clearingpendingbills. The ministry should emphasize the identified core poverty programs, such as Slum Upgrading and Roads 2000, because o f the effect they have in directly alleviating poverty. Inliaison with the Ministry of Finance, the ministry shouldimmediately review all externally funded projects with a view to reducing or stabilizing (in Kenyan shilling terms), re-scheduling or canceling nonviable projects, and reallocating donor finds to higher-priority projects. N o new projects should be started, other than 100 percent donor-financed projects, until the ministry has made adequate financial provision for existing projects according to the implementationprogram. The government should cease or modify its current policy o f funding the maintenance o f staff housing, with a view to reducing or ceasing such expenditures. A proposal should be developed to help civil servants obtain mortgages through the proposed civil servants' housing scheme to purchase government houses. There is a need to change the formula for payment to architects, engineers, and other design consultants, whose pay seems to be based on a percentage o f the estimated cost o f the proposedprojects. MINISTRYWATERRESOURCESMANAGEMENT OF AND DEVELOPMENT Structureof the Ministryand CurrentReformsandRestructuring 5.162 Through the 2002 Water Act, the Ministryo f Water Resources Management and Development was established as a separate entity in January 2003.. It was placed in charge o f a range o f functions related to water resources management. These include water resources management policy, water and sanitation services, water quality and pollution control, dam construction, flood control and land reclamation, wastewater treatment and disposal, water conservation, and irrigation. Before January 2003, the current Ministry o f Water Resources Management and Development and Ministry o f Environment, Natural Resources and Wildlife operated as a single ministry. 5.163 The ERS lists a number o f priorities related to the water sector, including (a) strengthening infrastructure to increase access to water resources, (b) enhancement o f private sector participation in the sector, (c) structural reforms to make water and sanitation provision autonomous, and (d) partnerships with community-based organizations to expand services to poor and rural communities. It specifically mentions implementationo f the 2002 Water Act as fundamental to achieving the stated goals. 105 5.164 The 2002 Water Act separates the ministry's water resources management functions from those o f water services provision. Many o f these functions are currently handled, or will be, by autonomous or semiautonomous bodies, with the ministry handling largely policy formulation and coordination. The autonomous and semiautonomous boards currently include the NWCPC National Irrigation Board, Kenya Water Institute, Water Services Management Authority, and Water Services Trust Fund. 5.165 The ministry i s still undergoing restructuring, which i s expected to continue over the next two fiscal years. At the same time, the ministry has committed itself to implementing the reforms laid out inthe 2002 Water Act. 5.166 As provided in the 2002 Water Act, the ministry i s finalizing a Water Sector Investment Plan to be implemented over a three-year period as the main vehicle for realizing the institutional reforms and infrastructure development in the sector. The investmentplan adopts a programmatic approach to sector development. The focus is on improving management o f water resources and extending water services coverage in the country, especiallyto the poor. 5.167 The investment plan i s expected to cost K Sh 8-10 billion annually over its three- year-period. This i s much more than the K Sh 4.2 billion set aside for the ministry inthe development budget, that is, largely from donor funding (revised estimates) for 2003/04. Thus, substantial increases in donor funding will be needed to meet the plan's requirements, because it will be very unlikely that the GoK can meet the additional fundingrequirements on its own. Already, at least K Sh 12.24 billion has beenpromised by bilateral and multilateral donors for the sector, to be made available from July 2004 onward. For the investment plan targets to be met in each o f the three years o f its expected duration, sustained efforts will be needed to attract sufficient donor funding over the next three years. 5.168 In the longer run, however, the ministry's resource requirements are expected to fall drastically once the reform and rehabilitation works are completed. At that point, the ministry' role i s expected to shift to policy formulation and coordination, not investment. Challengesinthe Sector andExpenditurePriorities Kenya faces a natural legacy of very limited freshwater resources and rainfall variability. It receives a renewable supply of freshwater o f less than 650 m3per person per year making it one o f the most water-scarce countries in the world. Much o f the rain falls in less than 20% o fthe country; the rest is aridand semi-arid. 5.170 Kenya has a legacy from the recent past o f poor management o f water resources. The country's water resources have been mismanaged through unsustainable water and land use policies, laws, and institutions; weak water allocation practices; growing pollution; and increasing degradation o f rivers, lakes, wetlands, and aquifers and their catchments. Infrastructure i s mostly old and dilapidated, and the country's irrigation potential is 540,000 hectares, but only 87,350 hectares have been developed. 106 5.171 The use o fwater occurs inmany different sectors, thus a multisectoral approach i s needed to manage water sector reform. This coordinated multisectoral approach ideally would involve such sectors as livestock, agriculture, energy, and manufacturing, among others. 5.172 The ministry faces two major challenges in the next three years. First, the decreasing water resources have to be better managed and have to be allocated more equitably to the various (competing) users. Second, the country's national fresh water resources must be sufficiently developed to meet growing demand and make it more accessible to Kenya's urban andrural populations. 5.173 The govemment has identifiedthree priority areas for its investment: 0 Water and catchment conservation, protectionand construction o f dams 0 Construction and rehabilitation o furban and rural water supplies 0 Construction and rehabilitation o f irrigation and drainage infrastructure for agriculture. AggregateExpenditureTrends Table 5.19. Trends of Combined Expenditure by All Departments and Agencies Belongingto the MinistryfromJanuary2003 Onward(K Sh billion) 2000101 2001/02 2002103 Recurrent 1.6 2.5 2.2 Development 1.6 1.2 1.o Total 3.3 3.7 3.2 Total as % of GoK 1.9 2.2 1.7 Total as % of GDP 0.39 0.4 0.3 1 Source; Ministryof Water resources. 5.174 Because the ministry began operations in its present form only in January 2003, aggregate expenditure trends should be analyzed by looking at the combined spending over time o f all entities belonging to the ministry from January 2003 onward. This i s what was done in computing the figures for tables 5.19, 5.20, and 5.21. 5.175 Aggregate expenditure for the departments and agencies o f the newly formed ministry is very small relative to total expenditure. Aggregate expenditure for these entities summedto less than 2 percent o f total GoK expenditure annually inrecent years, as shown intable 5.20. 107 Table 5.20. Three-year Breakdown of Expenditure (Actual) by Function for All MinistryDepartmentsandAgencies fro January 2003 Onward KSh Million) I 2000101 2001102 2002103 Recurrent Develop- ment Total Recurrent Develop- ment Total General administration andplanning 125 500 625 221 283 504 462 24 486 Rural, urban, and Special Water programs 1,326 647 1,973 2,123 551 2,674 1,587 516 2,103 Flood control, irrigation, and land reclamation 54 104 158 75 15 90 71 33 104 National Irrigation Board 122 10 132 91 85 176 82 61 143 NWCPC - 312 312 - 251 251 - 341 341 KenyaWater Institute 23 35 58 23 6 30 24 5 29 Total 1,649 1,609 3.258 2,534 1,191 3,725 2,225 980 3,206 -Notavailable. Note: The table i s a breakdown of the data by subvote. There is not necessarily a one-to-one correspondence between subvotes and departments or agencies; some subvotes incorporate funding to two or more entities or agencies Table 5.21. Total(Actual) Expenditureby EconomicClassification,2000/01-2002/03 (KSh Million) 2000101 2001102 2002103 Salaries and wages- 926 1,149 1,178 O&M 718 1,381 1,040 Plant and eauiDment 6 4 8 2000/01 and 2002/03 for the sub vote^^^ that were a part of the ministry from January 5.176 Recurrent expenditure did not show a significant rising or falling trend between 2003 onward, except for the National IrrigationBoard, for which it fell, leading to staff and operations spending cuts. As shown intable 5.20, recurrent funding for the National Irrigation Board fell from K Sh 122 million in 2000/01 to K Sh 9lmillion in 2001/02, then remained low at K Sh 82 million in2002/03. This ledto (a) a retrenchment of 569 National Irrigation Board staff in February 2002, resulting in total employees of fewer than 100; (b) a stalling o f irrigation and cropping operations in Ahero, West Kano, and Bunyala; and (c) an accumulation of the National Irrigation Board's pending bills, which currently stand at K Sh 282 million. The reasons for this fall in recurrent funding to the National IrrigationBoard, and its implications, need to be examined more closely; if the lower levels imply an excessively low level o f spending on key operations or maintenanceactivities, this may be cause for concern. 5.177 Over the period 2000/01-2002/03, expenditure on salaries and wages remained approximately constant. Spending on O&M, although not apparently low relative to salaries ad wages, showed significant fluctuations over the same period. As table 5.21 25Fundingduring the year is allocated onthe basis of subvotes, where each subvote consists of one or more departments or agencies within eachministry. 108 shows, expenditure on salaries andwages rose from K Sh 926 million in2000/01 to about K Sh 1.1billion inboth 2001/02 and 2002/03. Spending on O&M rose from K Sh 718 million in 2000/01 to K Sh 1.3 billion in 2001/02 and then fell to K Sh 1.0 billion in 2002/03. These are not low amounts relative to the figures for salary and wage payments, but the reasons for the fluctuations need to be examinedmore closely. EfficiencyandEffectivenessofExpenditures 5.178 Although a substantial increase in funding is needed for the sector, especially to implement its infrastructure enhancement needs, it i s also fundamental to ensure that existing resources are used in a more efficient manner. This implies the allocation o f water resources to the uses that have the largest benefits at the margin, in the most cost- effective manner possible, while ensuring that equity and environmental goals are met. Some steps that need to be taken to achieve this are discussed below. 5.179 The procedures for the application and processing o f water permits have been lengthy and ineffective, promoting illegal abstractions; enforcement has also been hampered by a lack o f funding, insufficient data, inadequate cooperation among water users and managers, and corrupt practices. Greater funding and better management are needed for enforcement o f compliance with water permits and improvement inthe permit applicationprocess. 5.180 Allocation would be improved by reinvigoration o f the water resources monitoring network; this would facilitate the allocation o f water on a factual and transparent basis. Costs can be reducedby usingmodern recording and communications technologies, such as automated flow-monitoring equipment and data transmissions that do not require regular visits by staff. 5.18 1 Water user charges (for raw water) and pollution penalties were introduced inthe 2002 Water Act to support some o f the key water resource management functions. It i s imperative that these charges are handledappropriately. The charges should be usedonly for their intended purpose, water resources management, and not to support other activities such as funding water supply and sanitation schemes. In irrigation districts, O&M activities need to include prompt repairs o f leaks in distribution systems to build confidence among those paying water user charges that their contributions are being used effectively. 5,182 Fulluse should be made o f measures such as water conservation, water reuse, and demand-sidemeasures to enhance rationalization inthe use o fwater. It is not enoughjust to increase infrastructure investment and implement other supply-side measures; demand- side measures are essential as well. 5.183 As o f December 2003, the ministryhad total pendingbills o f K Sh 806 million, o f which 68 percent are in recurrent and 32 percent in development expenditures. The recurrent expenditure bills are largely for utilities services, such as water, electricity, telephone, and so forth. 109 5.184 There are a total o f 25 programs andprojects inthe ministry, 17 o fwhich are core poverty programs. None o f these 25 projects are stalled, but there are a number o f weaknesses inproject implementation. These are largely common to other ministries as well; they include limited or delayed cash releases, cash hnd problems at district treasuries, lack o f monitoring and evaluation, weak and slow expenditure reporting systems, andthe inability o fbeneficiaries to sustain project activities. 5.185 There i s a need to substantially improve monitoring and evaluation o f the ministry's programs and projects. The ministry's Central Planning Unit, in conjunction with the Finance Unit and the implementingunits, has the responsibility o f undertaking monitoring and e; yet this role has not been effectively carried out because of the lack o f adequate personnel in the Unit.The only information available about the efficiency and effectiveness o f expenditure currently comes from the implementing units; yet these are often not objective in evaluating their own performance.. To address this, the ministryi s intheprocess ofestablishing its ownMonitoring andEvaluationUnit. 5.186 There are a number o f weaknesses in the situation with human resources, including inadequate numbers o f professional and technical staff and inadequate training. A training needs assessment should ideally be undertaken regularly to determine the training needs and wants o f the technical officers and professional officers in the ministry. The training should focus on technical and managerial skills and address support staff. 5.187 Performance indicators are being developed to stimulate water resources management reforms, such as assessing the linkage betweenprovision o f water services and poverty reduction. Yet the lack o f consistent and reliable data i s a barrier to the development and use o f good perfonnance indicators. The situation in data management i s now so badthat data are often recorded and handled manually. 5.188 Monitoring o f water quality would also be enhanced if the water resources monitoring network were reinvigorated. This was discussed before as a means to improve the water allocation process; aside from this, it would also facilitate water quality monitoring. 5.189 Conflicts could be reduced if catchment area advisory committees and river water use associations were encouraged as a priority. Legislation should ideally be introduced to provide these with the necessary authority. Conclusions 0 Development spending will need to be raised substantially to meet the sector investment plan requirement o f K Sh 8-10 billion over its three-year period. The revised estimate o f K Sh 4.2 billion for 2003/04 is already much more than the levels o f previous years, and already at least K Sh 12.24 billon have beenpledged , by donors from July 2004 onward, but a sustained effort will continue to be 110 needed to attract sufficient donor funding over the next three years to meet the goals o f the investmentplan. There seems to be a lack o f consistency in spendingon O&M, and the reasons for this needto belookedat inmore detail. The fall in recurrent funding for the National Irrigation Board needs to be examined more closely. If this implies excessively low spending on O&M activities, this may be cause for concern. Enhance funding for, and administration of, the process by which water permits are allocated and enforced, so that delays are reduced and there i s better enforcement. The water resources monitoring network needs to be reinvigorated to ensure better monitoring of water quality and water use (with the latter leading to better and more rational allocation). Water user charges (for raw water) need to be handled appropriately. For example, they should only be used for their intended purpose, water resources management, rather than other activities. Supply-side measures are not enough; full use should be made o f measures such as water conservation, water reuse, and demand-side measures to enhance rationalization inthe use o fwater. Monitoring and evaluation need to be strengthened to effectively observe and assess the implementation o f projects and programs against identified performance indicators. A new Monitoring and Evaluation unit i s to be set up by the ministry, and this needs to be adequately funded. Data keeping and storage need to be significantly improved, and the ministry needs to use information technology more effectively for this purpose. This will facilitate the monitoring o f the performance indicators that are currently being developed and facilitate overall monitoring and evaluation. Establish catchment area advisory committees andriver water use associations. 111 6. SUMMARY OF KEY RECOMMENDATIONSAND THE PER PROCESSIN2004/05 6.1 A number o f recommendations have been made at appropriate places in the text close to analysis and findings. The next section pulls together the key set o f recommendations that should form the reform program in the medium term and be monitored through regular mechanisms o fthe PER Steering Committee. DEVELOPING IMPLEMENTINGA CREDIBLEMEDIUM-TERMFISCALSTRATEGY AND 6.2 To mitigate the risks o f and meet the challenges to the medium-term fiscal strategy, the following actions are recommended: To counter the unpredictability of the volume o f external aid, continuous dialogue with development partners focused on strengthening information sharing and harmonizing and streamlining the practices andprerequisites for aid should be vigorously pursued. Meanwhile, budgeting should be based on conservative estimates o f external aid based on past experience. When ' assigning development ceilings, a planning reserve could be kept unassigned, to be assigned later, close to budget finalization, based on better information about aid availability. The fiscal impact o f the health system change (National Social Health Insurance Scheme [NSHIS]) is likely to be large but has not been carehlly worked out. The decision to implement the NSHIS should be contingent upon a full understanding o f design and realization issues and the public expenditure costs. Debate on the NSHIS could be carried out inthe context o f developing a fiscally sustainable health sector strategy. There i s an urgent need to systematically collect information on financial operations o f public enterprises and local authorities with a view to remain alert to, and adequately plan for, their demands on the central budget (contingent liabilities). This should also allow the government to monitor the consolidated public sector deficit apart from the fiscal deficit o f the central government, which i s currently monitored. There i s a need to revisit the proposedwage enhancements for teachers. 6.3 The medium-term fiscal framework should be revised based on a hardheaded look at feasible tax reforms and their likely yield. It would be prudent to compute financing requirements on the relatively more conservative increase in revenues. On external assistance, it would be better to explore what might be a plausible range o f outcomes. 112 Some alternative scenarios, including such alternative debt reduction rates as stabilizing debt ratios, should be explored. 6.4 An explicit Cabinet endorsement o fthe medium-term fiscal framework andpolicy assumptions behind the framework (as part o f the Budget Outlook Paper around October and a Budget Strategy Paper inMarch) would be essential for providing credibility to the overall resource envelope and ministerial ceilings. More civil society engagement in budget strategy would help to build support for difficult decisions. PRIORITIZING EXPENDITURE ACHIEVE ERS OBJECTIVES TO 6.5 The government cannot rely only on additional resources becoming available to restructure spending. Much can be done within the existing resource envelope, but that would require hard budget choices to be made. Specifically, the attempts should be focused on: (a) Curbing unproductive general administration expenditure, which has been rising over the past few years: The ongoing functional reviews should be used to identify such expenditure. These reviews, however, need to be carried out with a clear objective to cut expenditure, rather than an open- ended identification o f "core" functions without an explicit budget constraint. The government should revisit the reorganization o f ministries from an administrative cost andpolicy coordination point o f view. (b) To limit transfers to parastatals, the strategy for restructuring or privatizing for-profit parastatals should be implementedvigorously. There i s also a need to review the rationale for the existence o f many semiautonomous government agencies that carry out regulatory, regional development, and social service hnctions. Many o f these can be merged with parent ministries or closed down in a future reorganization. The review can be conducted as an extension o f the ongoing functional review inministries or a stand-alone study. 6.6 Mucho fthe expenditureprioritization withinkey line ministries should take place around core performance indicators being identified as part o f the ERS monitoring and evaluation framework. The process would involve setting targets for these core indicators, then identifying inputs in the form o f programs and interventions-the core priority programs-that are "essential" for achieving the targeted results. The identified core priority programs should thenbe funded sufficiently and predictably. This process o f identifying the results-to-inputs chain and providing adequate funding is likely to be an iterative one. Once identified and fully accepted by all concerned parties, the core priority programs could take the place of the core poverty programs used at present to prioritize expenditure. 113 6.7 For other programs not covered by the core priority programs, the Ministry o f Finance should take the lead, jointly with line ministries, in developing O&M norms linked to maintenance and service delivery standards. 6.8 Stalled projects identified for completion should be given priority over starting new projects. Other than donor-financed projects, no new projects shouldbe started until there i s adequate financial provision for existing projects. The target should be to resolve all stalledprojects by 2006-07. 6.9 N o w that a full cycle o f wage enhancements has been completed, the government should develop a medium-term pay policy and set up accompanying institutional mechanisms to replace ad hoc processes for wage adjustments. STRENGTHENING PEM 6.10 The government is updating the Enhanced Financial Management Action Plan in the context o f the PEM-AAP update. This updated action plan will provide the overall framework for strengthening PEM. The following paragraphs recount some important areas o f reform that should be given urgent attention. 6.11 Integrating the MTEF and annual budgeting processes. The most effective way to integrate the MTEF and annual budget processes would be to bringthem together institutionally inthe Treasury, as well as within line ministries. 6.12 Reforming the budget formulation process. Key new processes that emerge from the findings o f the MTEFReview are: (a) Preparation o f a Budget Outlook Paper inOctober (b) Early preparation o f MPERdMTEF submissions during October to December (c) Developing a Budget Strategy PaperMTEF using MPERsMTEF submissions and any other analytical inputs by March. The Budget Strategy Paper should be approved by the Cabinet and include ministerial ceilings. (d) Dissemination o f the Budget Strategy Paper to stakeholders and civil society. 6.13 The important elements of the proposed process reforms are engagement o f the political executive and senior management inthe following year's budget priorities early inthe budget cycle and development of a credible statement of spending priorities and ministerial ceilings that are seen as bindingby all concerned. 6.14 Functioning and constitution of Sector Working Groups. The role and hnctioning o f SWGs should be sharpened. Their terms o f reference should be more narrowlydefined, keepinginmindthat SWGs are coordination instruments and not self- standing government bodies with regular staff or a secretariat. There is also a need to ensure systematic interface with stakeholders inSWGs outside the budget cycle. 114 6.15 Expenditure arrears. Although settling old expenditure arrears (from 6efore June 2003) should no doubt be a priority, there i s also an urgent need to implement an action plan to ensure that new arrears do not accumulate. Such an action plan would include, among other things (a) announcing a transparent policy for clearing accumulated arrears and introducing the rules and regulations to prevent accumulation o f fresh arrears, (b) reviewing contract administration procedures, (c) establishing a project implementation and monitoring system, (d) institutingprocurement laws and regulations, and (e) establishing a system for managingpendingbills. RECOMMENDATIONS FOR SPECIFIC MINISTRIES Ministry of Education, Science and Technology 6.16 Drawing on the proposed staffing norms study, the options and costs o f both addressing the immediate imbalances in teacher-student ratios, as well as the long-term requirements for increased teacher numbers, should be quantified. Inthe meantime, there should be increased funding for temporary teachers to replace those that are absent because o f long-term illness and in-service teacher training for the current primary teachers. 6.17 An analysis should also be undertaken to examine the possibility o f the decentralization o f teacher recruitment, as well as management structures and incentives that facilitate a more efficient use o f teachers. The results o f this analysis are needed to address the shortage o f teachers in an optimal manner. In the meantime, increased funding i s probablyjustified for temporary teachers inareas where there are excessively high student-teacher-ratios and in-service training for current teachers to adjust to the new curriculum. 6.18 The secondary education strategy, including policies on (a) bursaries, (b) the establishment o f new schools, and (c) the appropriate proportion of primary students passing on to secondary education, shouldbe developed. 6.19 The following should be developed for the university subsector: (a) a strategy for expansion and public financing, (b) expenditure projections that take into account the proposed salary increases for university lecturers, and (c) identification o f effective means o fproviding incentives to universities to respondto national priorities. 6.20 The plans to significantly increase fhding for TVET should be suspended until a thorough study o f effectiveness o f spending i s carried out and options for greater on-the- job, employer-financed, andprivate sector training are explored. Ministry of Health 6.21 Immediate recommendations for expenditure allocations inthe coming budget are the same as those made in previous years: to increase the allocations for spending on lower-level service delivery facilities and staff and drugs andcontain spendingon the two 115 central hospitals. But, according to current evidence, there i s little basis for determining the optimal levels o f spending. 6.22 The PER work o f the past two years has not been sufficient to allow definitive decisions on reallocation of expenditure in the health sector. In the coming year, the following program o fwork should beundertaken: (a) A more detailed analysis o f actual spending compared with budgets at a disaggregated level by line item within each program (b) A specialized study to determine what services are actually being delivered at each level and type o f facility (health posts, provincial and district hospitals, and the Kenyatta National and M o i Referral Hospitals), including an assessment o f time use and expenditure within the facilities to estimate the real distribution o fcurrent spending by level o fcare (c) An analysis o f use o f facilities and the reasons for utilization and bypassing behavior (d) An assessment o f absorptive capacity constraints related to drugs, staffing, and flow o f funds to identify whether it makes sense to increase financing and the barriers that need to be overcome to do so (this could draw upon existing or ongoing work within the health sector) (e) A true expenditure tracking surveythat traces the flow o f specific "buckets" o f funds (and in the case o f drugs and medical supplies, commodities) through the system, to determine where leakage is taking place and its magnitude (f) Analysis o f the impact o f health fees on exclusion o f the poor and possible reforms to compensate for this (g) A risk analysis o fthe proposedhealthinsurance scheme. 6.23 It is not realistic to expect this work to be done as part o f a PER exercise conducted over a fairly short time in the runup to the budget. Work should start early as part o f a broader consolidated effort geared toward health sector budget reform, which can lay the basis for both expenditure reform and a common program o f donor support for increased health sector spending. If this is done well, bythe time o fthe next PER, the government should be ina position to make informed decisions regarding (a) revised health funding levels for the next budget, (b) proposed reallocations, and (c) complementary measures needed-for example, in institutional reforms, staffing and incentives, systemic or management changes, and so forth. RECOMMENDATIONSFORTHE PER PROCESSIN2004/05 6.24 The following recommendations are made to strengthen the PER process and its elements in2004-05: 116 (a) The terms o f reference for MPERs should have both generic elements and customized elements. Examples o f generic elements include (a) analysis o f updated and reconciled expenditure data at the subvote level, explaining any significant changes in expenditure trends; (b) reporting on performance indicators; (c) identifying policy changes, new programs and projects, and their expenditure implications inthe mediumterm and for the next budget year; and (d) reporting on the functional reviews and identifying the scope for savings over the medium term. The customized elements would be specific to each ministry and would be agreed upon between the PER Secretariat and the ministry at the beginning o f the process in September.' Ways should be found to merge MTEF submissions with MPERs. Line ministries should not have to make two submissions on expenditure strategy for one budget. (b) The calendar of the PERprocess should be completely harmonized with the annual budget and MTEF. The PER Report should be prepared in end- March or early April so that there can be a good discussion o f its findings among all stakeholders. (c) Action on PER recommendations shouldbe monitored through the Steering Committee. Short- and medium-term action plans could be developed. (d) Jointly with line ministries, the Budget MonitoringDivision should develop a historical database on expenditures by subvote. This will allow consistent analysis o f expenditures, notwithstanding several reorganizations o f ministries. Line ministries should use the same data for their own analysis in MPERs. (e) Apart from other issues, the next PER could focus generally on the issues o f efficiency o f public service delivery, the performance orientation o f budgeting(how it can be made more results-based), the extent o f contingent liabilities, the costs and benefits o f reorganization o f ministries, and a systematic review o f the main findings o f ongoing functional reviews inkey ministries. 'Ofcourse some ministries that serve mainly administrative finctions may not have any customized elements. 117 ANNEXES 118 Annex 1.DeficitImplicationsof AlternativeDebtTargets Sustainability i s intimately tied to the government's intertemporal budget constraint, which has a straightforward interpretation. Ultimately, expenditures have to be paid for, even though deficit financing enables this to be deferred for a time. However, deferral comes at a price, because interest rates are positive. Developing this simple insight involves some algebra, which is not reproduced here. What emerges from this algebra i s a very simple relation: Ifthe initial ratio o f aggregate net present value (NPV) o f debt to gross domestic product (GDP) i s bythe commercial interest rate i s r, and the economy i s growing at g, then the primary Heavily Indebted Poor Countries Initiative (HIPC)- consistent deficit, adjusted for seigniorage(s), must satisfy the equation: where6is the percentage rate o f decrease in b-that is, it i s positive if it i s intended to lower this ratio, zero if it is intended to maintain it, and negative if it i s intended to raise it. All this refers to the overall debt ratio. There is also the issue o f composition. The government's avowed intention is to use net external borrowing to retire domestic debt. The computations in a debt sustainability paper prepared by the International Monetary Fund in 2003 envisages a slow rise in the ratio o f NPV o f external debt to GDP up to 2010, accompanied by a more rapid fall in the domestic ratio, so that the overall ratio declines steadily up to 2010, and this decline accelerates thereafter. With this background, consider the current and likely figures for Kenya. First, consider the figures in the IMF debt sustainability paper. The NPV o f aggregate government debt (external and internal) is given as equal to 53.5 percent o f GDP in 2002. The nonconcessional rate o f interest used appears to be 5.12 percent. We assume that the Kenyan shilling exchange rate drifts slowly down so that this i s effectively the real interest rate. Meanwhile, the govemment's own real growth projections average around 3.8 percent over the next three years. Hence, the required adjusted primary surplus to GDP ratio would be 0.7 percent to stabilize the debt ratio, or 2.3 percent to reduce it at 3 percent per year, the rate that is built into the early years o f the IMFdocument.' The second step i s to consider what i s implied by the government's current projections. Figures are from the Economic Recovery Strategy for Wealth Employment and Creation 2003-2007 (ERS); they are averaged over the coming three-year period and given in percent o f GDP. On this basis, revenue i s projected at 20.3 percent o f GDP, outright grants at 1.9 percent. Expenditure on goods and services averages 24.7 percent, and on interest, 2.2 percent. '(.051 - .038).535 = .013 x .535 = .007 for the constant ratio; (.013 + .03).535 = .023 for the case where the debt ratio i s reduced at 3 percent per year. 119 Hence, the conventional overall deficit after grants i s 4.7 percent, and the primary deficit i s 2.5 percent. However, the conventional deficit needs to be adjusted for the grant content o f external borrowing (which might run a high as 1.6 percent) so that the HIPC- consistent primary deficit would be only 0.9 percent. Seigniorage will cover 0.6 percent, or two-thirds, o f this, leaving the adjusted primary deficit at 0.3 percent of GDP. This wouldimplya rise inthe overall debt ratio of around 2 percent per year.2 Assuming this outcome is unsatisfactory, there are three possible types o f acc~mmodation.~For concreteness, suppose that the goal i s to (a) stabilize the debt ratio or (b) reduce it at 3 percent per year as inthe debt sustainability document. The first type o f accommodationis arise ingrant or other concessional finance. The scale o f the increase ingrants would have to be equal to 1.O percent o f GDP for stabilization or 2.6 percent for reduction-roughly a 50 percent increase or a 150 percent increase, respectively, on current projections. If this mechanism were to work via concessional finance with a 50 percent grant element, these increases would need to be doubled. The second type o f accommodation is an effort to raise domestic revenue. The govemment is committed to maintaining revenue at 21 percent o f GDP. This would be worth an average 0.7 percent o f GDP, roughly two-thirds o f the gap for stabilizing the debt ratio. To close it completely would require revenue to be raised to 21.3 percent o f GDP or to 22.9 percent for debt reduction. The third type o f accommodation is a reduction in expenditure on goods and services. Stabilization would require this to be cut from 24.7 percent to 23.7 percent o f GDP; debt reductionwould require it to be cut to 22.1 percent. All of these accommodations are severe, particularly because they are the required average outturns over the next three years. O f course, inpractice, they may be combined invarious proportions, reducingthe severity of each taken separately. But even so, there appears to be a major challenge to the design o f fiscal strategy inKenya. It remains to briefly consider the debt composition issue. Under the projections quoted here, the NPV o f external debt would fall relative to GDP. Hence, the option of, for example, maintaining the current aggregate debt to GDP ratio would involve a firther rise in domestic debt relative to GDP. Ineffect, there are additional problems concerning composition andthose attaching to the aggregate that have been discussed here. -.0031.535 - .013 =-.02. We assume that interest payments are not susceptible to muchpolicy-induced change inthe short run. 120 Annex 2. Medium-TermFiscal Frameworkand Strategy inthe ERS Economic Prospects Table 1shows the economic forecasts for the period 2003-07 for key economic variables provided in the ERS. This forecast i s consistent with the expected external environment, especially trading partner growth and price levels, prices o f key commodity exports and imports, and expected growth in external support, which will be used to boost govemment investment and reduce domestic financing needs. Economic growth i s expected to rise from 1.8 percent in 2003 to 4.9 percent in 2007. Investment growth i s expected to be the primary driver for economic growth: investment levels are expected to rise from 14.6 percent o f GDP to 24.8 percent during this period. Domestic savings are expected to rise from 14.9 percent to 19.7 percent over the period, implyingthat there will beneed for substantial external inflows for the investment levels to be achieved. Private consumption i s expected to be low as the govemment focuses priority expenditure on investments and social services and the private sector keeps the proportion o f profit disbursement low inpreference for investments. Exports and imports are expected to be a major factor inthe growth outcome. Currently, export earnings are constrained by low prices o f commodities that dominate Kenya's exports. The ERS has proposed policy incentives considered sufficient to spur export growth and efficiency gains in production to improve competitiveness. Imports are mainly driven by the performance o f the other components o f the GDP, especially investments and exports. Table 1.Forecastsfor Key EconomicParameters 2003 2004 2005 2006 2007 GDP growth (% real) 1.8 2.5 3.7 4.1 4.9 Investments (% o f GDP) 14.6 20.4 23.3 24.4 24.8 Savings (% of GDP) 14.9 15.7 17.6 18.5 19.7 Export growth (% annual) 8.2 2.0 4.8 9.9 12.3 Growth inprivate consumption (% annual) 14.7 -9.4 -3.7 -1.3 0.2 Import growth (% annual) 9.4 18.2 10.5 9.3 8.8 Underlying inflation (%) 9.8 3.5 3.5 3.5 3.5 Source: Fiscal StrategyPaper, 20031042006107, MPND. Monetary Framework Kenya's monetary policy focuses on maintaining stability in the general price level and a well-functioning and stable market-based financial system. The current policy stance targets keeping overall inflation below 5 percent annually while targeting underlying inflation of 3.5 percent. Inthe framework, broad money supply (M3X) will be the intermediate target, and reserve money will serve as the operating target. A conservative monetary policy will allow Kenya's underlying inflation to remainbelow 3.5 percent, comparable with the forecasts for trading partner inflation rates. This will 121 allow for the maintenance of a stable nominal exchange rate policy without riskingreal appreciation of the Kenyan shilling. Low inflation and extemalization o f the deficit are expected to allow the rate o f interest on government debt to remain below 6 percent inthe medium term. Over 2003-07, money supply i s projected to rise at 8.1 percent per year, consistent with the projected real economic growth (table 2). Both net foreign assets and domestic credit to the economy will support the money supply growth projected. Net foreign assets o f the bankingsystem are projectedto rise by 13.6 percent, while domestic credit is expected to rise by 4.9 percent by December 2007. Private sector credit i s programmed to accelerate by 14.9 percent, while credit to the government will decline by 37.6 percent inthe years untilDecember 2007. The government will graduallyreduce its borrowing from domestic sources, paving the way for increased lending to the private sector. A flexible exchange policy pursuedby the monetary authority is fundamental to ensuring that the country's balance o f payments remains sustainable and the country has an adequate level o f foreign exchange reserves. Exchange rate interventions will be limited to smoothing short-term volatility and the variability o f donor flows, affecting external debt payments andmaintaining a sound level o fthe net internationalreserves. Table 2. MonetarySurvey(AnnualGrowthRates) Jun. 03 Dec. 03 Jun. 04 Dec. 04 Jun. 05 Dec. 05 Jun. 06 Dec. 06 Jun. 07 Dec. 07 1.Net foreign assets 6.6 8.4 13.6 16.9 19.9 25.5 30.2 22.5 16.5 13.6 2.Net domestic assets 12.4 6.5 5.0 4.3 3.6 1.1 -1.3 1.1 3.6 5.0 3.Domestic credit 12.0 7.4 4.5 4.1 3.7 1.6 -0.4 1.7 3.7 4.9 4.Government(net) 30.7 16.2 3.9 -1.8 -7.3 -14.4 -22.0 -25.8 -30.5 -37.6 5.Rest oftheeconomy 4.9 3.6 4.8 6.9 8.9 8.6 8.3 11.1 13.7 14.7 6. Otherpublicsector -4.7 -19.8 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 7. Private 5.2 4.4 4.9 7.0 9.0 8.7 8.5 11.3 13.9 14.9 8. Otheritems(net) 10.6 11.7 2.3 3.1 4.0 4.1 4.2 4.3 4.5 4.7 9. M3X 10.9 7.0 7.1 7.5 7.9 7.9 7.9 8.0 8.1 8.1 Source: CentralBank of Kenya. Revenue Projections To achieve the fiscal objective o fmaintaining the revenue to GDP ratio above 21 percent, the Kenya Revenue Authority plans to implement significant tax and administrative reforms. The objectives o f tax reform are to improve transparency and efficiency o f taxation, strengthen tax collection, expand the tax base, and harmonize tax systems within the East African Community (EAC). The keyreforms include expansiono fthe tax base, effective enforcement in revenue administration, computerization o f the tax administration, strategic capacity building, harmonizing and rationalizing tax regimes, integrating trade arrangements within the EAC and the Common Market for Eastem and Southern Africa, and implementing a regional transit cargo control strategy and strict enforcement o fthe Traffic Act. 122 Table 3. Proiected Revenue and Grants. 2003/04-2006/07 2003104 2004/05 2005106 2006107 Total revenue 237,993 247,876 256,218 267,457 Grants 20,831 22,5 16 27,269 25,258 Total revenue and grants 258,824 270,392 283,487 292,715 RevenueiGDP 21.2 20.3 19.5 18.9 GrantsiGDP 1.9 1.8 2.1 1.8 Revenue and grantsiGDP 23.1 22.1 21.6 20.7 Source: IMF and Kenyan authorities. Table 3 shows revenue and grants projections in the medium term. Revenues are expected to continue falling ExpenditureProjections In view of the declining performance on the revenue side, the bulk of Kenya's fiscal strategy will need to focus on expenditure reduction, restructuring, and reform. Whereas Kenya's revenue performance at more than 21 percent o f GDP has been above par for low-income countries, Kenya's expenditure levels at more than 26 percent o f GDP have been significantly above that for low-income countries, yet the public investments level, at less than 2.5 percent o f GDP, i s below the recent Sub-Saharan African performance. The country also has a poor public expenditure management (PEM) record, with only 3 o f 16 PEMindicators considered acceptable. The forecast expenditure path impliedby the revenue forecasts above is detailed intable 4. Overall expenditure i s expected to rise from 26.2 percent o f GDP to 26.5 percent of GDPby2006/07. More resources will be shifted toward the development vote (from 15.7 percent in2003/04 to 25.1 percent in 2006/07) and core poverty (from 8.3 percent to 8.6 percent over the same period). The civil service reform program is expected to lead to a reduction in the wage bill from 32.5 percent to 31.4 percent; interest payments are expected to decline (from 11.3 percent to 7.6 percent), as are transfers to parastatals and universities (from 5.4 percent to 5.0 percent)-all considered benefits o f the anticipated reforms. Table 4. Projected Expenditure Ratios, 2003/04-2006/07 2003/04 2004105 2005/06 2006107 Wagesiexp. ("A) 32.53 31.27 32.27 31.35 O&Mlexp. (%) 31.03 29.77 29.13 28.17 Transfersiexp. (%) 9.16 9.27 7.62 7.73 Interesvexp.(%) 11.35 8.34 8.17 7.6 Dev.1exp. (%) 15.66 19.72 22.81 25.15 Core poverty programsiexp. (%) 8.34 8.36 8.72 8.64 Transfers to parastatals & universitiesiexp. (%I 5.49 5.14 5.24 5.03 Wagedrevenue (%) 37.93 38.55 36.34 35.63 Core povertyprograms (% o f GDP) 3.93 4.0 4.0 4.0 Development expendituresiGDP (%) 4.24 5.42 6.01 6.69 Wages/GDP ("A) 8.8 8.6 8.5 8.3 Source; Fiscal StrategyPaper MPND. 123 The above projections are based on the following underlyingassumptions andmeasures: Wages are expected to decline from 8.7 percent o f GDP in2003/04 to 8.5 percent by 2005/06. In2003/04, the government awarded substantial increases to the disciplined forces (police and prison staff) and began implementation o f the teachers' salary award. N o awards have as yet beengiven to civil servants. Achieving the 8.5 percent target by 2005/06 will require that any awards to be provided to the civil servants or any additional awards (other than the remaining phases o f the teachers' awards) will be matched by a proportionate downsizing o f the public service. To this end, 21,388 civil servants are forecast to leave the service over a period o f three years through a Targeted Voluntary Early Retirement Scheme (TVERS) to enable civil service wages to be raised as part o f an overall pay and benefits strategy. The fiscal strategy for interest payments i s based on the premise that the government will be able to raise substantial additional external resources and thus allow for a reduction inboth domestic borrowing levels andinterest rates. The Central Bankwill also continue with its policy o f lengtheningthe term structure o f domestic debt, Nonwage ministerial recurrent expenditures (excluding transfers to subvented bodies) depend on exchequer releases and appropriations in aid (AM) expenditures. The strategy assumes that these expenditures will be maintained constant in real terms, other than for education, health, and core poverty expenditures. Transfers to parastatals and universities are expected to be maintained at their 2003/04 levels in real terms, with any increases in personnel emoluments being catered for through downsizing. Kenya currently operates a noncontributory pension scheme. However, pension payments have been rising at an unsustainable rate, and there will be a need to reduce these expenditures. The government is planning to introduce a contributory pension scheme in the 2004/05 fiscal year. The fiscal strategy assumes that pension payments will continue to rise at 15 percent per year, but an early introduction o f the contributory scheme could lower this growth. The main objective o f reprioritizing and selecting the core poverty programs was to allocate and ring-fence resources to those programs that have a direct impact on reducing poverty levels in the medium term by improving access o f the poor to infrastructural and social services (especially rural infrastructural services, education, and health), enhancing the capacity o f the poor to participate in productive service activities, and improving the governance and security environment. The major components o f guaranteed loans and capital transfers have tended to be government repayments o f parastatal debt. However, the government has adopted an ambitious strategy to restructure the National Bank o fKenya. It i s proposedthat over 2003/04 and2005/06, K Sh 12billion be used inthis endeavor. The Government is committed to raising the level o f health expenditures to at least 12 percent o f total expenditures as part o f its strategy for providing adequate health services. Overall recurrent health expenditures were 6.7 percent o f total recurrent expenditures in 2002/03, and achieving the 12 percent target by 2010 will require that recurrent health expenditures to grow by at least 7.5 percent faster than overall expenditures. The fiscal strategy assumes that these health expenditures will be 124 focused on nonwage, nontransfer expenditures and will thus enable the rapid increase inbasic health services. Inthe recent past, development expenditures inKenyahave declined as apercentage o f total expenditures and as a percentage of GDP. Implementingthe ERS will require a massive increase in development expenditures. The strategy has an ambitious target o f raising development expenditures from K Sh 43.3 billion to K Sh 92.7 billion, an annual average o f 23.6 percent. Over 2003/04-2006/07, a total o f K Sh 280.0 billion will be available for implementing development projects inthe ERS. Over the 1 9 9 0 Kenyabuilt up a substantial inventory o f stalled projects as lax fiscal ~ ~ control led to the initiation o f more projects than could be sustained by the development budget. As a result, by 1997, the completion rate for projects was as low as 3 percent. Stalled projects-those projects that were initiated but not completed and are currently not receiving funding-mushroomed. By 1999, the government had a total o f 164 stalled projects with an estimated original contract cost of K Sh 31.4 billion, accrued expenditures o f K Sh 13.3 billion, and an estimated completion cost o f K Sh 13.2 billion for those that could be completed and K Sh 2.4 billion for those that required termination. By 2003, the stalled projects were estimated to have increased to 207. The government will review all stalled projects with a view to terminating as many as possible and completing only those that fall in the ERS core priority areas or that have potential for meeting ERS core objectives. It i s estimated that reviewing the stalled projects will cost K Sh 90 million. Deficit Financing Kenya's external aid policy provides a framework for effective multilateral and bilateral cooperation geared toward increasing external inflows and improvement o f disbursement rates. The current fiscal strategy has built insignificant foreign assistance needs to fillthe resource gap needed to implement the ERS. Financing the deficit will focus on maximizing external concessional borrowing consistent with maintaining the external debt to GDP ratio o f less than 36 percent. External financing over 2003/04-2006/07 i s expected to total K Sh 198.3 billion, of which K Sh 26.6 billion will be for rescheduling. Net external borrowing over 2003/04-2006/07 will total K Sh 121.7 billion (table 5). Table 5. Deficit and Financing Projections-Some Key Ratios, 2003/04-2006/07 2003104 2004105 2005106 2006107 Overalldeficit/GDP (%) -2.5 -6.6 -2.9 -2.2 Net foreignborrowing(as % of GDP) -0.1 0.4 0.9 0.8 Netdomesticborrowing(as % of GDP) 2.8 1.8 2.3 1.4 Netdomesticdebt/GDP (YO) 24.9 23.4 21.7 20.0 Source: IMF andKenyanauthorities. To close the financing gap, domestic financing will total K Sh 251.8 billion over 2003/04-2006/07. Over and above financing the existing domestic debt (rolling over), a net domestic borrowing o fK Sh 41.4 billion will be accessedthrough additional domestic borrowing. 125 The outcome of the deficit and financing strategy i s expected to be an external debt to GDP ratio that remains below the 36 percent level achieved in 2002/03 and a domestic debtto GDP levelthat declines from 24.3 percent in2002/03 to 20.8 percent in2006/07. 126 Annex 3. Changesinthe Pay andEmploymentStructureof CivilService The civil service inKenya i s organized injob groups A to Y. Statistical annex table A.19 gives a detailed breakdown o f the number o f civil servants in position by job group in 2002. As shown in figure 1, the current structure o f the civil service i s concentrated around levels F to H.As a result o f the Civil Service ReformProgramme, the structure o f the civil service has changed inmajor ways over the last decade. Before the retrenchment program was introduced, job groups A to C constituted the highest numbers in the civil service, accounting for close to 33.3 percent o f the total number o f filled positions. Steps were taken duringthe retrenchment program to address the bottom-heavy structure o f the civil service. As a result, the share o fjob groups A to C has been declining steadily, and by February 2004, they accounted for 6.6 percent o f total central government employment. The share o f employment o fjob groups D to E has also declined, and job groups A to E account for 18.7 percent o f total employment, compared with 52.4 percent in1990. Figure1. Changes in Pay andEmploymentStructure of Civil Service between 1985 and 2003 1985 1993 abc de f gh jkl m-z abc de f gh jkl m-z Jot lob groups Employment 40 1 abc de f gh jklm-z Job groups 2003 The share of job group F in total public sector employment has been rising, and by February 2004, it accounted for 33.4 percent o f the total. However, the majority o f those injob group F (87 percent) are insecurity (figure 1and statisticalannex table A.19), just as the majority o fmedical and other government personnel are concentrated injob groups Gto J. 127 Annex 4. Status of MillenniumDevelopmentGoalsin Kenya Table 6 shows the status o f MDGs for Kenya. There are eight main goals, relating to reduced income poverty, improved basic health and education outcomes, gender equality, andaccess to safe water. Table 6. MDGsfor Kenya Baseline 2000 levela Target Goalby 2015 (1990) (2015) Halve the proportion o f people living o n less than $1 43.3% 51.8% 21.7% per day Achieve 100%primary school completion 63 (1986) 81 (2001) 100 Achieve gender equity ineducation Reduce under-age-five mortality by two-thirds 98.9 111.5 33.0 Reduce maternal mortality by three-quarters 590 - 147 Halt and begin to reverse the spread of HIV/AIDS Halt and begin to reverse the incidence of malaria and other major diseases Halve the proportion o fpeople without access to safe 45% 57% 72% drinlungwater a. Datato bereplaced (or complemented) with most recent measuredlevels for indicators. Achievement o f the poverty MDG will be difficult in the absence o f much more robust growth than has characterized the recent past inKenya. The determinants o f this are only partially related to public expenditure. They depend as much or more on the policy environment, governance, and external factors, including terms o f trade and weather. Nonetheless, investmentsininfrastructure, education, andinselected public interventions inareas such as agriculture are certainly going to beneeded. To quantify these, however, and to definitively link the levels o f spending with poverty outcomes, is very difficult. Kenya should be able to achieve the education MDGs. After the introduction o f free primary education, the system appears to be on track to reach close to 100 percent primary school enrollment, and it should subsequently be able to reach 100 percent completion. To do this, however, will require improving quality to keep children in school. The proposed measures discussed inchapter 5 for example, to reallocate teachers, relieve classroom overcrowding, and support schools in slums and low-income areas- should make this possible. The hnding requirements are not likely to exceed the levels foreseen inthe projections. The health-related MDGs will be difficult to achieve, not least because conditions have deteriorated since 1990 (the somewhat arbitrary base year against which the MDGs targets are defined), thus the magnitude o fprogress required is substantially larger than in some other countries. It i s difficult to quantify the direct public spending implications, in part because many determinants o f health outcomes-food and nutritional intake, hygiene, and patient behavior, for example-lie outside the scope o f direct public health service interventions. Nonetheless, there are a number o f areas in which public services can have a significant impact on the health MDGs: immunization, family planning 128 services, and malaria control being among the most obvious, but even in these cases, what i s required i s not only (or even mostly) public spending, but adequate staffing and management o f outreach services. As pointed out inchapter 5, increased spendingon the health sector alone will not materially contribute to achieving these basic health outcomes unless it i s directed to primary health care interventions and accompanied by associated measures to improve service delivery management and absorptive capacity. The water supply MDG would potentially be achievable, but not without substantial additional investment.Note that there are ambiguities with respect to the definitions o f effective coverage used that make it difficult to compare estimates o f current coverage and the level o f expansion needed by 2015 to reach the target. But there i s no doubt that substantial increases in coverage would be required to attain the MDG. Chapter 5 discusses the Water Sector Investment Plan that Kenya i s about to embark on, at an estimated cost o f K Sh 8-10 billion annually for three years. If implemented as planned, this will go a longway toward increasingthe proportionofthe populationwith safewater access. It will be vital to attract sufficient donor fundingbecause the government will be able to fund only a small portion o f the required investments (see chapter 5). O f equal importance to investments inwater supply will be the associated measures to ensure that the assets created are kept functioning through community mobilization for hnding and managing operations and maintenance. 129 Annex 5. MTEFSector Working Groups:CompositionandTerms of Reference Composition of SWGs The membership o f each sector i s as follows One chairperson-ne Permanent Secretary to be chosenby consensus One sector convener-Ministry o fPlanningandNational Development Secretariat for the sector Budget Supply Deparbnent representative ExternalResources Departmentrepresentative MTEFPRSP Secretariat representative Relevant ministries for the sector Developmentpartners Private sector Civil society Eight SWGs are organized as follows: (4Agriculture and Rural Development Agriculture (b> Livestock Development andFisheries Development (4 Cooperative Development andMarketing (4 Environment andNaturalResources (e> Lands andSettlement Physical Infrastructure (a> Roads, Public Works and Housing (b) Energy (4 Transport andCommunications (d) Water Resources Management andDevelopment (e> Local Government Human ResourceDevelopment (a> Health (b> Education, Science andTechnology (4 Gender, Sports, Culture and Social Services (d) Labour andHumanResource Development Trade, Industry and Tourism (a> (4 Trade andIndustry Information and Tourism Public Administration (a> Finance 130 Planning and National Development Directorate o f PersonnelManagement Foreign Affairs Regional Development Public Service Commission Controller and Auditor General National Assembly Electoral Commission. Public Safety, Law and Order (a> Judicial Department (b) Justice and Constitutional Affairs (4 Office o f the President (d) State House (e) Office o f the Vice President and Ministry o f Home Affairs (f) Office o f the Attorney General National Security (a) Department o f Defence (b) National Security Intelligence Service (4Information Communication Technology Finance (b) Planning and National Development (c) Environment and Natural Resources (d) Information and Tourism (e> Transport and Communications Terms of Reference of SWGs 1, Review the MTEFbudget and establish cross-sector reallocations anddeviations. 2. Define and articulate the sector clearly, including establishing spending needs for purposes o f the MTEFprocess. 3. Coordinate all activities leading to the development o f a sector-wide plan and production o f a sector report. 4. Identifysector development objectives. 5. Identifysector strategies inlight o f the known constraints. 6. Identifythe overall sector priority or priorities. 7. Analyze cost implications o f policies and strategies in the sector and identification o f sector-wide resource needs, including potential sources o f funds. 8. Identifypotential outcomes inlight o f the constraints, that is, identify key outputs and outcomes o f sector priority strategies and ensure that they are incorporated withinthe sector-wide plan. 131 9. Identify inter- and intrasectoral linkages. 10. Identify existing priorities and programs within the sector to tackle the sector priorities. 11,Identify activities inthe sector requiredto operationalize strategies for key cross- sectoral issues. 12. Identify performance targets, monitorable indicators, and modalities to monitor them and suggest responsible agencies. 13. Ensure that expenditure allocations at all levels are in line with the subsector plans, which have been derived from the sector-aide plan and priorities. Thereafter, individual vote plans must be translated into subvotes or heads for resource allocationpurposes. 14.Justify all proposed projects inthe sector investment program. 15, Monitor implementation o f the budget, including participation in the midterm review for the sector. 132 Annex 6. List of FundAccounts Establishedunderthe ExchequerandAudit Act Iffice of the President I(a) MechanicalTransport and Plant Renewals Fund(National . , Youth Service [NYS]) 1 (b) MechanicalTransport andPlant Maintenance Fund(NYS) . . (c) Government Press Fund I(d) Strategic Grain Reserve Trust Fund ~ dinistryo fHome Affairs (a) Prisons Industries Fund (b) PrisonFarmRevolving Fund dinistryof Finance (a) Government Clearing Agency Fund (b) Treasury MainClearance Fund (c) Kenya Local Loans Support Fund (d) Civil Contingencies Fund (e) PetroleumDevelopment Levy Fund (0RuralEnterprisesFund (g) Exchange RiskAssumption Fund (h)ProvidentFund (i) European Widows and Orphans Scheme 6) Asian Officers Family Pensions FundAccount (k)Asiatic Widows and Orphans PensionFund I( 1) RuralDevelopment Fund (m) Constituency DevelopmentFund IMinistryo fAgriculture, Livestock andFisheries (a) Hides and Skins Cess Fund . . (b) Demonstration Farms Fund (c) Agricultural Information Centre Revolving- Fund _ , - (d) Veterinary Vice DevelopmentFund Ministry of Cooperation and Marketing ,(a) Development Management Supervision and Liquidation Fund IMinistry o f Health \(a) Health Care Services Fund I(b)Medical Supplies Fund I .. Ministry o f Roads and Public Works I(a) Stores and Services Fund (b) RoadMaintenance LevyFund Ministry o f Water Resources and Development (a) Mombasa Water Works Renewals Fund . . (b) Minor Water Works RenewalFund (c) Water Service Fund National Assembly (a) Parliamentary Mortgage Scheme Fund (b)NationalAssembly Car LoanFund Ministry o fEnergy (a) Rural Electrification Programme Fund Electoral Commission o f Kenya (a) Electoral Commission Car Loan Scheme Fund Ministry o f Lands and Settlement (a) Township, Roads and Drains Fund Ministryof LocalGovernment (a) Local Authorities Transfer Fund 133 , t -d 4 3 x I -r -r-I- n m a P I v, m I. ~ x1 eA -A c P 2 I3Im STATISTICAL ANNEX 138 Table A.l. Summary of Key Macroeconomic Indicators 1998 1999 2000 2001 2002 2003 GDP (market prices) (K Sh 596.6 billion) 639.1 685.4 767.4 850.0 968.4 Real GDP growth rate (%) 1.8 1.4 -0.2 1.2 1.2 1.8 GDPper capita (constant, 1982 3,556 3,527 3,425 3,399 3,362 3,348 prices, K Sh) Gross investment/GDP (%) 20.1 18.8 17.9 16.7 15.2 14.6 Populationgrowth rate (%) 2.5 2.4 2.4 2.0 2.3 2.2 Consumer price index (%) 6.6 5.8 10.0 5.8 2.0 9.8 Average interest rate for 91-day 11.1 20.5 13.5 10.8 8.4 3.6 Treasury Bills (%) Growth inmoney supply (M3) (%)a 3.3 2.8 0.8 2.6 8.6 0.6 Growth indomestic credit (%) 7.1 5.9 2.1 0.3 7.9 -0.1 Imports (annual % change) 3.7 4.4 20.1 17.1 -11.2 9.4 Exports (annual % change) 0.6 1.1 9.8 9.7 14.7 8.2 Current account balance (as % o f GDP) 4.1 -1 .o -2.6 -3.9 -1.6 0.5 Exchange rate: K Sh/US$ (end o f 61.9 72.9 78.0 78.6 77.1 76.1 Year) Total domestic debtIGDP (%)b 28.2 26.7 30.1 27.6 27.8 29.9 Total external debt/GDP (%)b 42.6 53.8 52.9 47.7 39.4 36.4 a Broad money, M3, i s money supplied by the Central Bank, Commercial Banks and Non-bank Financial Institutions (NBIFs). Central govemment outstanding debt. Source: Central Bureau of Statistics, Economic Survey 2004, Central Bank of Kenya. Table A.2. Real GDP Growth by Sectors, 2000-03 (Constant 1982 Prices) 2000 2001 2002 2003 Non-monetary 1.2 1.7 1.7 1.7 , Monetary -0.3 1.2 1.1 1.8 Enterprises and non-profit institutions -0.6 1.3 1.1 1.8 Agriculture -2.1 1.3 0.8 1.5 Forestry -2.0 0.9 0.8 0.8 Fishing -2.1 0.8 0.5 3.O Miningandquarrying 0.9 1.o 0.8 2.5 Manufacturing -1.4 0.8 1.2 1.4 Buildingand construction -1.5 -0.5 0.3 2.2 Electricity and water -4.1 1.5 1.2 2.0 Trade, restaurants, and hotels 1.o 1.3 1.6 1.4 Transport, storage, and communication 2.0 3.2 2.6 1.5 Finance, insurance, real estate, business services 0.4 1.o 0.8 3.0 Ownership o f dwellings 1.4 1.8 1.5 1.7 Other services 0.5 1.o 1.o 1.8 Less: imputed bank charges 1.2 1.9 2.0 0.9 Private households (domestic services) 2.4 2.6 2.3 2.4 Producers o f government services 0.7 0.7 0.9 1.6 Total GDP -0.2 1.2 1.2 1.8 Source: Central Bureau o f Statistics, Economic Survey 2004. 139 Table A.3. Central GovernmentFinancialOperations 2000/01-2003/04 (K Sh2003104 Million) 2000/01 2001/02 2002103 actual actual actual projected 192.313 197,768 210,750 237,993 ~Revenue Income tax 55;829 59;034 70,452 76,735 Import duty (net) 28,726 21,584 18,477 20,910 Excise duty 28,318 32,077 35,643 41,464 Value-added tax 50,298 50,871 56,135 57,719 Investment income 2,478 767 1,243 4,221 Other 12,232 14,585 12,349 16,107 AIA 14,432 18,850 16,45 1 20,837 Expenditureand netlending 232,621 228,980 265,947 286,980 Recurrent expenditure 198,641 204,026 222,421 243,925 Interest payments 31,035 30,384 36,026 32,064 Domestic interest 23,232 23,744 27,567 25,645 Foreigninterest due 7,803 6,640 8,459 6,419 Wages and benefits (civil service) 68,119 78,125 85,087 96,340 Civil service reform 6,095 1,537 0 200 Pensions, etc. 6,136 8,995 9,450 14,067 Other 52,465 61,235 69,367 78,045 Ofwhich: operations and maintenance 33,531 39,157 44,101 52,153 Ofwhich: Transfers 14,974 16,978 18,359 21,490 Defense and NSIS (National Security Intelligence 16,847 21,980 22,933 23,209 Service)a Pending billsb 1,972 1,770 -442 0 Drought relief expenditures 15,972 0 0 0 Development and net lending 33,980 24,954 43,526 43,055 Domestically financed 5,685 10,873 12,778 15,957 Foreign financed 15,341 15,381 21,951 25,730 Net lending 3,957 1,862 1,669 1,368 Pending billsb 3,191 -3,162 7,128 0 Drought Development Expenditure 5,806 0 0 0 Civil Contingency Fund 0 0 0 0 Balance(commitment basis,excludinggrants) -40,308 -31,212 -55,197 -48,987 Grants 24,080 6,823 14,942 20,831 Foodrelief grants 12,444 0 0 0 Project grants 5,681 5,350 14,484 16,448 Program grants 5,955 1,473 458 4,383 Balance(commitmentbasis, includinggrants) -16,228 -24,389 -40,255 -28,156 Adjustmentsto cashbasis 6,650 -2,832 7,163 518 Balance(cashbasis,includinggrants) -9,578 -27,221 -33,092 -27,638 Financing 9,575 26,621 32,101 27,638 Net foreign financing 12,489 -1 1,250 -10,340 -1,383 Project loans 14,042 10,03 1 7,467 9,282 Program loans 4,045 0 0 7,191 Financial defense lease loan adjustment 0 2,064 1,803 2,515 Repayments due -33,277 -26,156 -20,724 -19,429 Change inarrears 5,886 2,811 1,114 -6,114 Rescheduling 21,793 0 0 5,172 Privatization proceeds 0 955 0 0 Bank restructuring costs 0 0 0 -2,500 Expenditure arrears securitization -3,538 -2,788 4,48 1 0 Net domestic borrowing 624 39,704 46,922 31,521 Ofwhich: excluding bank restructuring costs 624 39,704 46,922 29,021 and excluding expenditure arrears securitization 4,162 42,492 51,403 29,021 Financinggap (stat. discrepancyfor outturns) 4 600 991 0 Memorandum items Total project support 32,167 15,381 21,951 25,730 Total identified and unidentified gross extemal support 42,167 16,854 22,409 37,304 Nominal GDP 839,534 926,039 1,018,188 1,123,561 Primarv budeet balance 21,457 3,163 2,934 4,426 Stock i f doiestic debt, net (end o fperiod) 164;203 202;775 247,706 279;227 aDifferfrom authorities' numbers because o f different accounting treatment o f finance defense leases. bThe fiscal accounts are on a cashbasis (with the exception o f foreign interest due). Adding accumulation o fpending bills and subtractingcash repayment o f them adjusts to a commitment basis. Source: MoF ,IMF estimates and projections. 140 TableA.4. CentralGovernmentFinancialOperations2000/01-2003/04 (% of GDP) 2000/01 2001102 2002/03 2003/04 Revenue 22.9 21.4 20.7 21.2 Income tax 6.6 6.4 6.9 6.8 Import duty (net) 3.4 2.3 1.8 1.9 Excise duty 3.4 3.5 3.5 3.7 Value-added tax 6.0 5.5 5.5 5.1 Investment income 0.3 0.1 0.1 0.4 Other 1.5 1.6 1.2 1.4 AIA 1.7 2.0 1.6 1.9 Expenditure and net lending 27.7 24.7 26.1 25.5 Recurrent expenditure 23.7 22.0 21.8 21.7 Interest payments 3.7 3.3 3.5 2.9 Domestic interest 2.8 2.6 2.7 2.3 Foreign interest due 0.9 0.7 0.8 0.6 Wages and benefits (civil service) 8.1 8.4 8.4 8.6 Civil service reform 0.7 0.2 0.0 0.0 Pensions, etc. 0.7 1.0 0.9 1.3 Other 6.2 6.6 6.8 6.9 Ufwhich: operations and maintenance 4.0 4.2 4.3 4.6 Of which;Transfers 1.8 1.8 1.8 1.9 Defense and NSIS (National Security Intelligence Service)a 2.0 2.4 2.3 2.1 Pending billsb 0.2 0.2 0.0 0.0 Drought relief expenditures 1.9 0.0 0.0 0.0 Development and net lending 4.0 2.7 4.3 3.8 Domestically financed 0.7 1.2 1.3 1.4 Foreign financed 1.8 1.7 2.2 2.3 Net lending 0.5 0.2 0.2 0.1 Pending billsb 0.4 -0.3 0.7 0.0 Drought Development Expenditure 0.7 0.0 0.0 0.0 CCF 0.0 0.0 0.0 0.0 Revenue/Expenditure Measures 0.0 0.0 0.0 0.0 Balance (commitment basis, excluding grants) -4.8 -3.4 -5.4 -4.4 Grants 2.9 0.7 1.5 1.9 Food relief grants 1.5 0.0 0.0 0.0 Project grants 0.7 0.6 1.4 1.5 Program grants 0.7 0.2 0.0 0.4 Balance (commitment basis, including grants) -1.9 -2.6 -4.0 -2.5 Adjustments to cash basis 0.8 -0.3 0.7 0.0 Balance (cash basis, including grants) -1.1 -2.9 -3.3 -2.5 Financing 1.1 2.9 3.2 2.5 Net foreign financing 1.5 -1.2 -1.0 -0.1 Project loans 1.7 1.1 0.7 0.8 Program loans 0.5 0.0 0.0 0.6 Financial defense lease loan adjustment 0.0 0.2 0.2 0.2 Repayments due -4.0 -2.8 -2.0 -1.7 Change inarrears 0.7 0.3 0.1 -0.5 Rescheduling 2.6 0.0 0.0 0.5 Privatization proceeds 0.0 0.1 0.0 0.0 Bank restructuring costs 0.0 0.0 0.0 -0.2 Expenditure arrears securitization -0.4 -0.3 -0.4 0.0 Net domestic borrowing 0.1 4.3 4.6 2.8 Ufwhich: excluding bank restructuring costs 0.1 4.3 4.6 2.6 and excluding expenditure arrears securitization 0.5 4.6 5.0 2.6 Financinggap (stat. discrepancy for outturns) 0.0 0.1 0.1 0.0 Memorandum items Total project support 3.8 1.7 2.2 2.3 Total identified and unidentified gross external support 5.0 1.8 2.2 3.3 Primarv budgetbalance 2.6 0.3 0.3 0.4 Stock of domestic debt, net (end o fperiod) 19.6 21.9 24.3 24.9 aDifferfrom authorities numbers because of different accounting treatment o f finance defense leases. bThefiscal accounts are on a cash basis (withthe exception o f foreign interest due). Adding accumulation of pending bills and subtracting cash repayment o f them adjusts to a commitment basis. Source: MoF, I M F estimates and projections. 141 TableA.5. Shares inTotalExpenditureandNetLending(%) 1994/95 1995/96 1996/97 1997198 1998/99 1999/2000 2000101 2001102 2002103 Recurrentexpenditure 78.9 81.2 84.3 83.0 81.8 86.4 85.4 89.1 83.6 Interest payments 23.3 24.9 20.7 19.6 20.3 16.2 13.3 13.3 13.5 Domestic interest 16.6 17.4 15.8 15.6 16.1 11.5 10.0 10.4 10.4 Foreign interest due 6.8 7.6 4.9 3.9 4.2 4.8 3.4 2.9 3.2 Civil service wages 31.4 30.7 28.7 32.0 32.0 36.4 29.3 34.1 32.0 Civil service reform 2.0 1.3 0.6 0.4 0.4 0.2 2.6 0.7 0.0 Pensions, etc. 1.7 2.1 2.8 2.4 2.6 2.8 2.6 3.9 3.6 Other 14.6 19.1 24.7 22.0 21.4 24.8 22.6 26.7 26.1 Ofwhich: O&M 15.6 14.4 17.1 16.6 Ofwhich: Transfers 7.6 6.4 7.4 6.9 Defenseand NSIS (National Security 5.4 6.1 6.2 5.2 5.6 6.8 7.2 9.6 8.6 IntelligenceService) Pending bills 0.4 -3.0 0.6 1.4 -0.6 -0.9 0.8 0.8 -0.2 Drought relief expenditures 0.0 0.0 0.0 0.0 0.0 0.0 6.9 0.0 0.0 Developmentandnet lending 21.1 18.8 15.7 17.0 18.2 13.6 14.6 10.9 16.4 Domestically financed 8.2 4.5 5.1 5.8 4.5 2.8 2.4 4.7 4.8 Foreign financed 10.3 11.4 9.4 9.6 8.3 10.1 6.6 6.7 8.3 Net lending 0.3 1.2 0.8 0.8 3.9 1.o 1.7 0.8 0.6 Pending bills 2.4 1.6 0.4 0.7 1.5 -0.4 1.4 -1.4 2.7 Drought development expenditure 0.0 0.0 0.0 0.0 0.0 0.0 2.5 0.0 0.0 Source: MoF 142 SI m o o 0 6 M N "! N > ~ 0 0 - 0 w m r - r -Wmd0F0ONd- 0 0 - o - o o o d Z N 5 I? m 3 Q I O O ~ d - m 0 -0"m o m ". e o - o o o 2 e o o m o m zmc-4 o w l m o m m d N ~ 0 0 0 0 0 3 m ~ N w e m N m 0 00m t- m w W W N * m e m m 1 0 0 0 0 o o m o m 2 - c i - 0 1 0Wo o m o o o o c o m oWo c c Table A.7. MinisterialRecurrentExpenditures(K Sh Million) 1999100 2000/01 2001/02 2002/03 2003/04 2004/05 Subvote & Ministry vote actual actual actual actual revised gross est. head RO1 Office o f the President 13,606 19,163 19,166 17,571 24,653 22,858 R02 State House 423 505 611 761 827 511 R03 Directorate o f Personnel Management (DPM) 670 6,493 2,152 1,409 1,072 2,533 R04 ForeignAffairs and International Co-operation 2,758 2,800 3,928 3,817 4,860 5,685 R05 Office o f the Vice-president (OW) and Home 7,756 7,465 (OW and Regional Development) 564 (Home Affairs) 3,199 3,571 4,255 4,534 R09 Regional Development 647 614 R06 Planning and NationalDevelopment 754 780 (Finance and Planning) 8,043 R07 Finance 4,602 6,533 6,864 8,414 15,071 R08 Defence 10,707 14,439 16,258 17,430 19,921 20,394 R10 Agriculture 4,886 5,795 4,762 3,663 3,434 3,593 R19 Livestock and Fisheries Developmenta 615 162 270 2,420 2,097 R11 Health 9,226 10,966 11,985 14,422 15,988 15,952 R12 Local Government 2,707 3,619 3,697 3,156 4,436 4,721 R13 Roads, Public Works and Housing 8,419 7,842 9,260 5,473 10,699 10,643 R14 Transport and Communications 882 1,686 1,657 1,863 1,933 2,628 R15 Labour and Human Resource Development 849 1,097 1,271 1,373 753 779 R46 Tourism and Information 1,314 1,800 1,839 729 1,185 1,034 R17 Justice and Constitutional Affairs 59 153 553 R18 Gender, Sports, Culture and Social Services 668 1,435 1,479 R20 Water Resources Management and 1,305 2,041 1,929 R21 Environment, Natural Resources and Wildlife 2,202 2,115 2,79 1 1,874 2,623 2,964 R22 Co-operative Development and Marketing 285 683 533 R16 Trade and Industry 1,839 1,756 2,136 1,942 R25 Attomey General 194 297 484 423 501 362 R26 Judiciary 365 1,013 1,141 957 1,352 1,338 R27 Public Service 43 70 117 118 181 137 R28 Controller and Auditor General 123 124 264 265 381 1,258 R29 NationalAssembly 1,306 1,922 3,296 3,700 4,672 5,500 R30 Energy 44 92 116 66 238 243 R31 Education, Science and Technology 45,800 48,103 52,589 60,938 70,927 78,707 R33 Electoral Commission 343 475 1,306 3,863 1,180 1,425 R36 Lands and Settlement 722 892 1,149 1,152 1,454 1,190 R45 NSIS 1,915 2,607 2,790 3,487 3,947 4,145 Grandtotal 117,305 144,634 155,749 165,724 203,656 221,061 a. Basedon Ministerial Public ExpenditureReview (MPER) estimates. Source: BMD. 146 Table A.8. MinisterialShares of RecurrentExpenditure (YO) 1999/00 2000/01 2001/02 2002103 2003104 2004/05 Subvote & Ministry vote actual actual actual actual revised gross est. head RO1 Office of the President 11.6 13.2 12.3 10.6 12.1 10.3 R02 State House 0.4 0.3 0.4 0.5 0.4 0.2 R03 DPM 0.6 4.5 1.4 0.9 0.5 1.1 R04 Foreign Affairs and IntemationalCo-operation 2.4 1.9 2.5 2.3 2.4 2.6 R05 O W and Home Affairs 3.8 3.4 (OW and RD) 0.3 (Home Affairs) 2.7 2.5 2.7 2.7 R09 Regional Development 0.3 0.3 R06 Planning and National Development 0.4 0.4 (Finance and Planning) 4.9 R07 Finance 3.9 4.5 4.4 4.1 6.8 R08 Defence 9.1 10.0 10.4 10.5 9.8 9.2 R10 Agriculture 4.2 4.0 3.1 2.2 1.7 1.6 R19 Livestock and Fisheries Developmenta 0.4 0.1 0.2 1.2 0.9 R11 Health 7.9 7.6 7.7 8.7 7.9 7.2 R12 Local Government 2.3 2.5 2.4 1.9 2.2 2.1 R13 Roads, Public Works and Housing 7.2 5.4 5.9 3.3 5.3 4.8 R14 Transport and Communications 0.8 1.2 1.1 1.1 0.9 1.2 R15 Labour and Human ResourceDevelopment 0.7 0.8 0.8 0.8 0.4 0.4 R46 Tourism and Information 1.1 1.2 1.2 0.4 0.6 0.5 R17 Justice and Constitutional Affairs 0.0 0.1 0.3 R18 Gender, Sports, Culture and Social Services 0.4 0.7 0.7 R20 Water Resources Management and Development 0.8 1.o 0.9 R21 Environment, Natural Resources and Wildlife 1.9 1.5 1.8 1.1 1.3 1.3 R22 Co-operative Development and Marketing 0.2 0.3 0.2 R16 Trade and Industry 1.2 1.1 1.o 0.9 R25 Attomey General 0.2 0.2 0.3 0.3 0.2 0.2 R26 Judiciary 0.3 0.7 0.7 0.6 0.1 0.6 R27 Public Service 0.0 0.0 0.1 0.1 0.1 0.1 R28 Controller and Auditor General 0.1 0.1 0.2 0.2 0.2 0.6 R29 National Assembly 1.1 I.3 2.1 2.2 2.3 2.5 R30 Energy 0.0 0.1 0.1 0.0 0.1 0.1 R31 Education, Science and Technology 39.0 33.3 33.8 36.8 34.8 35.6 R33 Electoral Commission 0.3 0.3 0.8 2.3 0.6 0.6 R36 Lands and Settlement 0.6 0.6 0.7 0.7 0.7 0.5 R45 NSIS 1.6 1.8 1.8 2.1 1.9 1.9 Grand total 100.0 100.0 100.0 100.0 100.0 100.0 Total as a percent of GDP 15.8 17.2 16.8 16.3 18.2 18.2 a. Based on MPER estimates. Source: BMD. 147 Table A.9. MinisterialDevelopmentExpenditure(KSh Million) 1999/00 2000101 2001/02 2002103 2003104 2004105 Subvote & Ministryvote actual actual actual actual revised gross est. head DO1 Office of the President 1,973 3,989 2,966 3,146 4,969 6,447 DO2 State House 19 35 36 104 273 250 DO3 DPM 186 121 203 170 655 862 DO4 Foreign Affairs and Intemational Co- 2 0 4 36 110 100 DO5 OVP and Home Affairs 586 955 (OW and RD) 0 (Home Affairs) 42 116 202 953 DO9 Regional Development 216 780 DO6 Planning and National Development 843 1,484 (Finance and Planning) 3,807 DO7 Finance 1,279 5,247 537 8,314 19,273 DO8 Defence 81 0 0 0 0 D10 Agriculture 300 856 991 1,384 3,126 4,301 D19 Livestock and Fisheries Developmenta 2,182 2,147 2,187 620 4,147 Dl1 Health 415 940 2,047 1,092 3,902 7,804 D12 Local Govemment 116 881 455 849 350 2,032 D13 Roads, Public Works andHousing 559 1,704 822 1,198 6,764 13,070 D14 Transport and Communications 292 438 532 101 579 1,901 D15 Labour and HumanResource 123 186 211 242 179 120 D46 Tourism and Information 112 68 71 753 919 749 D17 Justice and Constitutional Affairs 8 465 1,322 D18 Gender, Sports, Culture and Social 103 502 447 D20 Water ResourcesManagement and 650 4,169 5,325 D21 Environment, Natural Resources and 416 1,610 1,526 351 901 1,309 D22 Co-operative Developmentand Marketing 3 78 923 D16 Trade and Industry 117 161 370 677 D25 Attorney General 4 17 7 1 20 95 D26 Judiciary 1 10 13 10 164 287 D27 Public Services 0 0 0 0 0 0 D28 Controller and Auditor General 0 0 0 0 0 0 D29 National Assembly 0 0 0 0 0 0 D30 Energy 773 6,225 2,869 1,086 6,860 7,218 D31 Education, Science and Technology 392 533 998 2,439 8,484 4,77 1 D33 Electoral Commission 0 0 0 0 0 0 D36 Lands and Settlement 10 22 61 464 161 103 D45 NSIS 0 0 0 Grandtotal 7.095 25.180 16.815 19,111 54,580 86,752 a. Basedon MPER estimates. Source: BMD. 148 Table A.lO. MinisterialSharesof DevelopmentExpenditure(YO) 1999/00 2000/01 2001102 2002/03 2003/04 2004/05 Subvote & Ministry vote actual actual actual actual revised gross est. DO1 Office o f the President 27.8 15.8 17.6 16.5 9.1 7.4 DO2 State House 0.3 0.1 0.2 0.5 0.5 0.3 DO3 D P M 2.6 0.5 1.2 0.9 1.2 1.o DO4 ForeignAffairs and International Co-operation 0.0 0.0 0.0 0.2 0.2 0.1 DO5 O W and Home Affairs 1.1 1.1 (OW and RD) 0.0 (Home Affairs) 0.6 0.5 1.2 5.0 DO9 Regional Development 0.4 0.9 DO6 Planning and National Development 1.5 1.7 (Finance and Planning) 19.9 DO7 Finance 18.0 20.8 3.2 15.2 22.2 DO8 Defence 1.1 0.0 0.0 0.0 0.0 0.0 D10 Agriculture 4.2 3.4 5.9 7.2 5.7 5.0 D19 Livestock and Fisheries Development" 8.7 12.8 11.4 1.1 4.8 D11 Health 5.8 3.7 12.2 5.7 7.1 9.0 D12 Local Government 1.6 3.5 2.7 4.4 0.6 2.3 D13 Roads, Public Works and Housing 7.9 6.8 4.9 6.3 12.4 15.1 D14 Transport and Communications 4.1 1.7 3.2 0.5 1.1 2.2 D15 Labour and Human Resource Development 1.7 0.7 1.3 1.3 0.3 0.1 D46 Tourism and Information 1.6 0.3 0.4 3.9 1.7 0.9 D17 Justice and ConstitutionalAffairs 0.0 0.9 1.5 D18 Gender, Sports, Culture and Social Services 0.5 0.9 0.5 D20 Water Resources Management and 3.4 7.6 6.1 D21 Environment, Natural Resources and Wildlife 5.9 6.4 9.1 1.8 1.7 1.5 D22 Co-operative Development and Marketing 0.0 0.1 1.1 D16 Trade and Industry 0.7 0.8 0.7 0.8 D25 Attorney General 0.1 0.1 0.0 0.0 0.0 0.1 D26 Judiciary 0.0 0.0 0.1 0.1 0.3 0.3 D27 Public Services 0.0 0.0 0.0 0.0 0.0 0.0 D28 Controller and Auditor General 0.0 0.0 0.0 0.0 0.0 0.0 D29 National Assembly 0.0 0.0 0.0 0.0 0.0 0.0 D30 Energy 10.9 24.7 17.1 5.7 12.6 8.3 D31 Education, Science and Technology 5.5 2.1 5.9 12.8 15.5 5.5 D33 Electoral Commission 0.0 0.0 0.0 0.0 0.0 0.0 D36 Lands and Settlement 0.1 0.1 0.4 2.4 0.3 0.1 D45 NSIS 0.0 0.0 0.0 0.0 0.0 0.0 Grand total 100.0 100.0 100.0 100.0 100.0 100.0 Total as a percent of GDP 1.0 3.0 1.8 1.9 4.9 7.1 a. Based on MPER estimates. Source: BMD. 149 Table A.11. TotalMinisterialExpenditures(KSh Million) 1999100 2000/01 2001102 2002103 2003/04 2004105 Subvote & Ministry vote actual actual actual actual revised gross est. head 01 Office o f the President 15,579 23,152 22,132 20,717 29,622 29,305 02 State House 442 540 647 865 1,100 761 03 D P M 856 6,614 2,355 1,579 1,727 3,395 04 Foreign Affairs and Intemational Co- 2,760 2,800 3,932 3,853 4,970 5,785 05 O W and Home Affairs 8,342 8,420 (OW and RD) 564 (Home Affairs) 3,241 3,687 4,457 5,487 09 Regional Development 863 1,394 06 Planning and NationalDevelopment 1,598 2,264 (Finance and Planning) 11,850 07 Finance 5,881 11,780 7,40 1 16,728 34,344 08 Defence 10,788 14,439 16,258 17,430 19,921 20,394 10 Agriculture 5,186 6,65 1 5,753 5,047 6,560 7,894 19 Livestock and Fisheries Developmenta 2,797 2,309 2,457 2,411 6,244 11 Health 9,641 11,906 14,032 15,514 19,890 23,756 12 Local Government 2,823 4,500 4,152 4,005 4,786 6,753 13 Roads, Public Works and Housing 8,978 9,546 10,082 6,671 17,463 23,713 14 Transport and Communications 1,174 2,124 2,189 1,964 2,5 12 4,529 15 Labour and Human Resource Development 972 1,283 1,482 1,615 932 899 46 Tourism and Information 1,426 1,868 1,910 1,482 2,104 1,783 17 Justice and Constitutional Affairs 67 618 1,875 18 Gender, Sports, Culture and Social 771 1,937 1,926 20 Water Resources Management & 1,955 6,210 7,254 21 Environment, Natural Resources and 2,618 3,725 4,317 2,225 3,524 4,273 22 Co-operative Development and Marketing 288 761 1,456 16 Trade and Industry 1,956 1,917 2,506 2,6 19 25 Attorney General 198 3 14 491 424 521 457 26 Judiciary 366 1,023 1,154 967 317 1,625 27 Public Service 43 70 117 118 181 137 28 Controller and Auditor General 123 124 264 265 381 1,258 29 National Assembly 1,306 1,922 3,296 3,700 4,672 5,500 30 Energy 817 6,317 2,985 1,152 7,098 7,46 1 31 Education, Science and Technology 46,192 48,636 53,587 63,377 79,411 83,478 33 Electoral Commission 343 475 1,306 3,863 1,180 1,425 36 Lands and Settlement 732 914 1,210 1,616 1,615 1,293 45 NSIS 1,915 2,607 2,790 3,487 3,947 4,145 Grand total 124,400 169,814 172,564 184,835 258,236 307,813 a. Based on MPER estimates. Source: BMD. 150 Table A.12. MinisterialShares of Total ExDenditure (YO) 1999100 2000101 2001102 2002103 2003104 2004105 Subvote & Ministry vote actual actual actual actual revised gross est. head 01 12.5 13.6 12.8 11.2 11.5 9.5 ~~Office o f the President 02 State House 0.4 0.3 0.4 0.5 0.4 0.2 03 D P M 0.7 3.9 1.4 0.9 0.7 1.1 04 Foreign Affairs and International Co- 2.2 1.6 2.3 2.1 1.9 1.9 05 O W andHome Affairs 3.2 2.7 ( O W and RD) 0.3 (Home Affairs) 2.6 2.2 2.6 3.0 09 Regional Development 0.3 0.5 06 Planning andNationalDevelopment 0.6 0.7 (Finance and Planning) 6.4 07 Finance 4.7 6.9 4.3 6.5 11.2 08 Defence 8.7 8.5 9.4 9.4 7.7 6.6 10 Agriculture 4.2 3.9 3.3 2.7 2.5 2.6 19 Livestock and Fisheries Developmenta 0.9 2.0 11 Health 7.8 7.0 8.1 8.4 7.7 7.7 12 Local Government 2.3 2.6 2.4 2.2 1.9 2.2 13 Roads, Public Works and Housing 7.2 5.6 5.8 3.6 6.8 7.7 14 Transport and Communications 0.9 1.3 1.3 1.1 1.o 1.5 15 Labour and Human Resource 0.8 0.8 0.9 0.9 0.4 0.3 46 Tourism and Information 1.1 1.1 1.1 0.8 0.8 0.6 17 Justice and Constitutional Affairs 0.0 0.2 0.6 18 Gender, Sports, Culture and Social 0.4 0.8 0.6 20 Water Resources Management and 1.1 2.4 2.4 21 Environment, Natural Resources and 2.1 2.2 2.5 1.2 1.4 1.4 22 Co-operative Development and Marketing 0.2 0.3 0.5 16 Trade and Industry 1.1 1.o 1.o 0.9 25 Attorney General 0.2 0.2 0.3 0.2 0.2 0.1 26 Judiciary 0.3 0.6 0.7 0.5 0.1 0.5 27 Public Service 0.0 0.0 0.1 0.1 0.1 0.0 28 Controller and Auditor General 0.1 0.1 0.2 0.1 0.1 0.4 29 National Assembly 1.o 1.1 1.9 2.0 1.8 1.8 30 Energy 0.7 3.7 1.7 0.6 2.7 2.4 31 Education, Science and Technology 37.1 28.6 31.1 34.3 30.8 27.1 33 Electoral Commission 0.3 0.3 0.8 2.1 0.5 0.5 36 Lands and Settlement 0.6 0.5 0.7 0.9 0.6 0.4 45 NSIS 1.5 1.5 1.6 1.9 1.5 1.3 Grand total 100.0 100.0 100.0 100.0 100.0 100.0 Total as a percent of GDP 16.7 20.2 18.6 18.2 23.1 25.3 a. Basedon MPERestimates. Source: BMD. 151 TableA.13. TotalMinisterialExpenditures(% of GDP) 1999100 2000101 2001102 2002103 2003104 2004105 Subvote &Ministry vote actual actual actual actual revised gross est. 01 Office o f the President 2.1 2.8 2.4 2.0 2.6 2.4 02 State House 0.1 0.1 0.1 0.1 0.1 0.1 03 D P M 0.1 0.8 0.3 0.2 0.2 0.3 04 Foreign Affairs and International Co- 0.4 0.3 0.4 0.4 0.4 0.5 05 OVP and Home Affairs 0.7 0.7 (OVP and RD) 0.1 0.0 0.0 (Home Affairs) 0.4 0.4 0.5 0.5 0.0 0.0 09 Regional Development 0.1 0.1 06 Planning and National Development 0.1 0.2 (Finance and Planning) 1.2 0.0 0.0 07 Finance 0.8 1.4 0.8 1.5 2.8 08 Defence 1.5 1.7 1.8 1.7 1.8 1.7 10 Agriculture 0.7 0.8 0.6 0.5 0.6 0.6 19 Livestock and Fisheries Development 0.3 0.2 0.2 0.2 0.5 11 Health 1.3 1.4 1.5 1.5 1.8 2.0 12 Local Government 0.4 0.5 0.4 0.4 0.4 0.6 13 Roads, Public Works and Housing 1.2 1.1 1.1 0.7 1.6 1.9 14 Transport and Communications 0.2 0.3 0.2 0.2 0.2 0.4 15 Labour and HumanResource 0.1 0.2 0.2 0.2 0.1 0.1 46 Tourism and Information 0.2 0.2 0.2 0.1 0.2 0.1 17 Justice & Constitutional Affairs 0.0 0.1 0.2 18 Gender, Sports, Culture and Social 0.1 0.2 0.2 20 Water Resources Management and 0.2 0.6 0.6 21 Environment, Natural Resources and 0.4 0.4 0.5 0.2 0.3 0.4 22 Co-operative Development and 0.0 0.1 0.1 16 Trade and Industry 0.2 0.2 0.2 0.2 25 Attorney General 0.0 0.0 0.1 0.0 0.0 0.0 26 Judiciary 0.0 0.1 0.1 0.1 0.0 0.1 27 Public Service 0.0 0.0 0.0 0.0 0.0 0.0 28 Controller and Auditor General 0.0 0.0 0.0 0.0 0.0 0.1 29 National Assembly 0.2 0.2 0.4 0.4 0.4 0.5 30 Energy 0.1 0.8 0.3 0.1 0.6 0.6 31 Education, Science and Technology 6.2 5.8 5.8 6.2 7.1 6.9 33 Electoral Commission 0.0 0.1 0.1 0.4 0.1 0.1 36 Lands and Settlement 0.1 0.1 0.1 0.2 0.1 0.1 45 NSIS 0.3 0.3 0.3 0.3 0.4 0.3 Grandtotal 16.7 20.2 18.6 18.2 23.1 25.3 a. Basedon MPER estimates. Source: BMD. 152 TableA.14. Actual DevelopmentSpendingbyBroadFunctions 2001102 2002103 actual actual Office o f the President 2,966 3,146 Finance 537 3,807 Home Affairs 202 953 Local Government 455 849 Other administrative 263 329 Education 998 2,439 Health 2,047 1,092 Public Works 822 1,198 Energy 2,869 1,086 Water Resources 532 101 Agriculture 991 1,384 Environment andNatural Resources 1,526 351 Other 460 1,726 Total 14,668 18,461 Table A.15. Changes inthe DevelopmentBudget2002/03-2003/04 (Printed Estimates,KSh CurrentMillion) ~ 2002103 20003104 budget budget Change Office o f the President 9,644 5,900 (3,744) Finance 4,779 7,837 3,058 Home Affairs 672 442 (230) Local Government 2,552 1,080 (1,472) Other Administrative 802 2,137 1,335 Education 3,218 8,434 5,216 Health 4,662 5,116 454 Public Works 5,235 8,663 3,428 Energy 4,992 6,578 1,586 Water Resources 450 579 129 Agriculture 3,506 3,882 376 Environment andNatural Resources 3,592 980 (2,612) Other 2,365 3,290 925 Total 46,469 54,918 8,449 153 EE 4- 0 TableA.17. Budgetversus Outturnby Key Ministries,2002/03, andComparison with 2003/04 Budget Recurrent budget 2002/03 2003104 Actual % % increase on Budget Actual of Budget Budget 2002103 Actual Office o f the President 20,333 17,751 87 20,536 16 Finance and Planning 7,576 8,043 106 15,319 90 Foreign Affairs 3,827 3,817 100 4,405 15 Home Affairs 4,473 4,534 101 5,992 32 Local Government 3,804 3,156 83 4,467 42 National Assembly 3,359 3,700 110 4,672 26 Electoral Commission 4,661 3,863 83 1,213 -69 Other administrative 3,151 2,553 81 2,689 5 Judiciary and related 1,670 1,439 86 1,957 36 Defence and NSIS 17,338 20,917 121 21,755 4 Education 54,709 60,938 111 71,943 18 Health 13,652 14,422 106 16,005 11 Public Works 10,597 5,473 52 10,865 99 Transport 1,719 1,863 108 1,662 -1 1 Water n.a. 1,305 n.a. 2,117 62 Trade and Industry 2,299 1,756 76 2,196 25 Agricu1ture 6,976 3,663 53 6,163 68 Environment & Natural Resources 3,516 1,874 53 2,440 30 Other rural-related 1,282 1,437 112 2,246 56 Other 2,307 3,220 140 5,219 62 Total 167,249 165,724 99 203,861 23 ma. not available. 155 Table A.18. Forward-LookingMTEFCeilings,2003/04-2006/07 (K Sh Million)' 2003104 2004105 2005106 2006107 revised Recurrent Public Administration 12,412 12,246 13,340 14,007 Public Safety, Law & Order 33,748 33,136 34,793 36,532 Human Resource Development 80,109 82,639 88,424 94,614 Tourism, Trade & Industry 3,090 3,061 2,977 3,126 Agriculture & Rural Development 10,556 10,265 13,120 13,776 Physical Infrastructure 19,294 18,909 20,066 21,069 Information Technology 700 720 742 764 National Security 23,901 22,000 22,660 23,340 Subtotal 183,810 182,976 196,122 207,228 Other recurrent expendituresb 19,797 30,434 25,772 26,4 11 Total recurrent 203,607 213,410 221,894 233,639 Development Public Administration 2,554 1,824 1,915 2,011 Public Safety, L a w & Order 6,417 6,713 7,048 7,401 HumanResourceDevelopment 13,068 11,920 12,516 13,142 Tourism, Trade & Industry 1,234 836 878 922 Agriculture & Rural Development 4,887 5,270 5,534 5,811 Physical Infrastructure 18,723 25,718 27,004 28,355 Information Technology 79 61 64 68 National Security 0 0 0 0 Government transfers-CDF (Constituency Development Funds) 1,224 4,500 4,500 4,500 Total development 48,186 56,842 59,459 62,210 a.Includes AIA and wages. b.Other recurrent expenditures include amounts taken out off the top by the Treasury. Examples include wage adjustments, the National Assembly, Controller and Auditor General, the Kenya Revenue Authority (KRA), elections, grains and famine relief, civil service reforms, and so on. 156 Table A.19. CivilService Establishment,ExcludingTeachers, 2002 Job group Average Of which Security & medical Mean pay Inpost Medical Security Other % K Sh A 41 455 455 0.0 6,077 B 43 5,620 5,620 0.0 6,726 C 46 8,269 3 5 8,261 0.1 6,926 D 43 10,365 67 35 10,263 1.0 7,372 E 42 14,138 81 51 14,006 0.9 8,272 F 35 62,360 34 54,461 7,865 87.4 9,736 G 42 26,224 2,602 5,763 17,859 31.9 12,133 H 41 18,723 8,135 2,415 8,173 56.3 14,753 J 42 25,537 11,187 2,365 11,985 53.1 17,119 K 43 8,829 3,602 378 4,849 45.1 22,981 L 42 6,502 608 176 5,718 12.1 33,661 M 44 2,237 247 55 1,935 13.5 39,858 N 45 1,085 236 41 808 25.5 53,473 P 48 401 112 12 277 30.9 73,790 Q 50 184 5 1 178 3.3 66,537 R 52 140 3 137 2.1 115,841 s 52 58 1 5 52 10.3 453,784 T 52 16 1 15 6.3 296,656 U 52 16 16 0.0 379,667 V 60 935 910 25 97.3 24,115 W 48 10 5 5 50.0 14,069 x 50 904 865 39 95.7 22,886 Y 46 9 9 0.0 22,504 Total 193,017 26,923 67,544 98,550 48.9 13,665 A-G share (YO) 66.0 10.4 89.3 65.3 14,063 46.1 A-J Share (%) 89.0 82.1 96.4 85.7 45,882 73.2 K-Y share (%) 11.0 17.9 3.6 14.3 38,605 26.8 157 TableA.20. FiscalOutcomes,2000/01-2002/03 (KShMillion) 2000/01 2001/02 2002103 3Yr Target Actual Deviation Target Actual Deviation Target Actual Deviation Ave. deviation% A. Revenue (1 +2 +3) 211,431 216,393 4,962 220,635 203,436 -17,199 235,527 230,134 (5,393) -3 1. Ordinary revenue 174,666 175,481 815 186,457 175,746 -10,711 192,382 190,591 -2 o fwhich Import duty (net) 29,633 28,726 -907 23,151 21,584 -1,567 21,118 18,477 -7 Exciseduty 28,738 28,318 -420 36,197 32,077 -4,120 38,082 35,643 -7 Income tax 53,157 53,429 272 55,297 55,862 565 61,198 66,744 4 Value added tax 50,503 50,298 -205 55,445 50,871 -4,574 57,449 56,135 -4 Investment income 982 2,478 1,496 2,346 767 -1,579 1,762 1,243 -12 Other 11,653 12,232 579 14,021 14,585 564 12,773 12,349 2 2. AIA 18,413 16,832 -1,581 23,810 20,867 -2,943 24,600 24,601 -7 3. Grants 18,352 24,080 5,728 10,368 6,823 -3,545 18,545 14,942 -3 o/w Cash 9,204 7,476 -1,728 2,750 2,563 -187 4,373 4,262 -12 AIA 9,148 16,604 7,456 7,618 4,260 -3,358 14,172 10,680 2 B. Expenditure and net lending 240,214 232,921 (7,293) 249,531 225,760 (23,771) 278,004 264,144 -5 1. Recurrent expenditure 203,019 198,941 -4,078 205,572 200,807 (4,765) 230,396 220,618 -3 O/w Domesticinterest 21,533 23,232 1,699 23,456 23,744 288 26,767 27,567 4 Foreign interest due 9,060 7,803 -1,257 7,899 6,640 -1,259 8,459 8,459 -10 Pensions and so forth 6,328 6,136 -192 9,768 8,995 -773 12,921 9,450 -15 Wages and salaries 69,861 68,119 -1,742 77,674 77,638 -36 85,744 85,087 -1 Civil service reform 7,979 6,095 -1,884 1,770 1,665 -105 1,600 957 -23 Faminerelief - - - - - 800 0 -100 Strategic grain reserve 1,500 1,400 -100 1,500 1,260 -240 1,000 500 -2 1 General electioniConstitutiona1 rev. - - - 1,100 1,100 - 4,000 3,402 -12 KRA 2,760 2,560 -200 2,693 2,693 - 2,791 2,329 -8 O&M, other 83,998 83,596 -402 79,712 77,072 -2,640 86,314 82,867 -3 2. Developmentand net lending 37,195 33,980 -3,215 43,959 24,953 -19,006 47,608 43,526 -26 C. Deficit excl. grants (commitment (47,135) (40,608) 6,527 (39,264) (29,147) 10,117 (61,022) (48,952) -17 basis) D.Deficit incl. grants(commitment (28,783) (16,528) 12,255 (28,896) (22,324) 6,572 (42,477) (34,010) -23 basis) E. Adjustmentto cashbasis -3,913 3,423 7,336 292 -5,085 -5,377 (5,688) (5,687) 40 F.Deficit incl.grants(cash basis) (32,696) (13,105) 19,591 (28,604) (27,409) 1,195 (48,165) (39,697) -18 G. Financing 32,696 13,105 (19,591) 28,604 27,409 (1,195) 48,165 34,779 -31 1. Net foreign financing: 26,633 12,489 -14,144 -5,179 -13,314 -8,135 (6,790) (12,143) -188 o/w Project loans, cash 9,338 8,719 -619 4,338 2,898 -1,440 4,473 2,191 -24 Project loans, AIA 8,045 5,323 -2,722 13,577 7,133 -6,444 10,687 5,276 -45 Program loans 14,387 4,045 -10,342 - - - 0 0 -72 2. b) Privatizationproceeds 7,572 - -7,572 955 955 0 0 -89 3. c) Net domestic financing: -1,509 616 2,125 32,828 39,768 6,940 54,955 46,922 1 158 Table A.21. Budget Outturn, 2000/01-2002/03: Deviation of Actual from Targets (%) 2000/0 1 2001/02 2002i03 2000i01- 2000101-2002iO3 three-year 2002i03 nominal change (%) actualitarget actualitarget actualitarget actualitarget actual target deviation (x) (%) (%) 2.3 -7.8 -2.3 -2.6 6.4 11.4 (5.0) 1. Ordinary revenue 0.5 -5.7 -0.9 -2.1 8.6 10.1 (1.5) o/w Import duty (net) -3.1 -6.8 -12.5 -7.4 -35.7 -28.7 (6.9) Excise duty -1.5 -11.4 -6.4 -6.4 25.9 32.5 (6.6) Income tax 0.5 1.0 9.1 3.5 24.9 15.1 9.8 Value added tax -0.4 -8.2 -2.3 -3.6 11.6 13.8 (2.1) Investment income 152.3 -67.3 -29.5 18.5 -49.8 79.4 (129.3) Other 5.0 4.0 -3.3 1.9 1.0 9.6 (8.7) 2. AIA -8.6 -12.4 0.0 -1.0 46.2 33.6 12.6 3. Grants 31.2 -34.2 -19.4 -7.5 -37.9 1.1 (39.0) oiw Cash -18.8 -6.8 -2.5 -9.4 -43.0 -52.5 9.5 AIA 81.5 -44.1 -24.6 4.3 -35.7 54.9 (90.6) B. Expenditureand net lending -3.0 -9.5 -5.0 -5.8 13.4 15.7 (2.3) 1. Recurrent expenditure -2.0 -2.3 -4.2 -2.9 10.9 13.5 (2.6) oiw Domestic interest 7.9 1.2 3.0 4.0 18.1 24.3 (5.6) Foreign interest due -13.9 -15.9 0.0 -9.9 8.4 -6.6 15.0 Pensions and so on -3.0 -7.9 -26.9 -12.6 54.0 104.2 (50.2) Wages and salaries -2.5 0.0 -0.8 -1.1 24.9 22.7 2.2 Civil service reform -23.6 -5.9 -40.2 -23.2 -84.3 -79.9 (4.4) Famine relief -100.0 -33.3 0.0 Strategic grain reserve -6.7 -16.0 -50.0 -24.2 -64.3 -33.3 (3 1.O) General 0.0 -15.0 -5.0 0.0 electionKonstitutiona1 rev. KRA -7.2 0.0 -16.6 -7.9 -9.0 1.1 (10.1) O&M other -0.5 -3.3 -4.0 -2.6 -0.9 2.8 (3.6) 2. Development and net lending -8.6 -43.2 -8.6 -20.2 28.1 28.0 0.1 C.Deficit excl. grants -13.8 -25.8 -19.8 -19.8 20.5 29.5 (8.9) (commitmentbasis) D. Deficitincl. grants -42.6 -22.7 -19.9 -28.4 105.8 47.6 58.2 (commitmentbasis) E.Adjustment to cash basis -1 87.5 -1,841.4 0.0 -676.3 -266.1 45.4 (311.5) F.Deficitincl. grants(cash -59.9 -4.2 -17.6 -27.2 202.9 47.3 155.6 basis) G.Financing -59.9 -4.2 -27.8 -30.6 165.4 47.3 118.1 1. Net foreign financing: -53.1 157.1 78.8 60.9 -197.2 -125.5 (71.7) oiw Project loans, cash -6.6 -33.2 -5 1.O -30.3 -74.9 -52.1 (22.8) Project loans, AIA -33.8 -47.5 -50.6 -44.0 -0.9 32.8 (33.7) Program loans -71.9 -24.0 -100.0 -100.0 0.0 2. Privatization proceeds -100.0 0.0 -33.3 -100.0 100.0 3. Net domestic financing: -140.8 21.1 -14.6 -44.8 7,517.2 -3741.8 11,259.0 Source: BMD Quarterly Budget Review. 159 BIBLIOGRAPHY Byaruhanga, C. V. 2004. "Review of the Medium Term Expenditure Framework: Draft Final Report." Ministryof Planning and National Development, Nairobi. Easterly, William. 1998. "When I s Fiscal Adjustment an Illusion?" World Bank, Washington, D.C. Government of Kenya. 1997. "Public Expenditure Review, Report of the Public Expenditure Secretariat." Nairobi. . 2000/01-2002/03 (various issues). "Quarterly Budget Review." Ministry of Finance, Nairobi. . 2002. "Public Expenditure Management: Developments since the 1997 Public Expenditure Report." Study conducted by Researchon Economic Policy Implementation and Management (REPIM), Northumberland, UnitedKingdom. . 2003a. Economic Recovery Strategy for Wealth and Employment Creation 2003-2007, Nairobi. . 2003b. "Public Expenditure Review 2003." Ministry of Planning and National Development, Nairobi. . 2004. "Investment Programme for the Economic Recovery Strategy for Wealth and Employment Creation2003-2007." Nairobi. International Monetary Fund (IMF). 2001. The Government Finance Statistics Manual. Washington, D.C. Kenya Institute for Public Policy Research and Analysis (KIPPRA). 2003. "Budget Mechanisms and Public Expenditure Tracking in Kenya." Study funded by the Deutsche Gesellschafi fir Technische Zusammenarbeit (German Agency for Technical Cooperation [GTZ]), Nairobi. Kithinji,K.,and R. Mukandala. 2003. "Tactics, Sequencing and Politics of Public Service Pay Policies in Developing and Middle Income Countries: Lessons from Sub-Saharan Afixa." Study sponsored by the World Bank and U.K. Department for International Development (DFID). Polackova, H. 1998. "Government Contingent Liabilities: A Hidden Risk to Fiscal Stability." World Bank, Washington, D.C. Republic ofKenya, Central Bureauof Statistics. 2004. "Economic Survey 2004." Nairobi. Republic of Kenya, Ministry o f Finance. 2003. "MTEF Budget Preparation Guidelines for the Period2004/05-2006/07." Treasury Circular 26/2003. Nairobi. Valentine, Theodore R., and John R. Wheeler. 2003. "Consultancy on Wage Bill Management and Civil Service Performance Enhancement for the Government o f Kenya." 160 WorldBank.2002. "AnAssessment of Local Service DeliveryandLocal GovernmentsinKenya, DraftFinalReport." Study commissionedbythe WorldBankandDFID, Washington, D.C. . 2003. "Kenya: A Policy Agenda Restore Growth." Report 25840-KE. Washington, to D.C. .2004a. "Strengthening the Foundationof Education and Training inKenya." Report 28064-KE. Washington, D.C. .2004b. "Project Appraisal Document for Northern Corridor Transport Improvement Project." Report 28826-KE. Washington, D.C WorldBankand IMF.2003. "Public ExpenditureManagementCountry Assessment andAction Plan: Kenya."Washington, D.C. 161