Report No. 30331-MG Madagascar Public Expenditure Review 2004 The Challenge of Poverty Reduction February 25, 2005 Public Sector Reform and Capacity Building Unit (AFTPR) Africa Region Document of the World Bank HIPC Highly IndebtedPoor Countries IDA IntemationalDevelopmentAssociation IFMIS Integratedfinancial management system IGF GeneralInspectoratefor Finances IMF IntemationalMonetaryFund JSA Joint StaffAssessment LFI Initial budgetlaw (loi de finances initiale) LFR Revisedbudgetlaw (loi de finances rectificative) MDG MillenniumDevelopmentGoal MECIE - Mise en ComptabilitC des Investissementsa 1'Environnement MEFB Ministry of Economy, Financeand Budget MENRS Ministry ofNationalEducationand ScientificResearch MENSUP -- Ministry of HigherEducation METFP - Ministry of Vocational andTechnical EducationandTraining M&E Monitoring and evaluation MINENVEF- Ministry of Environmentand Forests MINESEB Ministry of Basic and Secondary Education MOH Ministry of Health MTEF MediumTerm Expenditures Framework NEAP NationalEnvironmentAction Plan NGO Non-GovemmentOrganization NNP NationalNutrition Policy NPV Net Present Value ONE NationalOffice of the Environment OSF Observatoiredu Secteur Forestier PADR RuralDevelopment Action Plan PER Public ExpenditureReview PGDI Governanceand Institutional Development Project PIP Public InvestmentProgram PRDEAP Part des Recettes pour les Droits d'EntrCe dansles Aires ProtCgCes PRGF PovertyReductionand GrowthFacility PRS PovertyReductionStrategy PRSC PovertyReductionSupport Credit PRSP PovertyReductionSupportPaper PSIA Povertyand Social ImpactAnalysis RA RoadAuthority RMF RoadMaintenanceFund RNCFM RCseauNational des Cheminsde Fer Malagasy SMTM SociCt6 Malgachede Transport Maritime SOLIMA SolitanyMalagasy SSA Sub-Saharan Africa STA TechnicalSecretariat for Adjustment TSA Treasury Single Account UNDP UnitedNationsDevelopmentOrganization VAT Value AddedTax VPM Vice Prime Minister WB World Bank WBI World BankInstitute ZAP Administrativeand PedagogicalZones Vice President - GobindNankani Country Director ----James Bond Sector Director Paula Donovan Sector Manager Helga Mueller Task Team Leaders Guenter HeidenhofiChristos Kostopoulos The PER Team included Guenter Heidenhof and Christos Kostopoulos (Task Team Leader), Odile Keller, Emile Louis Rene Finateu, D. Randriamanampisoa, Sylke von Thadden, Gervais Rakotoarimanana, Sylvain Rambeloson (Public expenditure management); Jacob H. Bregman, Maria Eugenia Bonilla-Chacin, Patrick Ramanantoanina (Education); Willem van Eeghen, Janet Dooley, Anna van der Wouden (Country Management Unit); Susanne Holste, Noroarisoa Rabefaniraka (Transport); Kirk Hamilton, Martien Van Nieuwkoop, Bienvenu Rajaonson, Jean-Christophe Carret (Environment); Robert Keyftz, Luc Razafimandimby, Adam Schwartzman, Aurelien Kruse, Benu Bidani (Macroeconomic); Laza Razafiarison (Macro/Public Finance); Madeleine Chungkong, Mavo Ranaivoarivelo, Lanto Ramanankasina(overall support); Sanjay Pradhan, Brendan Horton and Stefano Patemostro are the peer reviewers. MADAGASCARPUBLICEXPENDITURE REVIEW TABLE OF CONTENTS EXECUTIVESUMMARY I CREATINGANENABLINGMACROECONOMICANDFISCALFRAMEWORK . ......................................................................................................................... .................1 i 1) MACROECONOMIC PERFORMANCEAND THE POLICY ENVIRONMENT .............................................. 1 2) POVERTYOUTCOMESAND CHALLENGES ......................................................................................... 4 3) 5 MADAGASCAR'S REDUCTIONSTRATEGY ........................................................................ THEROLE AND FINANCINGOFTHE STATE ....................................................................................... 4) FINANCIALSECTOR ........................................................................................................................ POVERTY 7 5) 12 I1 THE CHALLENGE OFRESOURCEALLOCATIONAND BUDGET EXECUTION .6) RECOMMENDATIONS ...................................................................................................................... .............14 13 1) ALLOCATING PUBLICRESOURCESINLINEWITH GOVERNMENT PRIORITIES ................................ 15 (A) (b) Comparative context.................................................................................................................. Past Allocations: Issues and Challenges .................................................................................... 15 22 (c) 23 (d) Coordination o f and ParticipationinThe Budget PreparationProcess...................................... Introducing a medium-term perspective to linkpolicy priorities with the budget.................... 25 (e) 30 (f) Revenue and Extemal Assistance Forecasting........................................................................... Planning for Recurrent and InvestmentExpenditures ............................................................... 33 (g) Comprehensiveness and Transparency ofthe Budget ............................................................... 34 2) ENSURINGBUDGET IMPLEMENTATIONINLINEWITH GOVERNMENT PRIORITIES ......................... 35 (a) (b) Credibility o f the Budget - sticking to BudgetedPriorities ....................................................... Budget Execution remains a Challenge ..................................................................................... 35 36 (c) (d) Expenditure Managementprocess............................................................................................. 39 Cash management And Commitment o f Expenditures.............................................................. 38 (e) Recommendations..................................................................................................................... -40 (0 ImprovingInternalandExternalControls................................................................................. 40 (g) 41 (h) ImprovingFinancialReporting.................................................................................................. Strengthening Public Procurement ............................................................................................. 42 (i)TransformingtheBudgetintoaManagementInstrument......................................................... 43 3) 46 I11 USE OFRESOURCESINTHE EDUCATIONSECTOR . EFFICIENTUSEOF HUMANRESOURCES.......................................................................................... ............................................................. 52 1) KEYPOLICY ISSUES ........................................................................................................................ 52 2) TRENDS EDUCATIONEXPENDITURES IN ......................................................................................... 54 3) BUDGET MANAGEMENT ..................................................................................................... ISSUES 56 4) 65 COMMUNITY PARTICIPATION ......................................................................................................... SECTORORGANIZATION................................................................................................................. 5) 68 6) RECOMMENDATIONS I V USEOF RESOURCESINTHE TRANSPORTSECTOR . ...................................................................................................................... 68 ............................................................. 71 1) KEYPOLICY ISSUES........................................................................................................................ 71 2) IMPROVING SECTORORGANIZATIONAND GOVERNANCE ............................................................... 73 3) CRITICAL PUBLIC EXPENDITUREISSUES ......................................................................................... 76 4) PRINCIPLEISSUESOF SUB-SECTORS .............................................................................................. 80 (a) (b) Air Transport.............................................................................................................................. 82 Roads ......................................................................................................................................... 80 (c) Rail Transport ............................................................................................................................ 83 (d) Maritime Transport and Port Management................................................................................ 84 84 V USE OF RESOURCESINTHE ENVIRONMENTSECTOR .5) RECOMMENDATIONS ...................................................................................................................... ....................................................... 86 1) KEYPOLICY ISSUES ........................................................................................................................ 86 2) SECTOR STRATEGIES ...................................................................................................................... 87 3) POLICY AND INST~TUT~ONAL FRAMEWORK ................................................................................... 90 (a) The Legislative Framework...................................................................................................... -90 (b) The Institutional Framework...................................................................................................... 90 4) CR1T1CALPUBLIC EXPENDITURE ISSUES ........................................................................................ 91 5) EFFECTIVENESS ENVIRONMENTALINSTITUTIONS.................................................................... 98 OF 6) ANUNFUNDEDMANDATE: ENVIRONMENTALIMPACTASSESSMENTS OF PUBLIC INVESTMENTS ................................................................................................................. 100 7) REVENUE SHARING WITH LOCAL COMM~ITIES......................................................................... 100 8) RECOMMENDATIONS .................................................................................................................... 101 Listof Tables Table 1-1: Madagascar's macroeconomic indicators (1995-2003)....................................................... 3 Table 1-2: Madagascar's poverty indicators (selected years 1993-2002)............................................. 5 Table 1-3: Madagascar's fiscal indicators ............................................................................................ 6 Table 1-4: Fiscal performance, 1994-2003 ........................................................................................... 7 Table 1-5: PRSP scenarios for poverty reduction................................................................................. 8 Table 1-6: The PRGFMacroeconomic Framework (percent o f GDP)................................................. 9 Table 1-7: Comparison o f PRSP and PRGFMacroeconomic and Fiscal Frameworks...................... 10 Table 11-1: Madagascar's budget allocationtrends selected years between 1997-2004 ..................... 17 Table 11-2: Madagascar's recurrent and investment allocationpatterns selected years between 1997-2004(Percent ofnon interest expenditure) ............................................... 18 Table 11-3: Distribution o f public allocations within the general administration............................... -21 Table 11-4: Execution o f actual public expenditures across different administrative categories .........21 Table 11-5: Budgetary allocations 1997-2004..................................................................................... 22 Table 11-6: 23 The 2003 budget process andproposed calendar.............................................................. Structure o f the Malagasy Budget..................................................................................... Table 11-7: 29 Table 11-8: Madagascar's past executiontrends (selected years between 1997-2003) ....................... 35 Table 11-9: Madagascar's recurrent and investment spending patterns selected years between 1997-2003(actual expenditures as percent o f budgetary allocations) ................36 Table 11-10: Civil servant rate for 100inhabitants insome African countries...................................... 46 Table 111-1: Number o fnew entrants inprimary school....................................................................... 53 Table 111-2: 53 Net enrollment across Faritany 2001................................................................................ Survival Rate and Gross Intake Rate inlast grade o f primary.......................................... Table 111-3: 54 Table 111-4: Madagascar- Education Sector: Central Government allocation and expenditure 55 Primary and secondary expenditures in2000 and 2003 ................................................... and private sector spending (selected years between 1997 and 2004).............................. Table 111-5: 57 Table 111-6: Budget allocationto primary education (selected years between 1997-2004) .................59 Table 111-7: Allocation o f recurrent budget expenditures to administration, primary and secondary schools andtransfers selectedyears between 1997-2003 ....................... -60 Table 111-8: Percentage of total public expenditure on primary and secondary school 63 Distribution o f primary teachers across zones 2002103 and 20031041.............................. across income quintiles..................................................................................................... Table 111-9: 64 Table 111-10: Executionrates (inpercent) o f non salary recurrent expenditures by education level .....66 Table 111-11: Studentheacher ratio inprimary school across years........................................................ 67 Table IV-1: Access o f populationby vehicle type (1999) and poverty rates (1999) ........................... 72 Table IV-2: The Size ofthe Transport Sector ...................................................................................... 73 Table IV-3: Governance features o fnew public agencies.................................................................... 75 Table IV-4: Madagascar's past allocations to the transport sector....................................................... 77 Table IV-5: Madagascar - execution rate o f the transport budget ....................................................... 78 Table IV-6: PRSP Expenditure Forecasts 2004-2006.......................................................................... 79 Table IV-7: Public Enterprises inthe Transport Sector ....................................................................... 80 Table IV-8: RoadNetwork Quality in2002 and Objectives for 2008 ................................................. 81 Table IV-9: 82 NEAP andPRSP objectives.............................................................................................. RoadMaintenance: Total requirements and sources o f funding ...................................... Table V-1: 89 Table V-2: Executingagencies for the NEAP..................................................................................... 91 Table V-3: Madagascar - Environment Sector: Central Government allocation and expenditure......93 Table V-4: Madagascar's recurrent and investment spending patterns inthe 95 Summary o f some institutional issues and reform status .................................................. environment sector over the years 1997-2003.................................................................. Table V-5: 99 List of Charts Chart 1-1: 1 Chart 11-1 a-c: Deconcentration o fpublic services ................................................................................... Per capita GDP growth and levelrelative to SSA .............................................................. 19 Chart 11-2: Chart 11-3 a-b: Comparison o f planned and actual revenues and donor funds,......................................... Annual growth o f non salary recurrent expenditures and investment 1998-2004.............31 Chart 11-4: Comparison o fpast budget allocationand executionbetween 1997 and 2003 .................33 37 Chart 11-5: Trends infiscal revenues and budget executionrate......................................................... 38 Chart 11-6: Distribution o f civil servants by sector in2004l .............................................................. -47 Chart 11-7 a - b: a Evolution and distribution o f the Malagasy civil service by functional level 1997-2004 48 Geographical distribution o f civil service agents 1997-2004........................................... b Qualification o fCivil servants by functional level....................................................... Chart 11-8: -49 Chart 111-1: Net Enrollment across quintiles inrural areas .1999 -2001 ........................................... 54 Chart 111-2: Public expenditure ineducation as percentage o f GDP 2000 ........................................... Chart 111-3: RepetitionRates for selected countries inSub-Saharan A h c a in200112002 ..................56 58 Chart 111-4: Madagascar -Execution rate o f the Education budget 1997-2003................................... Chart 111-5: Executionrate o f investment expenditures per source o f finding 1997-2003..................61 62 Chart 111-6: 63 Distribution o f expenditures by environmentaltheme (200 1).......................................... Benefit incidence o fpublic expenditure ineducation 2001.............................................. Chart V-1: 96 Listof Boxes BOX1-1 The 2002 crisis.................................................................................................................... 4 BOX1-2 2004 Priority Action Plan - A government-led reformprocess......................................... HIPC Debt Relief.............................................................................................................. 12 BOX11-1 15 BOX11-2 Programbudgeting ........................................................................................................... 24 BOX11-3 Developing a strategic decision making process .............................................................. 27 BOX11-4 28 Dual budget systems ......................................................................................................... The Budget Framework Paper .......................................................................................... BOX11-5 30 BOX11-6 The impact o fbudget support.......................................................................................... 32 BOX11-7 38 Performance or results-based management ...................................................................... Madagascar's legal instruments to modify the national budget........................................ BOX 11-8 44 BOX11-9 Common problems of monitoring & evaluation systems ................................................. 45 BOX11-10 51 Critical ingredients for a road maintenance fund.............................................................. Citizens report cards - a tool to asses public sector performance .................................... BOXN - 1 76 Annexes ANNEX 1: SUMMARY OFPUBLIC FINANCERECOMMENDATIONS ANNEX 2: MADAGASCAR'S POVERTY REDUCTIONSTRATEGY ....................................... .................................. 107 ANNEX 3: ASSESSMENT OF MADAGASCAR'S BUDGETMANAGEMENT CAPACITY 120 ...117 ANNEX 4: REFORMOF THE PUBLIC FINANCE MANAGEMENT SYSTEM ....................... 124 ANNEX 5: STATISTICAL TABLES (PUBLIC FINANCE) ........................................................... 130 ANNEX 6: THE PHASESAND PROCEDURESOF THE BUDGET CYCLE ............................. 135 ANNEX 7: ORGANIZATION STRUCTUREAND REFORMPROGRAMOF INTERNAL AND EXTERNAL CONTROL INMADAGASCAR .................................................... 137 ANNEX 8: MODERNIZATION OF THE LEGALPUBLICFINANCE FRAMEWORK 2004 ....................................................................................................... 139 ANNEX 9: THE MALAGASY EDUCATIONSYSTEM-STRUCTUREAND INDICATORS .................................................................................................................. 142 ANNEX 10: STATISTICAL TABLES (EDUCATION SECTOR) .................................................. 148 ANNEX 11: THE ENVIRONMENT SECTORINMADAGASCAR 150 ANNEX 12: STATISTICAL TABLES (ENVIRONMENT) ............................................................. .............................................. 154 i EXECUTIVESUMMARY This Public Expenditure Review (PER) analyzes the Government o f Madagascar's efforts to meet the challenge o f poverty reduction. The report will first evaluate the macroeconomic and fiscal framework within which the Government i s implementing its Poverty Reduction Strategy. Secondly, it will review the resource allocation and current budget management process inMadagascar and the extent to which the current budgeting and budget management process allows Government to effectively implement the PRSP. This section will also deal with overall budget execution in implementing Government policies and the use o f human resources. The final sections will analyze the effective use o f public resources inthree priority sectors which are crucial for the implementation o f the PRSP (education, transport and environment), and absorb the largest sectoral allocations.' It is envisaged to cover other sectors insubsequent PERs, on the basis o f a multiyear work program with Government and other development partners. For the next PER exercise, agriculture, nutrition and social protectionwould be strong candidate sectors. Expectations vis-&vis the new Government are enormous - both in Madagascar and abroad. At the same time the Government is facing signijicant challenges to turn around many years of failed policies and mismanagement. This report analyses a small but essentialpart of these challenges, in particular in the areas of macro-economic management, public finance and sewice delivery in selected sectors. The number of "issues" ident8ed in these areas are substantial and would raise questions of reform focus, prioritization and sequencingfor every Government, also in the developed world. The intention of the report is not to outline what does not work in Madagascar but toprovide a holistic analysis that constructively contributes to refinement and implementation of the reform policy of the Government. It is also important to note that the Government has not had much time to implement its reform program - the analysis in this report should be seen against the background of only two years of change. It is evident that the implementation of substantial changes will take more time. A. DEVELOPMENT CONTEXT Madagascar is a large country which is sparselypopulated. It spans about 587,000 square km, or approximately the size o f France and Belgium together, Kenya or Thailand. Population density in Madagascar i s about 28 people per square km, close to Mozambique's density o f 24 people per square km, but much less than Kenya (55) or Thailand (121). Madagascar is also mountainous and mostly rural. In2004, eighty percent o f the country's 17 million people lived inrural areas. Madagascar is one of the poorest countries in Africa. During the three decades until 2001, Madagascar's per capita income declined by 40 percent. It fell from US$402 in 1970 to US$242 in 1999, and i s expected at US$235 in 2004. This period stressed self-sufficiency and extensive state intervention in all sectors based on a socialist model o f society. Initial reforms to overcome these years o f economic mismanagement were launched in the mid-1980s. These reforms were aimed at liberalizing the economy and privatizing public enterprises. These changes, however, as well as similar reforms in the early 1990s were only piecemeal and not implemented decisively. Consequently, they had only marginal economic impact. ' A companion report to the PER was prepared, a Development Policy Review (DPR), with the objective of examining development policy reforms that Madagascar will need to implement to achieve pro poor growth. The PERcomplementsthe DPRby assessingpublic sector resource availability andprioritization. 11 Structural reforms brought Madagascar's economy to a turning point in 1997. The period 1997- 2001 i s characterized by more substantial reforms focusing on creating a favorable environment for private investment and on integrating Madagascar into the world economy. As a consequence ' economic growth picked up substantively - during this period the growth rate averaged 4.6 percent, representing the best economic performance in 40 years. Main sources o f economic growth were the activities o f the Export Processing Zones ("zones franches"), shrimp productiodexportation and tourism. The impact of these reforms on nationalpoverty reduction, however, was limited. Economic growth mainly benefited the urban population while poverty in the rural areas increased over the period 1997-2001. The perception that government services were flawed by corruption and serious governance problems contributedto a widespread dissatisfaction with the regime inpower. Poor Government performance contributed to the political crisis in 2002. After the general elections in December 2001, the incumbent President Ratsiraka contested the results and two parallel Governments were established, each with its own central bank, leading to a freezing o f Madagascar's assets abroad and a suspension o f foreign exchange. Duringthis period economic activity fell sharply and poverty increased significantly. The political crisis ended in July 2002 with the departure o f President Didier Ratsiraka, and the inauguration o f Marc Ravalomanana, a self-made business man, as President. B. CREATINGANENABLINGMACROECONOMIC AND FISCAL FRAMEWORK Since independence in 1960, Madagascar has suffered through a lengthy downward spiral and i s today one o f the world's poorest countries, with per capita GDP estimated at US$235 in 2004. Over the course o f three decades o f interventionist policies and state control, real incomes fell by 40 percent and the headcount poverty rate rose above 70 percent o f the population. Recently, structural reforms and improved macroeconomic management have started having a positive impact, givingrise to hopes that Madagascar has turned the corner. The period 1997-2001 saw sustained positive growth, which representedthe best outturn for more than three decades. GDP fell by 12.7 percent, however, in 2002 following a disputed election and political crisis. After a rebound in 2003, Madagascar was hit by a number o f exogenous shocks, including two cyclones in early 2004, a sharp depreciation inthe FMGand a collapse o fvanilla prices. The country continues to be vulnerable to risk. Social indicators are poor. Access to medical services, education, electricity and safe water i s low, and infant mortality i s high.Moreover, in spite o f better macroeconomic performance, overall poverty impacts have been limited because growth has been concentrated in a few export sectors and urban areas and has not spread to rural areas three fourths o f the population live. Indeed, while the nationwide poverty headcount fell from 73 percent to 70 percent in 1997-2001, rural poverty did not decrease and in 2002 jumped to an estimated 86 percent. Limited human capital and access to credit, and poor public services depress agricultural productivity and there are few opportunities for wage employment. A third o f the rural populationhas no road access. Government's development plan is overly ambitious. In 2003, Madagascar finalized an ambitious Poverty Reduction Strategy Paper (PRSP) which proposes to reduce poverty by half in 10 years, from 70 percent in 2003 to 35 percent in 2013. The plan centers around three strategic axes: (i)restoration o f the rule o f law and strengthening o f good governance, (ii)broad based economic growth, and (iii) human development and social protection. The PRSP envisaged that economic growth would be led by foreign investment in agriculture and in textiles, and strong public sector investment in infrastructure and social services. Yet Madagascar must overcome serious capacity constraints to ... 111 implement this agenda and public sector performance i s weak even compared to other low income countries. Despite recent efforts on the revenue side, the tax base remains highly concentrated, and customs administration i s hampered by slow processing and corruption. HIPC debt relief has significantly expanded the fiscal envelope, but poor budget execution remains a problem. Economic targets require more realism. Halvingpoverty ina decade implies an average growth rate o f 9.3 percent (compared to 4.6 percent in 1997-2001). The Joint Staff Assessment o f the PRSP concluded that while the macroeconomic and fiscal framework was basically sound, such a growth target was unrealistic and should be lowered. Recent reviews by the Bank and the International Monetary Fund (IMF) recommended lowering growth projections in the medium to longer term scenarios to a more plausible 6 percent. The most recent PRGF review (September 2004) projects above trend growth o f seven percent in 2005-06 during the recovery phase after the 2002 crisis, easing back to 6 percent in the longer term. In striking contrast, the PRSP anticipates growth accelerating from 6 percent in2003 to 8 percent in2006. Onthe basis o f an unexpectedly strong reboundof 9.8 percent in2003 (outpacing the IMF's program target o f 6 percent), the PRSP appears on track in the near term. But the PRSP's macroeconomic scenarios have a horizon o f 2006 and provide no details about performance beyond that point, making it difficult to evaluate longer term expectations. Nevertheless, the near term acceleration o f the PRSP compared to the steady state trend in the PRGF suggests there i s a significant divergence o f views about the longer term. At the very least, the Government needs to prioritize its spending plans and be prepared to adjust the program to maintain fiscal sustainability. Financial sector should be strengthened. The weakness o f the financial sector poses an additional problem for Madagascar's fiscal framework since it has precluded better financing o f the private sector. Private sector borrowing averaged 9 percent o f GDP over the period 1998-2002, less than half o f the average for African countries. To further stimulate private sector growth the financial sector needs to be strengthened, in addition to maintaining a stable macroeconomic environment an in particular controlling inflation. C. THE CHALLENGE OF RESOURCEALLOCATION AND BUDGETEXECUTION To improve resource allocation and budget execution, the Government has launched a comprehensive reformprogram, which i s in its early stages. The central instrument i s an annual Priority Action Plan, which sequences intended reforms in line with the absorptive capacity o f the public finance system. Recent achievements o f the government's action plan include a new Organic Finance L a w which introduces a modern institutional and procedural public finance framework, a new procurement code to address deficiencies inpublic procurement, the establishment o f a new internal control cadre inthe Ministryo f Economy, Finance and Budget (MEFB), and improvements inthe operational efficiency o f the treasury department. Despite these efforts, Madagascar meets presently only four out o f fifteen criteria which were developed to benchmark public finance systems in developing countries. Overall issues that the Government i s facing in the public finance management process are listed below and specific challenges in budget allocation, execution and human resources are presented in more detail insub-sections ofthe chapter. 1v Budgetary allocations are only partially in line with PRSP priorities. Allocations for primary education increased by 39 percent and for primary health care by around 15 percent over the period 1997-2004. Allocations for productive sectors (inparticular energy, agriculture and environment) fell substantively - from 33 percent of non-interest expenditures in 1997 to 26 percent in 2004. However, public resources allocated to the transport sector increased by annually 18 percent on average over the past five years. The share o f budgetary allocations to "governance" (defined by the Government as the sum o f allocations for general administration, defence and public security including the judiciary) has been reduced since 1999. General administration, however, continues to take the lion's share o f the budget (on average 36 percent over the observed period). Bringing Government closer to thepeople. The distribution o f budgetary allocations to sectors are not yet in line with the administration's intention to "bring Government closer to the people". It i s unclear to what extent the allocations for the general administration (about 33 percent o f non-interest expenditures in2004) contribute to achieving the goals and objectives o f the PRSP. At the same time, key PRSP sectors, inparticular the agriculture and health sector, have been marginalized with only a share o f the budget of 5.2 and 7.4 percent of the total budget in2004. Interms o f decentralization, the central Government maintains control over the bulk o fbudgetary resources; only 3 to 4 percent of the national budget i s managed by the communes, the only operational level below the central Government. The credibility of the budget continues to be low because of poor revenue forecasting and unbudgeted expenditures. The gap between revenue forecasts and revenue inflows i s substantial: revenues had to be adjusted downwards by 22 percent in2001 and 44 percent in2002. Main problems include lack o f a clear methodology and lack o f capacity in the relevant departments to accurately project inflows. Due to a significant improvement of the budget execution rate in 2003 (92 percent) this gap couldbenarroweddownto 4 percent. Budget execution rates need to be closely monitored. Budget execution rates have deteriorated since the end o f the 1990s from 96 percent in 1998 to 75 percent in 2001, mainly because o f weak revenue collection and o f budget implementation problems. Estimates for the 2003 budget execution show a 92 percent execution rate, which would indicate that Government's efforts to streamline the execution process begins to bear fruit. The improvement in execution can be mainly attributed to recurrent expenditures (salary and non salary expenditures) while shortcomings continue to exist in the implementation of public investment projects. Over the period 1997-2003 average execution rate for recurrent expenditures was 89 percent but only 68 percent for capital expenditure. HIPC funds (representing 6 percent of the Government's budget in 2003) which were executed based on simplified procedures reached an executionrate o f 99 percent. Budget execution problems, notably o f capital expenditures, are mainly caused by cumbersome and complex disbursement procedures. Government has begun to rationalize disbursement procedures and i s putting in place an integrated financial management system, which will streamline the expenditure management process. Budgetpreparation faces significant challenges. These are: (i)lack of reliable macroeconomic and a fiscal parameters guiding the preparation process; (ii) little involvement o f sector ministries in the preparation process, (iii) no systematic mechanism to discuss budgetary disputes-inter-governmental consultation about the budget i s limited to budgetary conferences; and (iv) strategic discussion and decision-making on fiscal and budgetary policies at the Cabinet level i s limited and ad-hoc. Inall, the allocation o f public resources inMadagascar needs to be based on a reliable expenditure management framework that clearly links policy priorities to the budget and prioritizes needs within a hard budget constraint. V The problem of separating recurrent and capital budgets. Although a consolidated budget i s presented to Parliament, the Government i s still operating on a dual budget system with different procedures, timelines and institutional responsibilities for recurrent and investment expenditures. This dual budget systemreinforces the inconsistenciesbetween investment and (non-salary) recurrent expenditures. The Government should consider to filly integrate the preparation o f recurrent and investment budgets. It should harmonize the preparation process and develop a uniform budget proposal, which addresses the cost implication o f investments within a medium-term envelope. Possible solutions. To address these problems the Government can (i) introduce a Budget Framework Paper, which outlines - early in the budget preparation process - the macroeconomic framework under which the budget i s being developed; (ii) enhance the role o f Cabinet by systematically ensuring strategic decision-making at the different stages o f the budget preparation process, (iii) strengthen the participation o f line ministriesbased on an effective planning and consultation process, and (iv) refocus budgetary conferences on policy objectives, prioritization o f needs within the hard budget constraints and results achieved on the bases o f resources allocated. To accommodate these changes, the calendar for the preparation o f the annual budget would need to be adjusted; in particular, the submission o f the full budget to Parliament could be delayed to ensure that Parliament can vote on a high-quality budget. Efforts made by the Government to develop a medium-term macro-economic andflscal perspective and to introduce program budgeting (as o f budget year 2005) are important steps to better linkpolicy priorities to the budget and to provide a medium-term perspective about the affordability o f Government policies. Experiences from other countries, however, indicate that the introduction o f program budgeting will pose a significant challenge for the public administration; the reforms should be complemented by a systematic program o f training and capacity building which aims at establishing or enhancing relevant capacities inline ministries. Execution of the budget is hampered by administrative bottlenecks. As processing payments in Madagascar i s complex with around 11 different stages that each requires a formal approval, "simplified procedures" to release funds are frequently applied. These procedures should, however, be reserved for exceptional cases. Inaddition, shortcomings in the procurement and control system increase the administration's transaction cost and provide the ground for substantial inefficiencies. A flawed system of financial monitoring. Key information is not available, in particular execution data on investment projects financed by donors. Budget reportingby line ministries i s at an even more embryonic stage: since 2000 only the social sectors (education and health) produced regular budget execution reports. In 2003, additional reports were prepared by the ministries o f agriculture, justice, water and forest, and public works to fulfill a Government commitment inthe HIPC context. Finally, there i s no systematic reconciliation process between the Treasury and the Central Bank. As a result o f these deficiencies the Treasury faces difficulties inclosing the annual accounts within the statutory timefiame and the Auditor General is not able to validate these accounts as requiredby the law. Over the past years, the treasury accounting system was strengthened to improve the closing o f Government accounts in a timely manner. As a result the final accounts for the years 1997 to 2001 have been sent for review to the Auditor General. The accounts for the years 2001, 2003 and 2004 are expected to be submitted in2005. Government reforms are underway to address these problems. Reforms to date focus mainly on improving the operational efficiency o f the Treasury. Government should also build on the joint budget allocation reports presented by key ministries in 2003 and develop an integrated reporting system that can serve different purposes (e.g. PRSP, business plan implementation, budget execution). v1 The need to transform the budget into a management instrument. The budget should be the main instrument to facilitate implementation o f Government policies. InMadagascar, however, the budget i s not seen as a reliable planning instrument because actual releases typically vary substantially from the allocated amount. As a consequence, the budget i s not used as a management instrument that guides the implementation o f policies and programs. There i s also no effective monitoring & evaluation mechanism inplace that systematically tracks implementation o f Government policies and programs. Introduction of basicperformance management under way. To strengthen the results orientation o f the administration Government introduced business plans for all line ministries which outline the work program for the respective year. Many business plans, however, lack realism and detailed costing; they also do not feed into the program budgets, which are prepared in a separate exercise. Government should integrate the different activities into a coherent results-based management framework. Under such a systempolicy-related information i s systematically collected at all levels o f Government, consolidated by a well-functioning mechanism at the centre o f Government and scrutinized for potential follow-up. The system should generate regular reports about the status o f the implementation of the Government program, which should be sent to Cabinet for discussion and strategic decision-making. D.EFFICIENT USE OFHUMAN RESOURCES Improvement of the efficiency of the civil service needed. The Malagasy civil service accounts for 146,000 agents in 2004, which represents 0.8 percent o f the population. Education employs the largest share of civil servants (around 44 percent), after public security (15 percent) and administration (12 percent). Several previous attempts to reform the civil service failed because they were overly ambitious and did not address underlying systemic problems. As a consequence the public administration continues to be flawed by governance problems, in particular corruption, inefficiencies and mismanagement. Government efforts to address these problems are under way. They concentrate on improving operational efficiency and performance orientation o f the administration. Govemment is also considering a comprehensive administrative reform program to change the existing institutional and procedural set-up. Govemment should learn from experiences in other countries in designing and implementing its administrative reform program. These experiences overwhelmingly indicate that comprehensive reforms o f the civil service are extremely difficult to implement and to sustain. A more promising approach would focus on interventions that target specific public services with large public-private interface (such as primary education, health centers, customs, land titling, and the lower courts). The objective would be to improve service delivery and produce visible results in the short term. While these reforms cannot substitute a long-term vision for the reformo f the civil service they keep the reform program focused on tangible results and, thus establish muchneeded credibility. The need to address structural issues.. Potential reform should also look at the composition o f the . civil service. As in many other developing countries the Malagasy civil service i s characterized by a relatively high number o f support staff, around 60 percent o f the civil service in 2004. At the same time, the intermediate level (inparticular technical specialists) who typically form the backbone o f a good civil service account for only 30 percent o f the total staff. The management level i s also rather small with less than 8 percent o f civil servants. This composition inpart explains existing bottlenecks in delivering public services - not enough qualified personnel is available to deal with the various administrative tasks. vii .,.and staffing at the service delivery level. It will also be important to increase the number o f civil servants allocated to the periphery to correspond to the vision o f an administration which i s close to the people. The number o f civil servants i s currently 2 per 100 inhabitants in urban centers while the ratio i s only 0.75 for rural areas. This ratio constitutes another explanation for existing service delivery problems -not enough civil servants are allocated to front-line services. Government should align human resources to budgetary priorities. Duringthe budget preparation process no systematic review o f the staffing allocations o f the sector ministries takes place. Staff and payroll estimates are based on the staffing plans prepared by line ministries in collaboration with the Ministryof Civil Service. However, at the level ofthe sector ministriesneither job descriptions nor a breakdown o f staff against goals and objectives o f the respective ministry or department exist. Line ministries should be required to prepare and regularly update recruitment and staffing plans as part o f the budget preparation process. These plans should be reviewed to correspond to Govemment priorities in separate manpower hearings with the MEFB and the Ministry o f Civil Service. The final plans should be submitted to Cabinet with a substantiation that the allocation o f human resources i s in line with the goals and objectives o f the Government. E.USEOFPUBLICRESOURCESINTHE EDUCATIONSECTOR The education sectors reforms yielded significant results, but more challenges are ahead. The Malagasy education system has many strengths; it can inparticular buildon a large network o f 19,000 primary schools that serve more than 3 million children. Policy reforms, during and after the 2002 crisis, such as the elimination o f school fees inprimary schools and the distribution o f textbooks and school kits,resulted into a substantial increase inenrolment. Among the key challenges are : Low survival rates. Enrolment for primary education has significantly increased over the last two years. The key challenge for the Govemment i s to sustain the enrolment increases and to ensure that the children who are enrolled complete the five years o f primary education. Of all children in school age only 40 percent complete the five year cycle o f primary education. Out o f 100 children who start primary school less thenhalf complete their five years o f primary education inthe year 2003/04. o Reduce drop-out and repetition rates. Repetition rates at the primary school level have been above 30 percent in the last years, the highest rate in Sub-Saharan Africa. Only 4 percent o f the children do not repeat a class and 25 percent repeat two or more times during primary school. Inthis context, private primary schools have significantly lower repetition rates (16 percent compared to 33 percent o f public schools in 2003-2004), which indicates that the quality o f education i s better at the private school level. o Improve efficiency and cost-effectiveness. Efficiency remains a key problem o f the education system. Of those who complete primary education, around 66 percent repeated at least a grade and only few o f them will continue to junior secondary education (44 percent). Cost-effectiveness i s low: To produce a 5th grader the Malagasy education system spends 2.7 times more than in a system without dropouts or repeaters. Diminish disparities and inequities in the system. Enrolment shows strong regional disparities. Enrolment rates in urban areas were above 70 percent in all provinces, the rates in rural areas were less than 50 percent. As the majority o f the poor live inrural areas this pattem indicates a regressive distribution o f primary school benefits. There are also large income variations in enrolment. From ... Vlll 1999 to 2001 the enrolment rates o f children belonging to the poorest income levels decreased while the ones belonging to the highest income levels increased. Moreover, the poorest districts are those with the largest studendpublic teacher ratios. Government faces difficulties to hire and maintain teacher in remote areas. Publicly paid teachers are much less allocated to villages where the average income level o f the populationi s low. Madagascar's commitment to "Education for all" provides a unique opportunity to tackle the challenges. T o confront these challenges the Government has formulated a new strategy to reform the education system and has committed itself to achieving "Education For All" (EFA) with a 100 percent school completion rate by the year 2015. This new strategy i s supported by the international community, which has pledgedto help Government reach its education goals. Per student allocation in primary education has hardly changed. Budgetary allocations for education have significantly increased in the last years. While in 1997 the budget allocation was about 2 percent o f GDP the 2004 budget allocated 3.6 percent o f GDP, which i s in line with the Sub- Saharan Africa average o f 3 percent o f GDP (in 2000). This increase mostly focused on basic education, especially the primary schools. Despite these increases the per student allocations in primary education remained almost unchanged with 15 percent o f the per capita GDP in 2000 and 2003. This i s due to significant increases in enrollment. At the same time GDP per capita allocations for secondary educationreached46 percent in 2003, from 28 percent in2000 suggesting that students insecondary educationbenefitedfrom the increase ofpublic resources, despite a drop inthe share of resources allocated to secondary education. Management information is limited. While the ministry collects standard information about the performance o f the school system in the country this data i s not produced in a timely manner to feed into policy decisions. Budget preparation i s typically based on outdated information, which poses significant problems inadequately determiningbudgetary needs. Decision-making remains highly centralized. The division o f responsibilities between the policymaker and service provider functions inthe education ministry should be krther improved. The school districts (CISCOs) are meant to be responsible for the provision o f primary education services; the ministry for the design, supervision and regulation o f policies for the sector. In the past, the ministryretainedmajor delivery functions: it directly managedteachers and controlledall investment to the schools. Evidence were the relatively high administrative overheads at the central level. Since 2003, efforts are under way to progressively transfer service delivery functions from the central govemment to the service delivery levels. District agencies should be strengthened. Budget execution rates o f non-salary expenditures at the CISCO level (on average 86 percent between 1997 and 2001) are lower than the rates o f the central ministry or o f the provincial administration, on average 92 and 90 percent over the same period. Limited (public financial management) capacity, lack o f equipment and the burdensome procurement and disbursement processes are key factors in this context. To address the low performance o f CISCOs key measures were launched to improve capacity; as a consequence the execution rate o f non salary expenditures improved to 91 percent in 2003. Expenditure tracking surveys have also found significant resource leakages at the CISCO level (55 percent o f the schools report leakages o f in-kindtransfers). Budgetary performance o f the CISCOs should be systematically strengthened. It will also be important to rethink the distribution o f budgetary resources among the districts and the communes on the basis o f needs, as determined by the number o f students or the conditions o f the school infrastructure. ix Capitalizing on the existing system of community participation. In Madagascar, parental associations (FRAMs) play an important (and widely acknowledged) role in school issues. FRAMs provide in-kind support for the schools, in particular related to in school maintenance. In some communities these associations hire teachers and pay their salaries. The role o f the FRAMs has been challenged by the creation o f FAFs (community-government organization), which manage financial transfers to schools from the central Government ("caisse d'kcole" program). The creation o f the FAFs raises questions about their relationship with the FRAMS; this relationship should be clarified to ensure effective coordinationo fthe support to the schools provided by the various actors. F. USEOFPUBLICRESOURCESINTHE TRANSPORT SECTOR Historically, Madagascar`s transport sector has been dominated by direct state intervention with heavy reliance on Government agencies and parastatals to provide services. With the exception o f road freight transport, Government had, directly or indirectly, managed and financed railways, aviation, airports, shipping, and ports. As a result, today's service i s poor, operating efficiency i s low, safety standards are inappropriate, maintenance has been neglected, and funding for capital investments has been difficult to mobilize. InApril 2000, the Government adopted a comprehensivetransport sector policy and strategy, which seeks to reverse past policies. It aims at (i) focusing Government's role on strategic planning, sector oversight and coordination; (ii) creating jointly controlled public-private and user-financed agencies for sub-sector management and regulatory functions; (iii)divesting operational activities to the private sector, through privatization and concessioning arrangements; (iv) developing the local private sector for works design and execution; and (v) rehabilitating transport infrastructure to appropriate levels. In implementing its new vision the Government i s facing important challenges which are summarized below: o Poor service quality in the road sector. Over the past 30 years, road quality inMadagascar has deteriorated: the country has lost about 1000 kilometers o f roads per year primarily due to poor maintenance. Today, 81 percent o f roads are in poor condition. The share o f paved roads i s 11.6 percent, compared to an average o f 13 percent for Sub-Saharan Africa, 16 percent for low income countries worldwide, and 53 percent for lower-middle income countries. On average, only about half o f the populationi s accessible by motorized vehicle. Air transport system in urgent need. Most airports require urgent rehabilitation to bring infrastructure to international safety and operations standards, inparticular with regard to safety and security equipment, telecommunications, radio navigation, expansion o f terminals and runways, and improvements inroad access. Port services inadequate. Of the country's 17 ports, only 4 are deep enough to permit unloading from large vessels. Cargo handling technology inthe country's main port, Toamasina, i s outdated: the port unloads containers at a rate o f 8 per hour compared to a rate o f 25 per hour in neighboring Reunion and 30 per hour inneighboring Mauritius. o Investment needs enormous. The PRSP assumes for the 2003-2006 period investments o f about MGF 7 trillion, or close to US$1 billion (with the exchange rate used inthe PRSP). This represents about 45 percent o f the total investment needs identified inthe PRSP. X Comprehensive reforms are under way to improve Operational efficiency of the sector: the Government merged three ministries and created a Ministry o f Transport and Regional Planning. The key role o f the ministryi s to design transport policy as well as to coordinate all activities inthe sector. Regulatory functions will be managed by autonomous agencies, which are spin-offs from old ministerial departments or existing agencies. These autonomous agencies include: the Road Management Fund (created 2003), the Civil Aviation Authority (created 1999), the Road Authority (planned for 2005) and the Port, Maritime and k v e r Agency (created 2000 but not yet operational due to lack o f funding). Under the new administrative framework Government also plans to gradually devolve some road maintenance activities, management o f smaller airports and ports to local authorities. Inline with the new policy the Government has begun to increase the role o f the private sector: Air Madagascar has been put under a private management contract; similar contracts are planned inother areas (railways, ports, airports). Key issues include: o Affordability and realism. Madagascar expects to allocate on average 2.7 percent o f GDP inthe period 2004-2006 to the transport sector. Against the background o f the past performance in the sector these allocations raise serious questions concerning affordability and implementation realism. In the 1997-2003 period, the Government budgeted transport expenditures of about 2.3 percent of GDP and executed only about 1.5 percent o f GDP. o Contingent liabilities. The total amount o f contingent liabilities and debt stemming from the transport sector i s unknown. This poses a significant risk for the Government's budget. These liabilities needto be clearly established and addressed by a debt management plan. Low budget execution rates. Between 1997 and 2003, on average only 71 percent o f the transport sector budgets were implemented. This i s mainly due to low execution rates o f the investment budget (the execution rate o f investments deteriorated from 96 percent in 1997 to 89 percent in 2003). Some o f the sub sectors (rail, maritime and ship) only achieved an average execution rate for the capital budget o f 44 percent in 2004. Main reasons for the execution problems include inadequate procurement and management capacity inthe Ministry o f Transport and Regional Planning as well as complex management and disbursement procedures under the different donor projects. Another problem i s the limited absorptive capacity o f the sector - in particular the private sector faces difficulties in implementing the ambitious Government program. To tackle this problem, the Government intends to invite more foreign firms to build and rehabilitate transport infrastructure. o Improvement of the Road Maintenance Fund. The road maintenance fund has not been adequately funded because o f inadequate allocations provided by the Government. Organizational and technical problems continue to hamper the efficiency o f the fund. For the fund to be more effective it will be important to rethink the funding arrangements based on experiences in other countries. It will also be important to update and rigidly enforce accounting and procurement standards. The fund shouldbe subjected to regular value-for-money audits. Theproblem of autonomouspublic agencies. Government intends to devolve responsibility for the regulation o f the road, air transport, marine, and rail sectors to autonomous agencies. Experience inother countries, however, indicates that the devolution ofauthorityto autonomous agencies is not a silver bullet. The track record o f autonomous agencies i s rather mixed; a particular danger i s to replicate the inefficient administration autocracy at the central Government at the level o f these agencies. International experience also indicates that these institutional changes will require significant implementation time. xi R The issue of capacity of local authorities. Government intends to devolve some activities in the areas o f road maintenance, airport management, port operation and management to local authorities. This devolution needs to be complemented by substantial investments in training and capacity buildingto enable the local communitiesto adequately fulfilltheir new responsibilities. R Inconsistency of budgetary data. The inconsistency o f allocation and execution data between the MinistryofFinance andthe sector ministryraises serious concems about the quality andusefulnessof expenditure information. These inconsistencies should be addressed on a priority basis to enable a transparent review o f implementationperformance against policy targets as specified inthe PRSP. G. USE OF PUBLICRESOURCESINTHE ENVIRONMENT SECTOR Madagascar's economy is in a vulnerable position due to the degradation of its key natural assets, The economy depends highly on the country's natural resources. Agriculture, forestry and fishing account for 32 percent o f GDP. Natural forests provide key environmental services while harboring Madagascar's unique biological resources, a crucial factor in the growth o f ecotourism on the island, However, the loss o f natural forest, primarily to tavy (slash and bum agriculture) and the use o f fuel wood and charcoal, has had a detrimental impact on Madagascar's biodiversity and hurts the growth potential inecotourism. Intackling these environmentalproblemspolicymakers are facing three major challenges: Growingpopulation and agricultural stagnation. The combination o f an expanding population and static productivity has generated pressure for agricultural expansion through conversion o f new land (mostly forests) to agricultural uses. R Poor incentivesfor sustainable property management. Forested lands are the property o f the State in Madagascar. Local communities typically have no claim on these resources and no formal rightto use them.As communities do not own forested lands they haveno incentives to provide better management. Slash and burn agriculture. Communities in the headlands of water basins practice slash and bumagnculture without considering the negative effects ondownstreamusers. The Government adopted a comprehensive environment sector strategy, manifested inthe National Environmental Action Plan (NEAP) and the PRSP. The NEAP was designed as a fifteen year investment program divided into three five-year phases (EP1 1991-96, EP2 97-02 and EP3 04-09). Since its publication in 1990, the action plan institutionalized several sector priorities such as the establishment o f protected areas, rural resource management, and environmental education and training. However, several problem areas require further improvement, for example: (i) the non- application o f policies and regulations, (ii) the reallocation o f public resources in favor o f local communities (at the moment around 70 percent o f the resources are allocated in favor o f the central level), and (iii) awareness buildingo fthe Malagasy population conceming environmental issues. Budget allocations do not sufficiently respond to needs. The environment budget comprises mainly o f capital expenditures, on average 91 percent o f total expenditures over the period 1997-2004; at the same time recurrent expenditures are chronically under-funded. This poses a major problem since the conservation o f natural resources i s dependent on the availability o f adequate resources for surveillance, control etc. xii Budget execution requires improvement. Budget execution, for example o f the executing agencies, i s hampered by lack o f standardized procurement processes, absence o f a harmonized financial management framework and weak financial management capacity. Budgetary information about executing agencies i s limited.As a consequence, the Ministry o f Environment and Forests i s not able to adequately monitor the implementation o f its sector strategy. Institutional reforms key priority. A key priority for the existing environmental program is the rationalization o f the overly complex and bureaucratic institutional set-up. Under the new institutional framework, (i) the Ministry for Environment and Forest will be responsible for policy-making and coordination, (ii) the operational work will be carried out by three specialized executing agencies (ANGAP, ONE and ANGEF), and (iii) responsibility for environment and natural resource more management will be devolved to local communities. At the same time the transfer o f funds to local communities will be made more effective. Experience in other countries indicate that the devolution o f management responsibility to local authorities requires significant investments in training and capacity building. Sustainable funding of protected areas needed. The long term funding o f the protected areas i s at riskas the relevant managementagency is likely to face fundingshortfalls duringthe implementation o f the current environmental program. It will be important to generate additional revenue to close the funding gap, in particular fi-om ecotourism and hotel taxes. The protected area system should be further developed to improve environmental protection and tourism rents. Environmental Impact Assessments alsofor public investments. While all large-scale investments, both public and private, might require Environmental Impact Assessments (EIAs), to date there has been little application of EIA for public sector investments. This situation presents a severe environmental risk, since large-scale public investment i s expected to average more than three times the value o f similar private investment until 2006. This also creates a negative image for the government because o f the double standard applied to private and public investments. Even where public investments carried out EIAs, the required permit fee has not been paid. EIA administration in Madagascar could be placed on a sustainable footing if payments for permits o f public investments are fully budgeted and paid for by all government services at the national and sub-national levels. H. CONCLUSIONAND OUTLOOK The findings o f this report expand and further deepen existing knowledge about the public finance system in Madagascar and its contribution to the delivery o f services in specific sectors. While they acknowledge Government efforts to improve the system o f public finance they confirm significant weaknesses in allocating and executing public resources at the sectoral and cross-sectoral levels. They also indicate that Government policies are not sufficiently reflected in the annual budget. In addition, the mechanisms to systematically monitor and evaluate implementation o f Govemment policies are inadequate; it i s therefore difficult to clearly establish status, impact and implementation problems o f ongoing or intended activities. As a result political decision-making i s based on insufficient information. The findings and recommendations o f the report call for the development of an integrated reform program, which prioritizes intended reforms in view o f the absorptive capacity o f the Government machinery. To facilitatethe development of such a programthe recommendationsin this report are summarizedin annex 1in short, mediumand long-term measures.At the cross-sectoral level it i s anticipated to feed the recommendations into the development o f the Priority Action Plan o f the MEFB for the year 2005. At the sectoral levelsthe recommendations o fthisreport will contributeto a ... Xlll refinement of government's interventions aimed at improving implementation of sectoral policies and delivery o f services. The government i s considering to conduct a public expenditure review on an annual basis. The review would focus on assessing budgetary implementation against policy objectives, and on determining strategic priorities for the subsequent budget year. -1- I,CREATINGANENABLINGMACROECONOMICANDFISCAL FRAMEWORK 1) MACROECONOMIC PERFORMANCEAND THE POLICYENVIRONMENT 1. One of the poorest countries in the world, Madagascar has formulated a PRSP aimed at cutting poverty rapidly through faster growth and improved govemance. The public sector plays a key role in achieving the plan's objectives, through appropriate policies and public investments in infrastructure, education and health. The results are underpinned by an ambitious macroeconomic and fiscal framework which, if achieved, will dramatically reverse the country's long term decline. An unexpectedly buoyant recovery after the 2002 crisis has created some space for the poverty reduction program which for now i s safely within the fiscal envelope. Nevertheless, it may be overly optimistic to project a continuation o f recent growth because this reflects in part a cyclical rebound following the crisis. Moreover, Madagascar has limited public sector capacity and must overcome a legacy o f poor implementation. Thus, Government should be prepared to reprioritize spending plans and make adjustments to ensure fiscal sustainability. 2. Madagascar has suffered through a lengthy decline. After independence, Madagascar achieved a decade o f modest growth, but then inward looking, static policies and poor economic management put the economy on a prolonged, downward spiral. Though such a descriptioncould apply to many countries in Sub Saharan Africa, Madagascar's decline was exceptionally steep. From 1965- 2004, real per capita GDP declined by over 40 percent from US$401 to an estimated US$235, or from 140percent o f the regional average tojust 70 percent (chart 1.1). Chart 1-1: Per capita GDP growth and level relative to SSA . . . . . . . . . . . . . . . .*0 1................................................. ....^.....^.......^.............`I.. I 1966 1971 1976 1981 1986 1991 1996 2001 O 6 ~ 1l G D P per capita(% growthf-relative to SSA x So. Afrida Source. Live Database and IMF 3. Attempts at structural reform began in the 1980s, but a stop-go approach limited their effectiveness. With support from the IMF and World Bank, Madagascar embarked on structural reforms in the mid 1980s, focusing on trade and exchange rate liberalization, market deregulation, parastatal, agricultural and financial sector reform, and private sector development. Sectors targeted by the reforms included fishing (through the introduction o f a transparent license system), mining (through a cadastre update and mining permit legislation), and tourism (through a waiver o f visa requirements). These initiatives briefly stabilized the economy inthe late 1980s, but widespread civil unrest and demands for political liberalization in the early 1990s halted progress and in some cases even reversed it. Even after -2- relative calm had been restored, political infighting and disagreements over policy continued to block substantive reform. 4. Macroeconomic management improvedmarkedly. Since the mid 1990s, numerous measures have dramatically improved the quality o f macroeconomic management. From 1994-97, broad money growth was brought under control, reducing inflation from 61 percent to 4.8 percent; the government deficit was reduced from over 8 percent o f GDP to 2.4 percent and official reserves were built up from less than one month o f imports to 3.5 months. In the external sector the maximum tariff rate was brought down to 30 percent and non-tariff barriers were eliminated. Regional links were strengthened in 2000 byjoining COMESA (though so far there has been little trade apart from Mauritius). The financial sector has been strengthened through increased central bank independence, enhanced prudential supervision o f commercial banks and restructuring o f state-owned banks which had experienced financial difficulties.2 5. Together, structural reforms and improved macroeconomic management underpinned an impressive turnaround in economic performance in the late 1990s (Table 1.1). From 1995, the growth rate increased steadily to reach 6 percent in2001, the fastest pace inmore than two decades, while CPI inflation largely remained in single digits during the 1995-2001 period. Investment and savings rates rose and current account and fiscal deficits narrowed. 6. ..although debt continued to place a heavy burden. Madagascar's external debt was exerting substantial pressure on the public finances and on the balance o f payments. With extemal debt at 101.2 percent o f GDP at end-1999, Madagascar's external debt stood at U S D 4.4 billion; in Net Present Value terms the debt amounted to USD 2.1 billion, 56 percent o f GDP, and 248 percent o f exports o f goods and services. Madagascar's extemal debt was unsustainable even after the full application o f traditional debt relief mechanisms (i.e. under Naples terms obtained in January 1999 from the Paris Club). Madagascar's debt to GDP ratio would have stayed above 150 percent through 2008; while the debt to exports ratio would decline from 11.7 percent in 2000 to 8.7 in 2009, and the debt service to govemment revenue ratio would decline from 23.1 percent in 2000 to 13.3 percent in 2009. Despite the good economic performance through the late nineties, Madagascar, being among the poorest countries inthe world, would not be able to pay for its imports and would still have to pay close to a quarter o f its government revenues to meet interest payments on extemal debt. Madagascar qualified for the enhanced debt relief under the HIPC initiative in December 2000, which effectively reduced debt service payments by about 1percent o f GDP from 2001 onward^.^ IMF 1999 See Decision Point Documentfor Enhanced Highly Indebted Poor Country Initiative - Madagascar, December 1,2000, IMF andIDA for fiuther informationon Madagascar's debt situation. -3- Table 1-1: Madagascar's macroeconomic indicators (1995-2003) 1995 1996 1997 1998 1999 2000 2001 2002 2003 GDP (% change) 1.7 2.1 3.7 3.9 4.7 4.8 6.0 -12.7 9.8 CPI Inflation(%) 49.0 19.8 4.5 6.2 9.9 11.9 7.4 15.8 0.2 Investment rate (% GDP) 10.9 11.6 12.8 14.8 14.9 15.0 18.5 14.3 16.0 Privateinvestment rate (% GDP) 5.2 5.0 6.3 6.9 8.0 8.3 11.2 9.4 10.2 National savings rate (x GDP) -3.3 -0.1 0.8 -0.6 2.6 2.7 9.9 3.3 5.3 Real effective exchange rate (% GDP) 42.4 53.9 48.5 48.9 47.6 52.4 57.8 62.5 65.6 Fiscal bal., incl. Grants (YOGDP) -8.0 -11.6 -5.6 -9.0 -5.7 -5.7 -6.8 -6.8 -5.5 Current account balance (% GDP) -8.4 -5.0 -5.5 -7.4 -5.4 -5.6 -1.3 -6.2 -4.9 Domestic debt 11.0 9.7 9.6 12.0 11.7 .. .. 14.1 13.9 Source: Live Database and IMF,IIIPC' PrelinaDocumelit & FifthReview of PRGF 7. A political crisis in 2002 severely disrupted the economy,though the impact was short lived and performance has remained strong. Once again, however, politics intervened when in December 2001 a disputed election brought the country to the brink o f civil war (box I.l).3 As a result, GDP fell by 12.7percent in2002, while inflationreachednearly 16percent andthe investment rate fell from 18.5 to 14.3 percent o f GDP (see table 1.1). The external current account balance worsened from -1.3 to - 6.2 percent o f GDP. On the fiscal side, revenues collapsed from 14 to 10 percent o f GDP, as did expenditures, from 17.6 percent to 15.7 percent (see table 1.3). However, 2003 saw an unexpectedly strong rebound, with real growth reaching 9.8 percent and inflation virtually eliminated - eclipsing the IMF's program target o f 6.0 percent growth and 6.8 percent inflation. Public expenditure recovered to 19.5 percent o f GDP while revenues (including grants) reached 15.4 percent o f GDP, exceeding the pre- crisis 2001 rates. Growth slowed in 2004, though remained resilient despite the dampening effects of two major cyclones, a collapse inworld vanilla prices hnd severe exchange rate volatility. Moreover, sporadic violence has continued to cast a shadow -- Several grenade attacks occurred as recently as mid 2004 - though their motivation and significance are unclear and so far the reform process has not been derailed. -4- Box 1-1The 2002 crisis I A contested first round of elections in December 2001 plunged Madagascar into a political crisis. The election pittedthe incumbent president, Didier Ratsiraka, against Antananarivo's popular mayor, Marc Ravalomanana. The count indicated Mr. Ravalomanana inthe lead but short of an absolute majority. Mr.Ratsiraka's insistence on a second round of voting triggered mass protests and a general strike by Ravalomanana supporters. Mediation attempts failed and Mr. Ravalomanana remained in the capital while Mr. Ratsiraka withdrew to his home base in Tamatave. In the ensuing standoff bridges were destroyed and roads blocked, isolating large parts o f the highlands from the coastal areas. A second Central Bank was established inTamatave which resulted in the freezing o f Madagascar's assets abroad, a halt to foreign exchange trading and the closure o f the treasury bond market for several months. Violent clashes, though rare, claimed several hundred lives. Domestic and foreign support for Mr. Ravalomanana grew steadily and in mid 2002 he was proclaimed president while Mr.Ratsiraka fled to France. The economic impacts o f the crisis were widespread, especially inthe EPZ where many companies closed their doors and temporarily laid off an estimated 100,000 EPZ workers. Services were also affected as roadblocks, damaged bridges and fuel shortages virtually shut down the transport sector, and tourism fell by as much as 95 percent. Poverty rose sharply during the crisis, particularly in urban areas where both informal sector and low-income formal sector workers were affected by declines inproduction and plant closures. However, the effects were also felt in rural areas where agricultural prices fell and many farmers were unable to get their crops to market. I Source :IMF 2003 8. Recent outturns in the face of such political and economic turbulence bolster optimism about the medium term.. Especially encouraging i s the resilience of private investment which remains . at more than double the level o f the mid 1990s as a share of GDP, while the fiscal deficit has narrowed and inflation remains tame. Nevertheless, it would be premature to disregard the 2002 downturn and extrapolate what appears otherwise to be a rising trend in growth. First, the surge in 2003 partly represents a cyclical recovery from the downturn rather than sustainable trend growth. Second, the political, weather and market related shocks o f the past few years are indicative o f risks which Madagascar will continue to face. 9. ...and the commitment to reform seemsflrm. On the whole, steady progress achieved under two successive governments over the last decade lends substantial credibility to the commitment to reform. A series o f IMF ESAF and PRGF programs has been in place since the mid-1990s and successive reviews have awarded generally satisfactory marks. The government submitted an I-PRSP in 2000, a PRSP in2003 and reached the HIPC completionpoint inOctober 2004.4 2) POVERTY OUTCOMESAND CHALLENGES 10. Madagascar's long decline was accompanied by a dramatic increase in poverty, particularly during the 1980s (Table 1.2). Ina sample o f seventeen Sub-Saharan African countries with data available in 1995-2000, Madagascar had the highest head-count poverty rate at over 72 percent. Other poverty indicators tell the same story. Infant mortality, estimated at 86 deaths per 1000births in 2001, i s high and immunization rates are l o w with only 36 percent of children fully immunized. Access to safe drinking water, especially in rural areas, remains very low. Net primary school enrolment rates increased from 63 percent in 1997 to 70 percent in2002. A draft PRSP was completed in2001, but was extensively revisedby the new government, not least to reflect the impacts o f the 2002 crisis. -5- Table 1-2: Madagascar's poverty indicators (selected years 1993-2002) 1993 1997 1999 2001 2002 Poverty Indicators Incidence o fpoverty (national) (%) 70 73 71 70 81 Rural (%) 74.5 76 77 77 86 Urban(%) 50 63 52 44 62 Service Delivery Indicators Net primaryschool enrolments'(%) 48 63 64 67 70 Re etition rate in rural primary school, 11 grade(%) P n.a. n.a. 28 37 38 Pre-natal consultation, % 78 77 69 63 Infant mortality (out o f 1000) 97 99 86 86 Immunization rate6(%) 43 36 38 36 36 Electricity connections (% o f households) 9 12 14 15 19 Water connections'(% o f households) 17 19 22.5 21 27 Sanitation connections8(% o f households) 35 33 45 52 55 Source: E P M 1997, 1999, 2001. Poverty aggregates between 1999 and 2001 are not entirely identical and small variations with PRSP data exist. The 2002 poverty rates were simulated on the basis o f the 2001 household survey, assuming household welfare changes in line with aggregate sector growth rates (for agriculture, industry, construction & services). These are updated based on numbers from INSTAT in the First Progress Report. Simulations for the poverty rate for 2003 show a decline to 74 percent. 11. Poverty in Madagascar is especially prevalent in rural areas where 77 percent o f the population was poor in 2001 compared to 44 percent in urban areas. Moreover, though faster growth in the late 1990s made inroads into poverty and the nationwide headcount rate fell from 73 percent in 1997to 70 percent in2001, the gains were achieved almost entirely inurban areas while ruralpoverty, if anything, increased. Rural poverty is associated with lack of access to transportation (about 80 percent o f rural households have no reliable transport services and a third have no road access at all) and limited credit, public services and human capital. As a result, agricultural productivity i s low and there are few opportunities for wage employment to augment household incomes. This underscores the challenge o f extending the benefits o f faster growth to more remote areas. 3) THEROLEAND FINANCINGOFTHE STATE 12. A weak public sector performance has contributed to the failure of Madagascar's development strategy. Inefficient revenue generation, poor budget execution and a lack o f strategic priorities which led to deteriorating infrastructure and erosion o f human capital are partly responsible for Madagascar's long term decline. Both current and capital expenditure are comparatively low, averaging around 11percent and 7 percent o f GDP respectively, and if anything these ratios have been declining since the early 1990s (Table 1.3). However, since 2001, HIPC debt relief and some efforts in tax administration have expanded the fiscal envelope, allowing a rise in spending averaging around 1 percent o f GDP. 'Netprimary school enrolment rates include public and private schools. Net primary enrolment rate for 2001 refers to the school year September 2000-June2001. The figures are based on administrative data and are different from those calculated from the National Household Survey. 'Access Complete vaccination cycle; 1997:DHS survey; 1999 and 2001: EPM. to private or public pipe (individual or common). 8See above. -6- Table 1-3: Madagascar's fiscal indicators (percent of GDP) 1990-95 1996-00 2001 2002 2003 Total revenue and grants 12.8 14.5 14.0 10.2 15.4 Total revenue 9.8 10.4 10.1 8.0 10.3 - Tax revenue 7.9 10.0 9.7 7.7 10.0 Grants 3.0 4.0 3.9 2.2 5.1 Total expenditure 19.6 17.2 17.6 15.7 19.5 - Current expenditure 12.8 10.1 10.3 10.3 11.4 - -Capital o f which interest 4.0 3.O 2.0 2.2 2.2 expenditure and net lending 6.8 7.1 7.3 5.3 8.3 Overall balance Including grants -6.9 -2.7 -3.6 -5.5 -4.1 Excluding grants -9.9 -6.8 -7.5 -7.7 -9.2 Primarybalance 1.3 3.7 1.8 -0.2 1.1 Financing 6.9 4.0 4.4 6.2 4.8 Domestic 3.4 1.o 1.9 2.4 1.8 Foreign 3.5 2.9 2.1 3.6 2.8 Privatizationreceipts 0.0 0.1 0.3 0.2 0.2 ~ Source: WorldBank, Live Data Base; IMF, SthPRGF Review 13. Even compared to other low income countries,fiscalperformance has beenpoor. Table 1.4 compares period averages in 1994-2003 for a sample o f fiscal indicators for Madagascar (column 1) and for up to 55 low income countries with data available (column 2). The final column shows Madagascar's percentile in the sample. For both government consumption and current revenue excluding grants as shares o f GDP, Madagascar i s in the bottom fifth o f the distribution, achieving barely half the average for all low income countries. The best performance i s on public spending on education which i s two-thirds the group average or the 26" percentile. Trade taxes represent over 50 percent o f current revenue compared to an average o f 26.7 for all low income counties, which i s the 98thpercentile. -7- Table 1-4: Fiscal performance, 1994-2003 Madagascar Low income Percentile Government consumption (% o f GDP) 7.9 14.1 15 Current revenueexcluding grants (% o f GDP) 9.9 17.9 18 Public spending on education (% of GDP) 2.4 3.7 26 Trade taxes (% of current revenue) 50.2 26.7 98 Real per capita GDP at PPP rates 736.0 1244.0 17 Source: World Bank, WDI and Live Database. Notes: Columnheaded `Low income' shows arithmetic average for up to 55 low income countries with at least one observation over the period 1994-2003. Eachcountry's value is the averageof all availableobservations. 14. Expanding revenue collection is crucial, though in the near term the possibilities are limited. As in many other Sub Saharan countries, subsistence agriculture and a large informal sector hamper revenue mobilization. The tax base i s small and highly concentrated, with 8 percent o f customs taxes and 28 percent o f inland revenue respectively deriving from the five largest contributors to each category o f tax. VAT and import taxes account for 70 per cent o f the country's tax receipts and only 200 enterprises pay approximately 75 percent o f company taxes. Thus, despite a low overall tax burden, rapidly increasing revenue collection runs the risk o f imposing a significant burden on savings and the flow o f funds to investment. 15. Nevertheless, there i s considerable scope for increasing efficiency. LMF structural conditionalities through the years have consistently focused on tax administration and limitingevasion and the proliferation o f exemptions and partial adjustments. In recent years a large taxpayer unit has been introduced, though its operational efficiency needs to be raised. Progress has been made in implementing the ASYCUDA++ customs system. But customs administration continues to be severely hampered by overly bureaucratic procedures, slow processing, administrative delays and pervasive corruption.' 16. Moreover, tax policy has been driven by political pressures or ad hoc responses to temporary shocks. Both before and after the 2001 election, various measures were introduced to stimulate private investment, including (i) creating a free tax regime (the zone franche) to benefit qualified foreign investors and domestic firms that export all their production, (ii)exempting domestic firms from company taxes, and (iii)temporarily lifting taxes and duties on investment goods. Though such measures may have contributed to the resilience of investment spending through the crisis, they also entail a cost in terms o f foregone revenue, or `tax expenditures'. In 2001, the revenue loss associated with incentives to promote private investment is estimated at MGF248 billion, or 1.5 percent o f GDP. Ofthis cost, 0.6 percent o f GDP is associatedwith incentives enjoyedby firms inthe zone franche, with the remainder accruing to the enterprises des droits commun. 4) MADAGASCAR'S REDUCTION POVERTY STRATEGY 17. Madagascar's PRSP mapped out a program to reducepoverty through rapid and sustained growth. The strategy entailed three strategic axes: (i) improving governance; (ii) promoting broadbased growth; and (iii) providing human and material security. The macroeconomic and fiscal framework was ambitious. Noting that rapid growth was the "sine qua non" of poverty reduction, the PRSP sketched out three scenarios. The first, proposed as a base case, aimed at reducing poverty by half in ten years 'IMF(Mai 2003) "RChabilitation des Administrations Douanikres et Fiscales". -8- implying an average growth rate of 9.3 percent. The second scenario would achieve the millennium development goal o f halving poverty by 2015 with average growth o f 8.0 percent. Finally, the third scenario, the l o w case, would maintain growth at the 1997-2001 average o f 4.5 percent (Table 1.5). The poverty strategy i s based on a combined growth pattern to be achieved across several economic sectors that would help to spread income growth to rural areas and thus would reduce extreme poverty. Scenario Outcome Average Investment Headcount growth rate poverty rate in 2003-15 2015 Base Reduce poverty by half in 10 years 9.3 24.8 29.8 Intermediate Reduce poverty by half by 2015 (as 8.0 21.0 34.2 per MDGs) Low Sustained growth at 1997-2001 4.5 12.0 46.1 average rate 18. T h e Joint Staff Assessment found the overall strategic focus o f the PRSP to be clear, with coherent sector development plans built onwork program agreements with the ministries involved. Most importantly, a high degree o f ownership and commitment was apparent. At the same time, the JSA noted that only limited costing had taken place, that in many cases the analytical underpinnings were weak and that there would be a need to prioritize and strengthen capacity to implement the program successfully. The JSA also proposed further structural measures to promote growth, including major investments in public transport infrastructure; a commitment to public-private partnerships to improve the management o f public enterprises; reforms in tax and tariff policies and land laws; and better governance. 19. With regard to the macroeconomic framework, the JSA argued the growth outlook was too optimistic given limited government capacity, possible financing shortfalls, institutional constraints and a weak financial sector unable to support the demand for credit. Indeed, the high case calling for average growth o f 9.3 percent for a decade - which the PRSP advanced as a baseline expectation - would be virtually unprecedented in any country in the world. Since the PRSP's macroeconomic scenarios have a forecast horizon o f only 2006 it i s difficult to assess the feasibility o f longer term projections. Nevertheless, while agreeing that the proposed macroeconomic framework was basically sound, the JSA proposed that the growth target be lowered. 20. Since the PRSP, macroeconomicperformance in Madagascar has been robust. Though the JSA viewed the PRSP's macroeconomic outlook as overly optimistic, the surge in growth in 2003 was not fully anticipated and growth in 2005-06 i s also expected to remain strong. The outlook i s premised on significantly higher investment, rising to nearly 25 percent o f GDP from an average o f around 15 percent in the recent past (Table 1.6). Bothpublic and private investment are expected to register gains thanks to improvements in the business climate, including better infrastructure; use o f public-private partnerships to raise efficiency in ports, airports, railways and telecommunications; more transparency intax and customs administration; and improved governance. Highpotential sectors or "growth poles" have been identified and are being targeted with measures to accelerate growth, including export processing zones, tourism and mining. 21. Current expectations also anticipate a strong public sector performance, with revenue rising 10.3 percent o f GDP in 2003 to over 12.7 percent by 2006, while the overall deficit (including -9- grants) narrows from 4.2 percent to 3.8 percent (Table 1.6). Expenditure i s also expected to rise, in particular capital spending which increases from 7.8 percent o f GDP to 11.3 percent. Table 1-6: The PRGFMacroeconomic Framework (percent of GDP) 2003a/ 2004b/ 2005b' 2006b/ GDP Growth (%) 9.8 5.3 7.0 7.0 Investment (% o f GDP) 17.9 23.7 25.3 24.8 Public 7.8 11.5 11.3 11.3 Private 10.1 12.2 14.0 13.5 Balance of Payments Current Account -6.0 -9.2 -7.8 -6.3 International Reserve (months of imports) 2.8 3.0 3.2 3.4 CentralGovernmentAccounts Total Revenue and Grants 15.4 19.8 18.3 18.5 Total Revenue 10.3 12.0 12.4 12.7 Grants 5.1 7.9 5.9 5.8 Total Expenditure 19.5 23.1 22.4 22.1 Interest expenditure 2.2 3.0 2.4 2.1 Current non-interest expenditure 9.2 8.5 8.7 8.7 Capital expenditure 7.8 11.5 11.3 11.3 Overall balance (commitment basis) Excluding grants -9.3 -11.3 -10.0 -9.5 Including grants -4.2 -3.4 -4.1 -3.8 Financing (cash basis) 4.8 4.3 4.4 3.8 Foreign (net) 2.8 5.1 4.4 3.9 Memo item: Non-interest expenditures inFMGbil. 5,872 7,928 8,765 10,325 Domestic debt/GDP 13.9 10.4 8.5 7.5 Interest payments (domest) /tax revenues 21.4 25.0 19.4 16.5 Source: IMF 2004 (51hPRGF) Notes: R/ estimate; b/projection 22. A comparison between the PRSP and the PRGF framework suggest that Government has ample leeway to finance planned expenditures. Since the PRSP i s implemented through the Government's annual budget process, it i s important to ensure consistency between the macroeconomic and fiscal framework with the consensus outlook o f development partners, especially the IMF. In light of the strong outturn in 2003, the macroeconomic and fiscal outlook in the most recent PRGF review currently exceeds the projections o f the PRSP (Table 1.7). For instance, total non-interest expenditures (recurrent and capital) for the period 2004-2006 under the PRSP are approximately FMG 24.6 trillion compared to FMG27 trillion under the PRGF framework (innominal terms). Inreal terms, the margin would be even greater since inflation has been lower than the PRSP's target." Both the PRSP and loThough the PRSP does not forecast public consumption or investment deflators. -10- PRGF frameworks are designedaround fiscal deficit paths that maintainMadagascar's debt sustainable, and, together with the impact o f the HIPC Initiative, result in eating up a much smaller share o f government revenues than inthe 1990s. 23. Madagascar's revenue efforts in 2004 have paid ojJ Relative to the PRSP (which was drafted in 2003), table 1.7 suggests that Madagascar has made significant efforts to raise revenues from about 10 percent o f GDP in 2001 (pre-crisis reference year) to about 12 percent o f GDP in 2004. The strength o f these revenues underlies Madagascar's ability to meet PRSP spending objectives (for the period 2004-2006) while maintaining a more prudent deficit reduction path. For 2004, the improvement in revenues has been achieved through better customs administration, and a reduction in exemptions. 24. Critical role of revenues in thefuture. Earlier inthis chapter, it was noted that Madagascar has lagged behind other low income countries in its tax efforts. Table 1.7 suggests that Madagascar can maintain spending levels in accordance with the PRSP plans only through significant increases in revenue effort, particularly since the fiscal space achieved through debt relief has already been incorporated into the expenditure envelope (see Box 1.2) and it would be imprudentto forecast higher external grant availability for a sustained period. Table 1-7: Comparisonof PRSP andPRGFMacroeconomicandFiscalFrameworks 2001 2002 2003 2004 2005 2006 FromPRSP(July 2003) Real GDP growth 6.0 -12.7 6.0 7.0 7.5 8.0 CPI inflation 7.4 15.8 8.0 6.0 4.9 4.4 Total Revenue & Grants (as % o f GDP) 14.0 10.2 15.0 14.4 14.9 15.3 olw Revenue 10.1 8.0 10.6 11.5 12.0 12.5 Total Expenditure (as % of GDP) 17.6 15.7 18.2 20.2 20.2 19.9 olw Interest Due 2.0 2.2 1.8 1.3 1.2 1.1 olw Current (excl. interest) 8.1 7.1 8.4 8.6 8.5 8.3 olw Capital 7.3 4.8 7.3 10.2 10.4 10.5 Overall Balance (comm..) (as % of GDP) -4.3 -5.5 -3.3 -5.9 -5.3 -4.6 From5thPRGFMIPCCP (October 2004) 2001 2002 2003 2004 2005 2006 Real GDP growth 6.0 -12.7 9.8 5.3 7.0 7.0 CPI inflation 7.0 16.3 -1.7 10.1 7.6 5.0 Total Revenue & Grants (as % of GDP) 14.0 10.2 15.4 19.8 18.3 18.5 olw Revenue 10.1 8.0 10.3 12.0 12.4 12.7 Total Expenditure (as % of GDP) 17.6 15.7 19.5 23.1 22.4 22.1 olw Interest Due 2.0 2.2 2.2 3.0 2.4 2.1 oiw Current (excl. interest) 8.3 8.1 9.2 8.5 8.7 8.7 olw Capital 7.3 4.8 7.8 11.5 11.3 11.3 Overall Balance (comm.) (as % o f GDP) -4.3 -5.5 -4.2 -3.4 -4.1 -3.7 Source: Madamscar PRSP.IMF fhPRGF Review Note: Differenices between the PRSP and PRGF for 2002 likely reflect data revisions 25. The longer term outlook may be problematic. Though near term prospects are, "still consistent with PRSP objectives" (IMF, 5th PRGF Review) there i s reason for concern about the consistency o f longer term projections. The PRGF Review (September 2004) anticipates the cyclical rebound from the 2002 crisis will continue with growth o f 7.0 percent in2005-06, then slow to a more -11- sustainable pace o f 6.0 percent in the longer term. By contrast, the PRSP anticipates that growth will accelerate progressively from 6.0 percent in 2003 to 8.0 percent in 2006. The growing divergence in growth forecasts, in light o f the PRSP's longer term poverty reduction goals, suggests that Government should monitor macroeconomic and fiscal performance carefully. In the event Government's PRSP expectations turn out to be too optimistic in the later years, spending plans may need to be adjusted to maintain PRSP priorities (if not the time table) without jeopardizing fiscal sustainability. Such close monitoring would also give government the opportunity to review the recurrent-capital expenditure mix which i s necessary too support the new investment that are being made, and i s an issue that comes up in subsequent chapters as historical spending on sectors i s reviewed. 26. Onepotential threat which must be taken seriously is a build up ofpublic sector contingent liabilities. The full extent o f contingent liabilities o f the parastatal sector cannot be determined without regularizing accounting and management procedures - out o f 126 public establishments (EPIC and EPA) only 50 produce financial statements. Despite the self-financingstatus o fthe EPICmany continue to receive transfers from the state, as do a number o f EPA that are in fact no longer active. Out o f 69 state-owned companies, 33 are in a critical condition; only 10 produce financial statements and 36 are no longer operational. As o f 2001 repayment o f loans granted by the government to public enterprises o f MGF 309 billion remained unpaid. Debt o f SOLIMA to the Central Bank amounts to MGF 360 billion, for which the government i s assuming responsibility from 2004." I' Obligations in respect o f the state owned companies will be offset to some extent by receipts from the liquidation o f a number of state-owned companies. The liquidation o f 8 o f these companies have generated MGF 365 billion. -12- Box 1-2HIPC Debt Relief Madagascar reached the decision point under the HIPC initiative in 2000 and the completion point in 2004. Based o n reconciled debt data, outstanding external debt stood at US$4,843 million at the end o f 2003. At completion point, total debt relief from traditional mechanism, HIPC and bilateral relief, the N P V of external debt was lowered to US$1,467 million, equivalent to 137 percent o f exports. Under the D S A assumptions, the ratios o f both debt and debt service to exports decline progressively in the medium to long term and are expected to remain easily manageable (Table 1.8)' Debt relief frees up considerable resources for social and other spending. Interim HIPC relief from 2001-04 was equivalent to around 1percent o f GDP and was distributed across both current and capital budgets, more or less as envisaged by the decision point criteria. 62 percent o f the total was allocated to current spending, approximately half to education where it funded teacher recruitment, incentives for rural teachers, development o f cultural centers, and materials and supplies. Around 40 percent went to health for personnel recruitment, immunization and equipment purchases. In addition, funding was made available for various institutional reinforcement activities as well as environmental and social programs. The allocation to capital projects included road maintenance, water, safety net support, electrification and monitoring (including a household survey). Madagascar's debt indicators 2003 2004 2005 2010 2020 After traditional debt relief 3,429 3,158 3,268 3,605 3,837 After enhanced HIPC relief 3,371 2,248 2,388 2,862 3,334 After additional bilateral relief 3,345 1,618 1,776 2,328 2,927 Debt service after full implementation 50 55 104 217 Memo items: NPV debt-to-exports (%) 136.6 153.7 135.5 69.1 30.8 Debt service to exports (%) 4.8 4.2 5.0 5.5 Debt service to revenue ("h) 10.0 10.6 12.1 10.8 5) FINANCIALSECTOR 27. An efflcientFnancia1 sectorplays an important role in allocating risk and investment and stimulating growth. However, financial markets in Madagascar are in a rudimentary state. The Malagasy banking system i s highly concentrated and uncompetitive, hampered by poor contract enforcement and lacking in modern risk management techniques. As a result, commercial lending is limited. Private sector credit extension averages 9 per cent o f GDP over the period 1998 to 2002, less than half o f the average for African countries for which data are available. Of the estimated 12,500 firms in the Malagasy economy, less than 200 make use of credit facilities from the banks. What credit i s available can be accessed only at a very high price, with real long term (i.e. two years) interest rates averaging over 10 percent in the last five years.12 As a result, the private sector relies primarily on retainedearnings and private savings to fundinvestment. l2 Longer term credit i s largely unavailable. Borrowing is thus primarily undertaken to fundworking capital -13- 28. Notably, the low level o f credit extended to the private sector does not appear to result from the government's monopolization o f a limited savings pool - if anything, financial markets are characterized by excess liquidity. For instance, at the end o f 2003 banks held MGF 305,9 billion in excess reserves with the monetary authorities, approximately 1per cent o f GDP. Such a result suggests either the financial sector i s unwilling to take risks or private borrowers see few opportunities for productive investment. 6) RECOMMENDATIONS 29. This chapter has reviewed Madagascar's recent development experience and the country's PRSP as a strategy to reversing the decline in per capita incomes. In order to achieve the country's PRSP objectives, Madagascar may wish to: o Carry out a one-time review of the medium term projections of the PRSP to ensure realistic output and revenueforecasts. The aim should be to provide a sound medium term expenditure envelope and establish priorities consistent with fiscal sustainability. This will not only allow Government to sequence investments better and assure adequate recurrent financing, but facilitate the annual budget planning process. It would also help government manage the donor community more effectively, which i s an issue discussed insubsequent chapters. Annually review the macro framework under the PRSP at the time of budget preparation. This exercise would allow Government to align the PRSP implementation with its rolling medium term expenditure framework which underlies the annual exercise o fbudget preparation and the PRGF fkamework. Maintain a steady and sustainable program of revenue enhancement so as to moderate any negative impacts on incentives for savings and investment, and private sector growth. o Develop/enforce a simplesystem to monitor contingent liabilities more closely.A first step i s to require public enterprises to present audits to Government, or to parliament, as part o f the budget preparation process. This i s critical as fiscal space gained through debt relief or higher revenues can easily be wiped out and PRSP implementation reversed, by adverse events that trigger unmonitored fiscal risks. A second step could be to develop a matrix o f fiscal risks which includes an assessment o f explicit and implicit contingent liabilities which can have a direct and indirect impact on public finances. The matrix should be monitored and updated at the time when the macroeconomic and fiscal frameworks are reviewed. -14- 11. THE CHALLENGE OFRESOURCEALLOCATION AND BUDGET EXECUTION 30. This chapter intends to provide conceptual and technical guidance for the public finance reform agenda. It will look at the appropriateness o f the resource allocation and the budget management processes. Furthermore, it will deal with the overall execution o f the budget given its importance for the implementation o f Government policies and the country's effective use o f human resources. 31. Recent public finance reforms have yielded some encouraging results: First, relatively good technical capacity in the Ministry o f Finance has been developed, although this capacity has not yet been extended to the sector ministries. Second, Madagascar has a relatively modem institutional and legislative framework for its public finance system--the old basic public finance law ("loi 0rganique")-- which was derived from the French system has been updated. Third, in 2000 Madagascar introduced a uniform and modem Chart of Accounts that facilitates classification as well as budget monitoring and evaluation. Fourth, Madagascar has developed, with its own resources, a computerized financial management system which integrates the different functions o f the budget planning and execution process. This system has been piloted in Tamatave and will soon be deployed to other regional treasuries. The experiences, which were obtained during the development and implementation o f this pilot system are valuable for the intended further modernization o f the public finance system in Madagascar. 32. Additional work, however, remains to be done to bring the public. finance system up to international standards. Areas o f concern are (i) weak capacity o f the line ministries in public financial management, (ii) lack o f transparency mainly caused by weaknesses in the financial reporting system, (iii)ineffective control ofpublic expenditures, and (iv) cumbersome procedures regardingbudget execution which in turn have adverse implications for service delivery. The weaknesses o f the Malagasy system are documented in various reports and studies, in particular the Country Financial Accountability Assessment (CFAA)'3, the Country Public Procurement Assessment Report (CPAR)'4, both completed in 2003, an independent audit o f the "chaines de dkpenses" o f the year 2000J5, and the reviews o f the Ministry o f Finance and o f the Vice Prime Minister's Office". A joint Bank-Fund HIPC evaluation mission found that Madagascar meets only four out o f 15 criteria which were developed to benchmark public finance systems in developing countries (see annex 3)." However, the mission pointed out that some reforms that are currently implemented ,e.g. the draft public procurement code, and the strengthening o f internal control institutions and procedures, would help correct existing weaknesses over the short term. 33. To address the shortcomings o f the system o f public finance the Government developed in collaboration with the donor community a "2004 Prioritized Action Plan" (see box 11.1 below and annex 4). Additionally, a coordination committee (Cellule de Suivi et de Coordination des RCformes l3World Bank (Juin2003), RCpublique de Madagascar- Etude de la Gestion des Finances Publiques (Country FinancialAccountablity Assessment). l4World Bank Report (December 2002), No. 26410, Vol. I& I1 - Country Public Procurement Assessment Report. l5Group 2AC (Mai- Juin 2000) "Audit des Fonds de Contrepartie de l'Appui a 1'Ajustement Structure1dans la Republiquede Madagascar" l6Bloch-Laink,B de la Biche JP Landau (Octobre 1999) "Audit du Ministere des Finances et du Ministbre du Budget, de la DCcentralisationet duDeveloppementdes ProvincesAutonomes" 17 World Bank (2004) "Memorandum and Recommendation of the President o f the IDA to the Executive Directors onAssistance to the Republic ofMadagascarunder the EnhancedHIPCDebtInitiative". -15- des Finances Publiques) was established in the MEFB to coordinate, monitor and evaluate the reform agenda. Box 11-12004 PriorityAction Plan A government-led reformprocess - In2003, the government decidedto develop an actionplanfor public finance managementreforms basedon assessments undertaken by the donor community (EU financial audit, CFAA, CPAR, IMF technical assistance reviews, etc.) and the Malagasy administration (CRROC report). The objective o f the action plan was to review the recommendations made by the different assessments and to consolidate them into a sequenced and prioritized reform program for 2004. The short-term timeframe o f one year was chosen to focus on a number o f selected activities and o n early demonstration o f results. The donor community agreed to provide technical support inline with the priorities determined by the government. The mainreforms activities o f the Action Plan are the following: 1 Establish institutional arrangements to support reform implementation, including a Reform Committee at MEFBlevel and a coordination mechanism with the donors. a Address cross-cutting dimensions of reforms by developing a training strategy for the MEFB and continuing the extension o f the Integrated Financial Management System; I Consolidate the budget preparation calendar and sequence the introduction o f program budgeting; 1 Streamline budget execution procedures, develop supporting guidelines and finalize the legal and regulatory framework for procurement. 1 Strengthen internal and external control institutions interms o f staffing, equipment and training. 1 Improve timeliness o f reporting by addressing the backlog in the consolidation o f national accounts and preparation o f the budget execution law (`Zoi de R2glement"). 1 Reinforce tax and custom administrations by finalizing regulations, formulating guidelines and increasing computerization. A review o f progress o f the 2004 Action Plan took place in October 2004 and showed that the majority o f the activities had already been implemented. The government i s in the process o f developing a Prioritized Action Plan for 2005. 1) ALLOCATINGPUBLICRESOURCESI NLINEWITH GOVERNMENT PRIORITIES 34. This chapter seeks to shed light on how Government resources are allocated inMadagascar, both across sectors and across different levels o f Government. It pays particular attention to the post- crisis period, 2003-2004 during which the PRSP has been implemented and looks at the degree o f prioritization o f public expenditure which the Government has been able to achieve. (a) PastAllocations: Issuesand Challenges 35. As the PRSP is the key policy document for the Government the following budgetary analysis i s based on the structure o f the PRSP which has three strategic axes: governance, growth and social sectors. The intention i s to review the budgetary allocations over time against the goals and objectives o f the PRSP. SectoralAllocation 36. Past allocations show shifting priorities. Table 11.1 below classifies Government expenditures over the 1997-2004 period into four categories: governance, growth, social sectors, and other, non-classified expenditures. It shows that between 1997 and 2004 substantial variations in -16- allocations are observed across all sectors suggesting shifting priorities o f Government within the period, Some general trends include": General administration, the main sub-sector within the governance category, takes the lion's share o f the budget, on average 36 percent over the observed period. Since 1999, the year with the highest share of budgetary allocation to general administration (43 percent), efforts have been undertaken to cut down public spending inthis area. Despite a slight increase in allocations to general administration in 2003 by 6 percent, the 2004 budget reflects a further downward trend inthe share o f the budget. Government's efforts to strengthen the social sectors are reflected in increasing allocations to primary education (average annual increase o f 49 percent for the period 1997-2002) and to primary health care (average annual increase o f 20 percent for the same period). Since the PRSP implementation began in 2003, the education sector benefited by a 41 percent growth rate in allocations, reaching the highest share o f budgetary allocations (around 18 percent) in 2004. However, public resources to the health sector only rose by 2.3 percent over the 2003-2004 period. P Allocations to the growth sectors (notably transportation, agriculture and environment) have not increased at the same rate. The overall substantial decrease o f budget allocations to the agriculture sector until the year 2002 can be explained by Government's decision to withdraw from the productive sectors and to refocus Government activities in agriculture on key policy issues.'' In contrast, since 1999 the share o f the transportation budget (8 percent) increased to 13 percent o f total budget allocations in 2001. The 2003 budget shows a cut across nearly all growth sectors (in real terms and as a share o f the national budget) while allocations to governance and social sectors were increased. In 2004 strong prioritization took place notably with a reinforcement o f the transport budget andpartially the agriculture sector. o It is noteworthy that in the past defense allocations remained reasonable, averaging slightly above 3 percent o f total expenditure between 1997 and 2002. It increased, though, after the 2002 crisis, following political unrest, to restore state authority throughout the country. "Seealsoannex 5, table 5.1 for referenceregarding therealannualgrowthrates insectoralbudget allocations ''1997-2004. Such measures included the termination o f the extension ("vulgarization") program in the agriculture sector since 1998. (This program aimed to provide technical assistance to farmers mainly in the highland. It was supported by many donors including the World Bank, IFAD and the European Union but was closed end o f the 90's.) -17- Table 11-1:Madagascar's budget allocation trends selected years between 1997-2004 (Percent of noninterestexpenditures) 1997 1999 2001 2002 2003 2004 Total 100 100 100 100 100 100 Governance 39.7 50.6 43.9 36.4 44.6 42.1 General Administration 32.1 43.2 37.9 30.1 35.5 32.8 olw budget, finance & eco. admin 19.7 31.3 25.1 19.0 26.0 24.5 Defense 3.4 3.1 2.4 2.6 3.9 4.1 Public security 4.1 4.2 3.5 3.7 5.2 5.2 olw Justice 0.5 0.9 0.7 0.7 0.9 0.9 Growth 32.8 23.5 28.2 34.7 25.3 26.4 Energy 3.2 2.1 1.7 1.4 1.5 1.3 Agriculture 10.4 7.0 5.8 6.1 4.9 5.2 Environment 3.1 2.0 2.5 3.3 2.4 2.3 Transport & Communication 11.9 7.8 13.3 13.4 11.2 12.9 olw Transport 10.9 7.5 13.0 13.2 10.8 12.6 Others' 4.4 4.5 4.9 10.6 5.3 4.8 Social 27.5 25.9 27.9 28.8 30.1 31.5 Education 10.5 11.7 12.5 14.4 15.2 18.4 olw Primary Education 2.1 3.2 2.3 5.1 6.4 6.1 Health 8.8 7.8 9.0 8.5 8.4 7.4 oiw Primary Health care 4.4 3.4 2.8 3.7 2.9 3.4 Others2 8.2 6.4 -. . 6.5 6.0 6.5 5.7 Source: MEFB and World Bank estimates 'Includes programs related to industry and other economic affairs Includes programs related to social protection, urbanism and culture Economic allocation 37. Government increased allocationsfor recurrent expenditures. Recurrent expenditures as a share o f the total budget increased by 20 percent over the period 1997-2002, compared to investment expenditures, whose share increased by 17 percent of the total budget over the same period (see table 11.2).2o Over the 2003-2004 period, the share o f recurrent expenditures increased by 19 percent, while the share o f investment expenditures increased by only 10 percent. Madagascar has also experienced significant variations o f foreign assistance over the 1997-2004 period, as indicated in table 11.2. Externally financed investments amounted to about 34 percent o f allocations in 1997 compared to only 24 percent in 2003. These fluctuations negatively impact budget planning and programming, Nearly 60 percent of capital expenditures were financed by donors over the period 1997-2004. 2o See also annex 5, table 5.2 for reference regarding the real annual growth rates o f current and capital budget allocations 1997-2004. -18- Table 11-2: Madagascar's recurrent and investmentallocationpatternsselected years between1997-2004 (Percentof noninterestexpenditure) 1997 1999 2001 2002 2003 2004 Total 100 100 100 100 100 100 CurrentExpenditure 43.5 48.8 48.4 47.6 54.7 56.5 Salary 20.9 20.2 18.6 21.3 27.3 26.5 Goods & services 12.7 20.1 10.8 9.7 14.1 13.3 Transfer 10.0 8.5 19.0 16.6 13.3 16.7 CapitalExpenditure 56.5 51.2 51.6 52.4 45.3 43.5 Internal resources 22.2 23.8 25.0 21.3 21.6 18.3 External resources 34.3 27.4 26.6 31.1 23.7 25.2 Memo External resources (as a % o f total capital expend.) I 60.7 53.5 51.6 59.3 52.3 58.0 Source: MEFB and World Bank estimates ResourceAllocation to service delivery levels" 38. Budget allocations favor the central Government administration over service delivery. Madagascar's public administration i s highly centralized. Local Governments do not receive sufficient resources to effectively respond to local service needs. Communes, currently the only operational administrative level below the central Government, account for only a small portion o f public expenditures. A recent report from the World Bank on decentralization found that only 3 to 4 percent o f the national budget i s managed by the communes.22 39. Madagascar currently allocates about 15 percent o f budget resources to service delivery levels (see chart II.l.a).23 At the same time, only 40 percent o f civil servants are located in outposts (see chart 1I.l.b). In contrast, in 2001, around 77 percent o f the population and the poor lived in the rural areas.24 21The service delivery level i s defined as the services provided by central Government. It includes deconcentrated services provided at the regional and district level (e.g. the health service at the district (SSD), CISCO) or services directly provided to the population often by public outpost or other local Government facilities. 22The communes have own revenues such as real estate taxes, market fees or fees for administrative work; see World Bank (March 2004), Madagascar Decentralization for W h e r details. 23This represents, however, a slight improvement compared to earlier years in which the total amount of public expenditures managed at the deconcentratedlevels was as low as 9 percent (1997). 24E P M 2002 -19- Chart 11-1a-c: Deconcentration of public services - Allocation and managementof funds and staffing- - a: Percentage of budget allocated to outposts b: Percentage of civil servants who work in outposts 70 60 50 40 1 30 20 10 0 1997 1999 2001 2002 2004 1997 1999 2001 2002 2004 +SgxlR +AAGW Source: WorldBank c: Percentage of budget managed by outposts 1997 1999 2001 2002 2004 Source: WorldBank -20- 40. Level of deconcentration. The social sectors are the most advanced in terms o f allocating resources to the service delivery level. In2004, the Ministries o f Education and Health had allocated 60 percent and 35 percent, respectively, o f their total budgetary resources to deconcentrated line agencies. In2004, both ministryemployed 90 percent and 60 percent, respectively, of total staffinfacilities and administration at the sub-national level. In contrast to the high degree o f deconcentration o f the Education and Health ministries, the Ministry o f Agriculture allocated only 20 percent o f its budget to local service agencies, which i s rather modest given the importance o f the agriculture sector for local development.25 41. Budgetary control of outposts concentrated at the central level. While deconcentrated line agencies are responsible for service delivery they lack the authority to manage their own budget (see chart 11.1.c.) With the exception o f the social sectors, o f the total central Government budget allocated for rural service delivery, only 10 percent i s managed by deconcentrated agencies. They accounted in 2001 for more than 90 percent o f the funds executed outside the central Government. The majority o f the management decisions such as investment planning and execution or the hiring and firing o f civil servants are handledby the center. 42. Conclusion - allocations to service delivery levels insufficient. This analysis suggests that the allocation o fresources to the service delivery levels i s insufficient, and therefore not inline with the intention o f the PRSP to bringGovernment "closer to the people". Sustainabilityof the PRSP 43. Government has recently prioritized expenditures towards education, transportation, and general administration. The new Government has placed a major emphasis on education, transportation and on general administration while maintaining fiscal restraint in the total expenditure envelope. Whereas total budgetary allocations increased by an average annual rate o f 16 percent between 2003 and 2004, allocations to education and transportation increased by 40 and 34 percent respectively between 2003 and 2004 (see Annex 5 - table 5.3). Though budgetary allocations to general administration increased only slightly in 2004, it remains the major beneficiary o f public resources in 2004 (general administrationtook up 33 percent o f total budget in2004). 44* Spending on general administration seems to be relatively high. Tables 11.3 and 11.4 illustrate the distribution o f budgetary allocations and actual expenditures across different categories of the general administration. The MEFBtakes the lion share o f total administration expenditures with 75 percent, or 25 percent o f the total non-interest expenditures o f the budget in 2004.26These resources finance the preparation and management o f the budget as well as transfers to the communes and collectivities (Autonomous Provinces and ex Fi~ondronana).~'Overall, it remains unclear how these resources contribute to the goals and objectives o f the PRSP. Inaddition, implementation performance i s weak (86 percent in 2003), in particular if compared to other spending lines within public administration(see table 2.4) which raises questions about administrative efficiency. 25 Staffing o f outposts, however, seem to be fairly developed with 61 percent o f staff working at the local facilities (see chart 11.1b). 26 From 2003 onwards, the line "Budget, financial and economic administration" includes transfers to communities (under MEFB) which previously were under "Administration o f the territory and the collectives". However, the Budget, financial and economic administration expenditures are high even if we abstract from the tranfers to the communities. 27 The transfers to the communes are two-fold: i)general transfer o f recurrent expenditures to communities (since 1996) and ii)specific transfers health and education transfers (since 2000). -21- Table 11-3: Distribution of public allocations within the general administration (inpercent) Annual Annual aver. aver. 1997 1999 2001 2002 1997-02 2003 2004 1997-04 1997-04 Total 100 100 100 100 100 100 100 100 Common general service of public administration 0.8 0.5 0.9 1.6 0.9 1.2 2.4 1.8 1.2 Executiveand legislativebody 4.5 3.3 10.2 9.1 7.5 7.0 7.1 7.0 7.4 Budget, jinancial and economic administration 61.4 72.4 66.1 63.3 63.2 73.3 74.6 74.0 65.9 Administration of the territory and the collectives 21.6 14.0 13.3 14.3 17.6 6.9 3.5 5.2 14.5 Administration of foreign affairs 4.8 5.1 5.3 5.9 5.4 6.0 6.0 6.0 5.6 Inspectionand others 0.8 0.8 1.0 1.2 1.0 1.1 1.0 1.0 1.0 Research 6.0 3.7 3.0 4.4 4.2 4.4 4.8 4.6 4.3 Other Administration CA 0.1 0.1 0.1 0.2 0.I 0.2 0.5 0.3 0.2 Memo Transfers to communes 5.8 0.8 3.7 8.7 4.I 7.1 6.7 6.9 4.8 Source:MFEB, WorldBank I Table 11-4: Execution of actual public expenditures across different administrative categories (inpercent) Annual Annual avg. 1997 1999 2001 97-01 2002 2003 97-03 Total 74.3 56.7 77.7 74.4 51.3 86.8 76.4 Commongeneral service ofpublic administration 69.9 71.4 105.4 103.1 57.6 67.3 97.1 Executiveandlegislativebody 94.7 83.1 61.0 87.1 52.8 93.1 88.I Budget, jinancial and economic administration 61.2 43.1 74.6 65.7 47.5 84.3 68.8 Administration o fthe territory and the collectives 97.9 95.9 88.1 96.0 42.5 94.7 95.8 Administration o f foreign affairs 77.1 83.0 107.6 68.8 87.2 99.5 73.9 Inspectionand others 95.0 87.5 82.1 85.3 54.5 100.6 87.8 Research 103.1 104.2 95.2 101.6 78.8 91.0 99.8 Other Administration CA 105.4 96.2 85.8 110.6 61.2 101.5 109.I Memo Transfers to communes 99.4 94.6 93.9 96.9 13.2 100.0 I 97.4 Source: MFEB, WorldBank 45. Facing significant trade ojJs in PRSP implementation. Previous analysis has shown that the country's macro framework, as outlined in the PRSP, is on track (see chapter 1). As an implication, however, Madagascar has had to prioritize its expenditures especially over the past two years. Govemment reduced expenditures, in particular in the growth sector, and increased expenditures especially in primary education, general administration and defense in 2003. In 2004, additional resources for education, transport, and general administration were financed by reducing allocations to -22- other sectors (agriculture, health and environment). These expenditure trade offs demonstrate the constraints o f the fiscal envelope inlight o f the PRSP agenda. @) Comparativecontext 46. This section analyzes Madagascar's allocation of public expenditures in the Sub-Saharan context. For comparative reasons it uses standard functional and economic classification. Additional (and more detailed) tables that further substantiate the findings of this section can be found in annex 5. 47. Total budgetary allocations in Madagascar amount to 19 percent o f GDP in 2004; this means that the volume o f public expenditure i s below average Sub-Saharan African (SSA) standards o f about 26 percent o f GDP (see annex 5, table 5.5). Expenditure characteristics are similar to other countries in the region: the evolution o f expenditures over the last seven years confirms that the most important elements o f the budget are allocations to the general administrationand to the social sectors; at the same time expenditures in the productive sectors declined (see table 11.5). In 2004, 10 percent o f GDP was allocated to general administrative services; these allocations appear to be fairly high.They have seen an increase o f more than 15 percent between 1997 and 2004, and represent nearly 50 percent o f the budget. On the other hand defense spending i s significantly below SSA average (3 percent o f GDP) with only 0.6 percent o f GDP in 2004. Social sector allocations have increasedby 16 percent between 1997 and 2004: the education budget (3.1 percent o f GDP in 2004) however, is below SSA standards (3.4 percent o f GDP in2004), funding for the health sector reaches 1.2percent o f GDP in2004 which i s below the SSA average o f 2.6 percent o f GDP. Allocations for the productive sectors have decreased from 5 percent o f GDP in 1997 to 3.5 percent in 2004 in line with Government policy to increase the role o fthe private sector. Only 0.8 percent o f GDP i s presently allocated to the agriculture sector which seems fairly low inview o fthe fact that poverty i s prevalentparticularly inrural areas. Table 11-5: Budgetary allocations 1997-2004. (Percent o f GDP) 1997 1999 2001 2002 2003 2004 GeneralAdministration 7.3 10.6 10.3 9.6 8.8 10.1 Administration 6.0 9.1 8.9 8.0 7.1 8.4 Defence 0.6 0.6 0.5 0.6 0.7 0.6 Public Security 0.7 0.8 0.8 0.9 1.o 1.o Economicservices 5.0 4.2 5.8 6.9 4.2 3.3 Energy 0.4 0.5 0.6 0.5 0.4 0.2 Agriculture 1.7 1.4 1.2 1.1 0.8 0.8 Environment 0.5 0.4 0.5 0.7 0.4 0.0 Industry 0.4 0.4 1.1 2.6 1.1 1.o Transport 1.7 1.3 2.2 1.8 1.3 1.o Other economicservices 1.2 1.5 1.4 1.3 0.8 0.62 Social Services 4.7 5.0 6.1 6.9 5.7 5.4 Education 1.8 2.4 2.8 3.4 2.9 3.1 Health 1.4 1.4 1.7 1.9 1.4 1.2 Other social services 1.4 1.1 1.6 1.5 1.3 0.9 Source: MFEB, WorldBank -23- 48. Similar to other countries in Africa, Madagascar i s under funding its non-salary recurrent budget. Although this part o f the budget has seen an annual increase o f 17 percent between 1997 and 2004 (see table 11.6) it represents only 7 percent o f total expenditure which is equivalent to 1.6 percent o f GDP. Salaries account for only 5 percent o f GDP which i s very low compared to 6-10 percent in many countries in SSA. Salaries constitute only 25 percent o f the total budget (against an average o f 40 percent in SSA). The investment budget i s financed to a large part by contributions from external sources (around 60 percent). Investments represent 8 percent o f GDP in 2004 (a decline of more than 2 percent since 1997). Debt service stands at 2.2 percent o f GDP in 2004 which i s in line with SSA standards. Table 11-6: Structure of the Malagasy Budget (Percent o f GDP) 1997 1999 2001 2002 2003 2004 Current Expenditure 7.8 10.2 11.2 11.6 10.6 10.9 Salary 3.7 4.2 4.3 5.2 5.3 5.1 Goods & services 1.1 1.5 1.9 1.9 1.8 1.5 Transfer 1.8 1.7 4.4 4.0 2.5 3.2 CapitalExpenditure 10.2 10.7 12.0 12.8 8.8 8.4 Internal resources 3.8 4.9 5.8 5.2 4.2 3.5 External resources 6.3 5.8 6.2 7.6 4.6 4.8 Debt service 2.0 1.8 1.1 1.9 1.7 2.1 Source: MFEB, World Bank (c) Introducinga medium-termperspectiveto linkpolicy prioritieswith the budget Issuesand CurrentPractices 49. Government's efforts to align the budget with PRSP policy priorities have focused on the introduction of a medium-term perspective o f the budget. These efforts are at an early stage, and could lay the foundation for the implementation of a mediumterm expenditure framework. 50. Government has recently introduced a medium termperspective to budgetplanning. Until 2003, macroeconomic and fiscal parameters were defined on an annual basis. While line ministries developed public expenditures programs - linking objectives, activities, resources and expected outputs over a three-year period - the absence o f a medium-term sectoral resource envelope and a coherent expenditure framework as well as low capacity in line ministries limited the impact of these programs on budgetary outcomes. The absence o f a medium-term perspective had a number o f consequences, including (i)short-term allocation decisions, (ii) information about the affordability o f existing and little planned policies over the medium term; and, (iii) misalignment o f investment and recurrent expenditures (recurrent costs were not budgeted in the medium-term, but integrated into project investment costs). 51. I n 2003, the Government introduced program budgeting. Three-year macroeconomic and fiscal parameters were developed for the first time by the Ministry o f Economy, Finance and Budget an initial stage - additional work will be needed to improve quality of fiscal forecasts, multi-year (MEFB) and were used by line ministries to develop "program budgets" (see below). The process is at -24- planning for sectoral expenditures, and costing o f policy initiatives, as well as to enhance the coherence between investment and recurrent expenditures. Box 11-2 Program budgeting Program budgeting has been introduced into the public finance system with the new Organic Finance Law in 2004. Detailed line-item budget will be replaced by a budget classification system that divides expenditure into missions (cross-cutting broad objectives ) and programs. Each program consists of a clear set o f objectives, related activities necessary to achieve those objectives and expected results. Expenditure are classified by interest on public debt, salaries, recurrent expenditure, structural recurrent expenditure, investment expenditure, exceptional recurrent expenditure, and financial operation. Increased flexibility i s being given to program managers to shift expenditure across these broad categories, with the exception o f salaries, as well as to carry over expenditure to the next budget year on the basis o f a joint decree by MFEB and the concerned ministry (again with the exception of salaries). Source: WorldBank 52. The introduction o f program budgeting will be phased. The budget for 2005 has been prepared on a line-item basis and includes a program classification for information. As o f 2006, the annual budget will be exclusively structured along a program classification. 53. The budget preparation 2005 has shown that the introduction o f program budgets poses significant challenges to the public administration; the submission o f the 2005 budget to Parliament was delayed because o f problems with the program budgets at the line ministry level. The complexity o f the new budget methodology and the implementation problems put addition strain on the limited capacity, in particular in line ministries. The one-year transition period from a line-item to a program budget appears particularly risky, inview o f the experiences inother countries.28 Recommendations 54. The efforts made by the Government to develop a medium-term macro-economic and fiscal perspective and to introduce program budgeting should be further refined. For program budgeting, the approach o f the Government to use a broad rather than specific program classification i s appropriate in light o f the limited capacities o f the line ministries to operationalize the new approach. The following recommendations take into account Government's strategy o f introducing program budgeting for all the ministries within a one-year transition period: The Government should give high priority to strengthen the overall medium-term perspective in the budgeting process with the objective o f developing more reliable sectoral and cross-sectoral resources envelopes. This would allow line ministries to plan and manage resources more effectively. o The introduction o f sectoral MTEF by line ministries should be discouraged. Sectoral MTEF are unlikely to contribute to a more effective overall budget allocation process inlinewith Governmentpriorities. Suchaninitiative also bears theriskto divert scarce resources at the level o f line ministries at a time where Government i s implementing significant public finance reforms. In order to facilitate the introduction of program budgeting the provisions o f the Organic Financial L a w concerning program budgeting should be specified as soon as ** see L e HouerodTaliercio, MTEF inAfrica -25- possible by regulations and technical guidance, especially regarding (i) definition o f the a program to ensure a consistent program classification across sectors and institutions, (ii) contentoftheprogram,and(iii) criteriatobeusedfortheclassificationof the the various categories o f expenditures within a program. o The Government should concentrate on enhancing the capacity o f line ministries inthe area o f public finance to enable the reforms to gain momentum. The decision o f the Government to strengthen relevant capacity in key ministries (education, health, transport) on a pilot-basis points into the right direction but mustbe complementedby a systematic program o f training and capacity building. MEFB should also organize a workshop with line ministries to draw lessons from the introduction o f program budgeting for subsequent budgetyears. o The introduction o f program budgets should go hand-in-hand with an improvement o f the performance orientation o f the line ministries and spending agencies. To this end, Government should systematically define results and introduce service standards. The business plans that have recently been introduced are an important first step; it is, however, important to ensure that these plans are refined and fully costed as well as regularly monitored and evaluated. (d) Coordinationof andParticipationinThe BudgetPreparationProcess 55. Effective participation o f relevant stakeholders and predictability o f the budget process are key elements to ensure that fiscal and sectoral priorities are inline with overall Government objectives. Issuesand CurrentPractices 56. Reliable macroeconomic and fiscal benchmarks are lacking. Among the most important weaknesses o f the budget process i s the lack o f reliable macroeconomic and fiscal parameters guiding the budget preparation process. Since 1996 the Government has tried to improve macroeconomic forecasting and the development o f a macro framework that would serve as the basis for determining the budget envelope in any given fiscal year. An important organizational reform was the transformation o f the Secretariat for Forecasting and Modeling (Secretariat Permanent pour la Prevision et la Modelisation), initially charged with the task o f developing the macroeconomic framework, into the Macro-economic Directorate (Direction GCnCrale de 1'Economie) o f the Ministry o f Finance. At the beginning o f every budget year the framework i s developed by this directorate with input from the Central Bank, the Statistics Office (INSTAT), the Customs Directorate (Direction GCnCrale des Douanes), the Tax Directorate (Direction GenQale des Impots), the Treasury (Direction GCnCrale du Tremor) and the Directorate for Commitments (Direction GCnCrale des DCpenses EngagCes). Despite these efforts the quality o f the macro-economic framework continues to be low. 57. The budget calendar varies on a year to year basis. InMadagascar, the budget formulation process does not allow for adequate strategic decision-making over policy priorities. The calendar and the stages o f the budget preparation process vary from year to year and are communicated to the relevant stakeholders on short notice. Table 11.7 below sets out the budget timetable that was followed in2003, alongside with aproposed adjustedbudgettimetable. 58. The main shortcomings o f the consultation process include: 0 Within government: Intergovernmentalconsultation concerning the budget proposal i s mainly limited to budgetary conferences. There i s no systematic mechanism to ensure that budgetary disputes are discussed at different hierarchical levels (Directors, Permanent Secretaries, Ministers) o f the -26- administration prior to the submission o f the final draft o f the budget to the Cabinet. Open issues are resolvedby the MEFB at the technical level or raised through ad-hoc interventionat the Cabinet level. 0 Line ministries: Although the MEFB has made some efforts to better associate line ministries inthe budget preparationprocess, participation of line ministries is not yet effective, and interactions between the MEFB and the line ministries do not sufficiently focus on determining the budgetary impact o f policy priorities. Line ministriestypically do not receive any information on their expenditure ceilings before the middle o f the year. They have only limited time to submit their budget requests, once the budget circular i s issued (in 2003, two weeks). Discussions between the MEFB and the line ministries remain largely focused on the baseline budget and the proposed modifications ("mesures acquises" versus "mesures nouvelles"), which does not encourage a focus on policy priorities and on the effectiveness o f expenditures within a coherent fiscal framework. 0 Cabinet: Strategic discussions and decision-making on fiscal and budget policies at the Cabinet level i s limited and ad-hoc. The Cabinet does not systematically discuss or agree upon the choices made by the MEFB at an early stage, especially in relation to macroeconomic and fiscal projections as well as sectoral and ministerial allocations. The Cabinet also does not seem to be systematically associated with the modifications o f the budget, following budgetary conferences or negotiations with the IMF. -27- Box 11-3 Developing a strategic decision making process The involvement of the Cabinet Effective involvementof the Cabinet in the budget formulationprocess allows to: (i) build a government consensus on national policy priorities and allocation of related expenditures; (ii)enhance transparency through open and direct discussions, and; (iii) obtain strong commitment from the members of the Cabinet on the decisions made in view to increasethe discipline inthe budgetformulationprocess. Institutional arrangementsto support involvement of Cabinet depend on country circumstances. In some countries, the Ministry of Finance plays a leadrole inpreparingCabinet meeting,while inotherspreliminary discussionstake place at ministry level to facilitate resolution of technical issues and allowing the Cabinet to focus on strategic issues. For example, in Sweden, discussions on the fiscal policy, expenditure envelope, and sector ceilings take place in a two-day Cabinet retreat, with all decisions being taken by the Cabinet as a collective body. In South Africa, a formal subcommittee of the Cabinet considers policy changes with budgetary implications before making collective recommendationsto Cabinet. This is used by Cabinet to make all the final decisions on Government's medium-term policies and spending priorities. InUganda, the process starts with a Cabinet retreat to discuss strategic issues such as the initialresource framework andthe key budgetissues andpriorities for the comingfiscal years. The preparationof a strategic document (the BudgetFrameworkPaper) also facilitates decision-makingof the Cabinet while increasing transparency of debates. Good practice suggests that a Budget Framework Paper which should be updated as the budget preparationprocessprecedes includes (i) a review of past performanceand prospects in global and regional economies; (ii) budget out-turns(including revenues, expenditures, debt, deficit) and progress on policies implementations; (iii) medium-term budget prospects, (iv) overview about anticipated revenues (revenue targets, proposed changes in tax policies); debt; deficit financing and expenditure priorities (targets, evolution in sectoral allocations and key expenditures categories) with reference to the policy priority of the Government and its reform program. Source: World BanWKorea Development Institute, Reforming the Public Expenditure Management System, March 2004; OD4 Implementing a Medium-term Perspective in Budgeting in the context of national PRS ,undated. 0 Parliament: The National Assembly has only limited capacity to review the budget. According to the MEFB the National Assembly has not made any amendment to the proposed budget in recent years. In addition, transmission requirements to Parliament limit budget preparation time. The Constitution states that the annual budget must be presented to Parliament by the end o f October. The October submission o f the budget restricts the preparation process, taking away valuable weeks that could be used for consultation and the arbitration process between line ministries and the Ministry o f Finance. The budget proposal i s typically subject to substantial adjustments following the IMF mission inOctober. As a consequence, Parliamentvotes on an outdatedbudgetary framework, which drawsinto question the usefulness o f the Parliamentary vote. Recommendations o Enhance role of Cabinet.The involvement o f the Cabinet inthe different stages o f the budget preparation process should be reinforced to enhance policy discussions and strategic decision-making that links priorities and available resources. Discussions at the Cabinet level should be informed by a Budget Framework Paper developed by the MEFB at the beginning o f the budget year (see box 11.4). This framework paper would be updated during the budget preparation process. Government should also negotiate with the IMF to better associate the IMF mission schedule with the budget preparation process. -28- Box 11-4 The BudgetFramework Paper Good practice suggests that a Budget Framework Paper which should be regularly updated during budget preparation process, includes: (i) reviewofpastperformanceandprospectsinglobalandregionaleconomies; includingtheexternal A sector, real sector and financial sector; (ii)Budget out-turns (including revenues, expenditures, debt, deficit) and progress on policies implementations. ,(iii) Medium-term budget prospects (iv) Overview about anticipated revenues (revenue targets, proposed changes in tax policies); debt; deficit financing and expenditure priorities (targets, evolution in sectoral allocations and key expenditures categories) with reference to the policy priority o f the Government and its reform program. Source: WorldBank 0 Strengthen participation of line ministries. Participation o f line ministries in the budget preparation process should be significantly strengthened. This does not necessarily require two cycles o f budgetary conferences, as envisaged by the MEFB; key to improved participation i s timely information by MEFB to the line ministries. Early information concerning the budgetary envelope, using multi-year projections, would allow the line ministries to initiate a more effective planning and consultation process. Once the budget circular i s issued, more time should also be given to line ministries to prepare and submit their budget requests. In addition, training needs o f line ministries in the area o f budget planning/costing should be identified and included inthe plannedtraining program o f MEFB. P Improve budget circular and introduce budgetframework paper. The budget circular should include specific information that would allow line ministries to understand the strategic policy priorities o f the Government within the hard budget constraint. The circular should be developed on the basis o f a Budget Framework Paper that should be presentedto Cabinet. a Refocus budgetary conferences. Budgetary conferences between the MEFB and the line ministries should move away from the discussion o f the baseline budget and its proposed modifications to an interaction that i s focused on policy objectives, prioritization o f needs within the hard budget constraint and results achieved on the basis o f resources allocated. LI Adjust budget preparation calendar. To improve the quality o f the preparation o f the annual budget the Government should adjust the calendar and the stages o f the budget preparation process. Table 11.7below outlines the proposed budget process based on good practices inother countries. Delay submission of full budget to Parliament. Inview of the statutory obligation to submit the budget to the National Assembly by the end o f the month o f October the Government could consider putting forward a budget framework paper instead o f the completed budget, which in turn could be presented to the National Assembly in November. This approach would free up time for a more participatory preparation process, integrating in particular Cabinet and the line ministries into strategic decision making. It would also ensure that Parliament can vote on a high-quality budget. LI Strengthen capacity of Parliament. The Government should continue to strengthen the capacity o f the National Assembly in the area o f public finance to professionalize interaction. Inthis context the Government could also organize an annual workshop with Parliament to discuss the draft annual budget or other selected issues. More fi-equent interactions with the MEFB on budget execution -29- issues, as envisaged by the n e w Organic Finance Law, would contribute to improving the capacity of Parliament over the medium term. Table 11-7: The 2003 budgetprocess and proposed calendar 2003 BUDGETPROCESS PROPOSEDCALENDAR February-June February-April reparation o f a Budget Framework aper by MEFB. June Oral communication to Firsthalfo f May Discussion o f the Budget the Cabinet Framework Paper by an InterministerialCommittee July Budgetary Conferences Mid-May Approval o f the Budget Framework Paper by Cabinet Beginning o f August Notification of the Second half o f M a y Issue o f the Budget Circular which i s budget envelopes derived from the Budget Framework Paper EndofAugust Submission by the line Endo fJuly Budget submissions by line ministries ministries September Preparation o f the draft August-September Budgetary Conferences and revision Budget L a w o f the Budget Framework Paper; preparation o f the draft budget by MEFB(based onupdate o f I macroeconomic and fiscal forecasting and on the results o f budgetaly-conferences) i October bodifications o f the I October Discussion o f the revisedBudget acroeconomic and ramework Paper and the draft fiscal framework, udget by an Interministerial following negotiations Committee ithIMF.Subsequent adjustments to the udget L a w made by he MEFB, without consultation o f the line nistries. I Endof October Submission o f the draft Endo f October Approval o f the Budget Framework Budget L a w to the Paper and the draft budget by National Assembly Cabinet; submission o f the Budget Framework Paper to Parliament November Submission o f the draft Budget to Parliament -30- (e) Planningfor RecurrentandInvestment Expenditures Issues and CurrentPractices 59. Madagascar uses a dual budget system.. Although a consolidated budget i s presented in .. the annual Budget Law, a dual budget system still governs the budgeting o f recurrent and investment expenditures which are prepared according to different timelines, procedures and institutional responsibilities. Box 11-5 Dualbudgetsystems Dual budgets - in which recurrent and investment expenditure are budgeted separately -emerged around the time o f independence as a way to manage a growing volume o f aid financing and to outline the Government-led development strategy based o n public capital investments. Development partners appreciated the dual budget system, as it facilitated the coordination o f aid and the identification o f hture projects. Today, it i s recognized that the development effectiveness o f public expenditures is hampered by the dual budget system, as the recurrent cost requirements o f investments are not rigorously estimated and thus, the long-term affordability o f policies is not properly considered. This leads to inadequate provisions for recurrent expenditures in subsequent budgets and unsustainable investment decisions. Moreover, as the investment budget i s driven by donor preferences, projects are not necessarily integrated with the Government's policies and priorities or linkedto the current budget. Dual budgeting contributes to problems inlinking planning, policy and budgeting which are the main cause for poor budgetary outcomes. Source: WorldBank 60. ...which hampers coordination of investment and recurrent budgets.At the level o fthe line ministries, budget preparation is made in parallel by the Directorate o f Financial Administration (Direction Administrative et Financiers, DAF) for recurrent expenditures and by the Directorate o f Planning (Direction des Etudes et de Planification, DEP) for investment expenditures. The consolidation between these two categories o f expenditures i s problematic: medium-term recurrent costs projections o f existing and planned investments are not made in a systematic manner; there i s also little information about the affordability o f existing policies and the sustainability o f present investment decisions. 61. ...and results in inconsistency of investment and non salary recurrent expenditures. As a result o f these separate processes investment and (non salary) recurrent expenditures are not fully aligned and consistent. Efforts have been made to better align the recurrent and investment budgets. Chart 11.2 below indicates that non salary recurrent and investment budgets have been increased by more or less the same ratio since 199ga2'Relative to 1998, non salary recurrent expenditures increased by about 230 percent while investment expenditures increased by about 180 percent until 2004. Subsequent analysis in the education and transport sectors (see chapters below) shows that the level o f recurrent expenditures in2003/2004 i s not sufficient, despite Government's commitment to improve the mix o f recurrent and investment expenditures. Furthermore, in the year 2004 capital expenditures increased sharply, even outpacing non salary recurrent expenditure growth. This raises concerns about the adequacy o f allocations for operations and maintenance. 29 N o n salary recurrent expenditures exclude transfers. -3 1- Chart 11-2: Annual growth of nonsalary recurrent expenditures and investment 1998-2004 (inpercent) 40.0 30.0 20.0 10.0 0.0 -10.0 -20.0 -30.0 +Nonsalaryrecurrent-+- Investment Source: WorldBank 62. Misclassification results in data inconsistencies. An adequate analysis o f public expenditures (which should serve as a basis for political decision making) i s difficult because o f problems with the classification o f recurrent and investment expenditures. It i s estimated that around 10-15 percent o f capital expenditures (mainly those that benefit from external assistance) are inreality recurrent expenditures. This indicates difficulties in adequately applying the existing budgetary nomenclature (Plan Comptable des Opirations Publiques, PCOP) as well as the functional classificationo f expenditures (classification according to sector organization). 63. Aid delivery modalities reinforce the dual budgeting system. The modalities o f aid delivery reinforce the dual budgeting system. The budgeting process o f externally-financed investment expenditures takes place without reference to a budget ceiling; it i s only constrained by the availability o f external assistance and related counterpart funds. As a consequence, little consideration i s given to the future cost implications o f investment projects and the ability o f the budget to sustain these costs. The shift towards budget support, as initiated by some key donors such as the World Bank, the African Development Bank and the European Union, facilitates the integration o f foreign assistance into the budget and, thus the consistency o f recurrent and investment expenditures. -32- Box 11-6 The impactof budgetsupport A recent study o n budget support describes how in Uganda the s h f i towards program lending changed the incentive structure in the public administration: "Because line ministries and districts no longer have the option of seeking directfunding from donors, or have it within a strict ceiling agreed in the MTEF process, they have a much stronger incentive to comply with budget directives. These now call for a more structured approach to policy development, the lynchpin what are being called in Uganda sector-wide strategies. Sector- wide strategies become in-turn the basis of Budget Framework papers, which are the sector working groups' submissions within the MTEF, and ultimatelyfor resource allocation decisions in the budget .... 'I The report also notes that budget support encourages donors to focus their attention o n certain cross-cutting issues, such as strengthening o f procurement. Improvements also seem to have been made in relation to domestic accountability: "Auditors General3 reports are said to be improving. The Public Account Committee is becoming more active in following up cases of misappropriation .... The press, even though partly Government-owned, is increasingly open in criticizing public corruption and the insufliciency of action to curb it." The report recognizes that these positive trends require additional elements or have independent origins, but stresses that program lendingprovides a general enabling environment. Source: Oxford Policy Management and Overseas Development Institute, General Budget Support Evaluability Stud; Phase 1 - Final Synthesis Report, Report to the UK Department for International Development (DFID), 30 December 2002 Recommendations Integrate capital and recurrent expenditures. Budgeting o f recurrent and investment expenditures should be fully integrated. The Government could consider different options: One option would be to harmonize the preparation o f the recurrent and investment budgets on the basis o f a unified preparation process at the line ministries and MEFB levels. This would involve in particular requesting the DAF and the DEP in the line ministries to develop a uniform budget proposals which addresses the cost implications o f planned investments within a medium-term envelope. A more radical option would be to integrate the divisions incharge o f planning recurrent and investment expenditures, in particular the DAF and the DEP. Such an approach would be consistent with the introduction o f program budgeting, but would require substantial organizational and procedural changes. Educate sector ministries. The MFEB should provide guidance and training to line ministries how to adequately determine the recurrent costs o f investments as well as to assess the medium-term implications o f existing projects and planned investments. These costs should be systematically integrated in the medium-term projections o f programs and into the annual budget. New investments should only be approved if medium-term projections confirm the financial sustainability (Le. availability o f resources to finance day-to-day operation and maintenance costs). Encourage donors to increase budget support. In view o f the experiences in other countries (see box 11.6 above) the Government should consider to intensify the dialogue with the donor community about the advantages o f budget support over project assistance. The shift to budget support reduces the incentive to maintain the dual budget system, facilitates strategic prioritization o f resource allocation, and increases the flexibility to determine budgetary priorities and tradeoffs. -33- (f) Revenue and External Assistance Forecasting Issues and Current Practices 64. Revenueforecasting in Madagascar is weak both in relation to Government revenues and external assistance.Chart 11.3 a-b shows that revenues and external assistance are systematically over- estimated, resulting in lack o f predictability and the build-up o f arrears. The large variation in 2000 between planned and executed revenues are due to the delays in the privatization program. While the year 2002 indicates the l o w level o f revenue collection during the political crisis, the following year 2003 illustrates the slow rebound o f the economy, which recovered only inthe beginning o f the second half o f the year. The tax and custom administrations prepare revenue projections at the beginning o f the budget preparation process and communicate them to the Economic Directorate o f MEFB, which i s responsible for the final projections. Little capacity exists to accurately estimate the expected revenue level by taking into account the implications o f the economic and fiscal policies o f Government. Frequent and ad-hoc revisions o f the tax policy further exacerbate the difficulty o f providing accurate revenue forecasts. Chart 11-3 a-b: Comparison of planned and actual revenues and donor funds, selectedyears between 1997-2003 (inpercentage ofnoninterest expenditures) 4500 0 4000 0 3500 0 3000 0 2500 0 I OP Dlnea 2000 0 re en.es < .I m o n M G F 1500 0 m A c .a 10 cr.01 n 1000 0 m o - f $ G ' 500 0 0 0 _ _ 1997 1999 2001 2002 2003 1997 1999 2003 2002 2003 I Source: IMF and WorldBank estimates 65. Chart 11.3.b also indicates problems to forecast foreign external assistance. The differences between planned and executed donor funds suggests not only difficulties in accounting and reporting o f donors funds but also an overly optimistic assessment o f the availability o f external financing and o f the (absorptive) capacity o f the administration to implement investment projects. No viable methodology and only limited capacity seem to exist to adequately project these inflows. The year 2003 i s an exception with high convergence between projected and actual foreign aid. As donor support shifts to budgetsupport, this gap will decrease. 66. Different disbursementprocedures pose problems. The modalities o f aid delivery, mostly project aid, also increase the difficulty in developing more accurate forecasts, especially as regulations o f project disbursements vary from one donor to another. The shift o f several donors to program lending acknowledges the limitations o f project assistance and recognizes the need to support the Government ina consistent manner with full Government ownership. -34- Recommendations 0 Further develop revenueforecasting capacity. The improvement o frevenue forecasting should be given highpriority by the Government. The relevant capacity in the MEFB in general as well as the tax and custom administrations in particular should be systematically strengthened. The Government should also develop a medium-term reform plan for the tax and custom administrations. This plan should be linkedto specific revenue targets and performance standards. 0 Improve forecasting of donor aid. Reliable forecasting of donor assistance i s required to improve the credibility and transparency o f the budget. These forecasts should take as a reference point average disbursements in previous years, but taking into account the move to budget support. The Government should aim at strengthening coordination between the MEFB and the line ministries to improve reporting and monitoring o f external assistance. Government should also urge development partners to provide timely information about the disbursement o f their assistance; this should also include a medium-termperspective about the intended support to improve budget predictability. (g) ComprehensivenessandTransparencyof the Budget Issues and CurrentPractices 67. Comprehensive coverage of all fiscal activities undertaken by the central Government i s essential to ensure that the budget i s transparent and can be used as an effective planning and co- ordination tool. Transparency o f the budget preparation process i s also key for ensuring a more participatory process within the Government. In addition, the presentation o f the annual Budget Law strengthens public understanding o fpolicy priorities. 68. Although the budget in Madagascar i s regarded as comprehensive it includes mainly aggregated categories o f expenditures and lacks certain information. For example, transfers to municipalities which include administrative, health and education allocations are not specified indetail. Transfers to state-owned enterprises are indicated in the budget, but no information i s reported on their gross expenditure and revenue, including revenue from earmarked taxes, user-charges and other sources of revenue. Moreover, little i s known about the activities o f some extra-budgetary institutions, in particular social security funds, and thus about the nature o f any riskposed by the broader Government sector to the sustainability o f central Government finances. Recommendation Improve budget transparency. Further information should be provided inthe Budget Law about the transfers to municipalities and state-owned enterprises. The Government should review ifthe existing classificationprovides a sufficient degree o f transparency. The Government should also clearly establish the fiscal risks that may potentially arise from activities o f extra-budgetary institutions. -35- 2) ENSURING BUDGETIMPLEMENTATIONINLINEWITH GOVERNMENT PRIORITIES 69. Problems with the implementation o f the budget are often at the root o f a country's fiscal failures. While it i s a prerogative o f Government to ensure that public monies are efficiently spent, in practice the administration faces many obstacles, which are discussed indetail inthis chapter. (a) BudgetExecutionremainsa Challenge 70. Low budget execution rates. Over the past years budget execution in Madagascar has been relatively l o w - it notably deteriorated since the end o f the 90's from 96 percent in 1998 to 75 percent in 2001 (see table 11.8) - mainly because o f weak revenue collection and o f budget implementation problems (see following sections). Some early estimates o f the 2003 budget execution show a 92 percent executionrate o f the total budget. This improvement - if confirmed i s an encouraging sign o f - Government's efforts towards more solid public financial management. It would also indicate that Government's efforts in 2003 to streamline the execution process, e.g. by setting up one-stop shops "guichet unique" and by simplifyingprocurement procedures, has yieldedpositive results. Table 11-8: Madagascar's past execution trends (selected years between 1997-2003) (inpercentage of total noninterest expenditure^)^' 1997 1999 2001 2002 2003 Total 84.2 79.7 75.0 43.3 91.9 Governance 78.9 62.4 80.1 57.0 89.1 General Administration 74.3 56.7 77.7 51.3 86.8 olw budget, finance & eco. Admin 61.2 43.1 74.6 47.5 84.3 Defense 98.9 100.9 97.6 93.0 95.3 Public security 97.9 93.O 93.3 78.1 99.7 olw Justice 93.3 99.8 94.2 57.6 103.3 Growth 89.7 97.4 72.5 22.7 98.2 Energy 33.7 73.4 69.9 8.2 63.0 Agriculture 78.7 79.9 72.7 23.5 84.6 Environment 124.0 205.3 144.0 41.7 85.2 Transport & Communication 94.8 92.0 51.1 21.5 89.3 01w Transport 95.3 91.9 50.8 21.0 89.2 Others 118.3 96.5 94.2 19.9 144.9 Social 85.4 97.5 69.5 50.7 90.9 Education 103.9 112.4 78.2 63.5 97.0 olw Primary Education 92.5 134.0 85.3 60.7 78.8 Health 85.4 88.8 68.1 37.7 97.2 olw Primary Health care 71.6 100.0 82.0 33.0 81.3 Others 61.9 80.9 54.7 38.2 68.9 Source MEFB and WorldBank estimates 30 See annex 5, table 5.7 for reference regarding the real annual growth rates of actual expenditures 1997-2004 -36- 71. Poor performance of public investmentprograms. The comparison o f past performance o f recurrent and investment expenditures points to severe problems with the implementation o f public investment projects. While execution o f salary expenditures i s relatively satisfactory achieving an average spending rate o f 96 percent over the observed period, the execution o f non salary and capital expenditures i s fairly low with an average execution rate o f only 82 percent and 70 percent during 1997-2003 (see table 11.9). Table 11-9: Madagascar's recurrent and investmentspendingpatterns selected years between 1997-2003 (actual expenditures as percent of budgetary allocations) ~~ Aver. 1997 1999 2001 2002 2003 1997-03 TOTAL 84.2 79.7 75.0 43.3 91.9 78.5 Current Expenditure 101.4 92.2 83.2 63.6 94.7 88.7 Salary 98.7 100.9 95.9 84.3 94.3 95.7 Goods & services 94.1 89.1 83.0 50.7 90.9 82.5 Transfer 114.0 89.2 69.4 42.3 97.3 86.0 CapitalExpenditure 71.0 67.9 67.3 24.9 88.5 69.2 Internalresources 45.5 33.2 66.5 40.9 54.4 52.0 Externalresources 88.9 102.1 69.9 11.7 119.7 83.8 Memo External resources (as a % o f actual capital expend) 75.7 78.3 52.9 29.5 70.8 64.1 Source: MEFB and WorldBank estimates 72. High execution rates for HIPCfunds. At the same time, execution rates for HIPC-funds, implemented since 2001, are superior. These funds represent only 6 percent o f Govemment's budget in 2003; their execution rate reached 99 percent. HIPC funds were broadly earmarked at the decision point which simplified the yearly identification o f expenditures. HIPC money (more than 60 percent) could also finance recurrent expenditures, which generally have a better execution performance. The application o f special budgetary procedures3' also contributed to a better performance o f HIPC resources. These procedures allow for the management o f funds outside the public finance control system. Additionally, they permit the disbursement o f funds even after the closure o f the budget year (2002 HIPC funds were executed untilmid2003). @) Credibilityof the Budget stickingto BudgetedPriorities - 73. The credibility o f the budget i s important for effective implementation o f Govemment's policies and programs. It i s challenged when unplanned, commonly referred to as unbudgeted expenditures occur or when revenue shortfalls duringthe year require an adjustment o f the budget. 74. The issue of unbudgeted expenditures. Unbudgeted expenditures occur when the treasury accepts payment orders under the "droit de requisition pr~cedure.~' " This procedure was extensively used by the previous regime to execute expenditures under special accounts. In 2001, the President's special fund was considerably overdrawn with a total spending o f US$25 million compared to an 31A special account was set up for the management o f the HIPC funds to which all funds are directly transferred from the treasury. 32Under the "droit de requisition procedure the Credit Manager with the authorization o f the MEFB can ask for " the release o f a payment that had initially been refused by the Treasury. -37- allocated budget o f US$400,000. In the aftermath of the political crisis in 2002 the new Govemment introduced regulations to limit the use of this procedure. It also agreed to regularly audit the special accounts by the Auditor General. Inthe absence o f national emergencies, unbudgeted expenditures are viewed as poor public management practice because they distort the original prioritization of the resource envelope as reflected in the budget and as approved by Parliament. Despite Govemment's efforts to limit the use of this instrument, recent trends are not encouraging: the 2003 budget outtum o f unbudgeted expenditure amount to more than MFG200 billion (nearly 1percent of GDP), drawing into question the credibility o f the budget. 75. Substantial gap between the approved and executed budget. Two major findings stand out when allocation and expenditure data are analyzed: the gap between original budget and final allocation (initial budget allocations plus mid-year adjustments to the original allocations) has been closed in 200333.At the same time, the gap between final allocation and execution increased substantially since 1999 (see chart 11.4). The largest gaps occurred in 2001 (25 percent) and in 2002 (57 percent). The year 2003 i s an exception with an execution rate of 92 percent. While this indicates progress, line ministries continue to be concerned about the budget as a reliable planning instrument, mainly because o f past experiences with unpredictable actual releases. The uncertainty about the available resources envelope negatively impacts the ability o f line ministriesto implement policies and programs. Chart 11-4: Comparison of pastbudgetallocationand executionbetween1997 and 2003 (in billionMGF) 8,000.00 7,000.00 6,000.00 5,000.00 4,000.00 3,000.00 2,000.00 1,000.00 1997 1998 1999 2000 2001 2002 2003 I annualallocation(inbillionMGF)ITotalannualexecution(inbillionMGF) Total Source: MEFB 76. Legal instruments to modify the budget. Govemment has two possibilities to legitimize unbudgeted expenditures or revenue losses: (i) the preparation of a revised finance law and (ii) the utilization o f Article 46 of the Organic Finance L a w (Article 20 o f the new Organic Finance Law, see box 11.7). In the past, Government mainly used the stipulations of Article 46 to legitimize the gap between actual allocations and the allocations made by the original finance law. The application o f Article 46, however, undermines the legitimacy of the budget as original priorities are altered without approval o f Cabinet or Parliament. 33Deviation from the initial finance law decreased from 12 percent in 1997 to 7 percent in 2001. For the year 2003 original credit allocations matched with final credit allocations. -38- Box 11-7 Madagascar's legal instrumentsto modifythe nationalbudget The revised budget law: Following the francophone model, Madagascar could submit two budgets to the National Assembly every year: a) the initial budget (loi de finances initiale, LFI) and b) the revised budget (loi de finance rectificative, LFR). The latter aims to accommodate changes and modifications by the Government during the first semester o f budget implementation and i s submitted once again for approval to Parliament. Typically, these changes or modification do not substantially alter the initial budget. Inrecent years, no LFRhas been preparedby the MFEB. Article 46 of the Organic Law of Finance: The mainreason for. neglecting the possibility to amend the budget on the basis o f a LFR i s the latitude provided by the provisions o f Article 46 (Law Nr. 63-015) o f the former organic law o f finance. This law permits to integrate de facto modifications to the budget o f the current year into the budget o f the following year. In principle the stipulations o f the law require an emergency situation; in reality the Government is using the law for all relevant budgetary modifications. Art 20 o f new organic law o f finance (Law N o 2004-007) restricts the ability to modify the budget: the new law requires Government to seek Parliamentary approval o f the emergency situation; this approval will be provided by the Office o f Parliament. ~ Source: WorldBank 77. Strong correlation between revenue forecasting and low budget execution. An important factor for the gap between the allocated and finally executed budget i s the adjustment o f projected revenues due to lower revenue inflows during the fiscal year. Revenues have been systematically over- estimated since the year 2000 (see chart 11.5).They were revised downwards by 22 percent in2001 and 44 percent in 2002. As a consequence, expenditures had to be significantly adjusted to minimize the impact of the revenue shortfall on the overall fiscal deficit. These adjustments, which are made on an irregular basis by the MEFB,reduce the resources envelope of line ministries and spending agencies. Chart 11-5: Trends infiscal revenuesandbudgetexecutionrate (inpercent) 100.0 80.0 60.0 40.0 20.0 forecasted 0.o revenue --cBudget -20.0 execution -40.0 -60.0 Source: MEFB (c) CashmanagementAnd CommitmentofExpenditures 78. The current cash management system contributes to low execution rates of the budget as ministries are unable to spend their budget allocations on a timely basis. The Treasury Department has -39- developed a cash management plan with the assistance o f the IMF. The plan i s implemented at the central Treasury inAntananarivo. It aims to balance the outflows and inflows o fpublic resources34.The plan faces several shortcomings, in particular problems with data reconciliation between the Treasury and the Central Bank, and difficulties to supply cash to remote treasuries. In addition, unreliable cash flow forecasting by line ministries makes it difficult to accurately determine payment needs for the whole Government. Consequently, the Treasury i s not able to meet payment obligations on a timely basis as well as any important demands or seasonal needs. Regional Treasury offices are short o f cash, which leads to difficulties inthe payment o f services and salaries. It i s expected that the implementation o f a country-wide integrated financial management system (IFMIS) will improve problems o f data reliability, completeness and consolidation. 79. Commitment procedures cumbersome. The ability o f line ministries to execute the budget i s further constrained by existing procedures concerning the commitment o f expenditures. When the budget i s approved (typically in December or in January) line ministries need to appoint Credit Managers who are responsible for the commitment o f expenditures. These nominations are often delayed (even for several months) because o f bureaucratic bottlenecks. Many line ministries are therefore not able to make commitments against the budget at the beginning o f the budget year. In addition, the commitment period i s limited until the end o f October (mainly to avoid over-commitment o f expenditures). (d) ExpenditureManagement process 80. Complex disbursement procedures. Processing public expenditures in Madagascar i s complex and long: the CFAA identified 11 different steps with each stage requiring a formal approval (see Annex 6). As the transaction costs are highthe use o f "simplified procedures" to release funds i s on the rise, which - according to public finance regulations - should only be used in exceptional circ~mstances.~~The CFAA estimates that such simplified procedures are used for up to 30 percent o f the expenditures. 81. "Flexible" investment budgets. Line ministries have certain flexibility regarding the investment levels incorporated inthe capital budget. On an occasional basis, they are allowed to spend more than the amount indicated in the Public Investment Program (PIE'); e.g. if (i)the rate o f implementation o f a particular project i s better than expected at the time o f the budget (and PIE') preparation, or (ii) domestic funds for the line ministry are unavailable. This "flexibility" carries if substantial risk o f raising the fiscal deficit. It also raises concern with regard to the comprehensiveness o f the budget. 82. Government has begun to address theseproblems: To rationalize administrative procedures it has merged the function o f the "sous-ordonnateur" and o f the Credit Manager. It is also preparing regulations for supporting documents ("nomenclature des pieces justijkatives "j' that are required for the disbursement process. Government also discourages the use o f simplified procedures such as the "droit de requisition and intends to limit the application o f these procedures to exceptional cases. A " new computerized integrated financial management system is under preparation which will further streamline and rationalize expenditure management procedures. 34 The plan contains revenue forecasting (from taxes and customs duties) and expenditure forecasting for non- discretionary expenditures (salaries, pensions, andpublic debt). 35This refers notably to the application o f the "droit de requisition" (see section on the credibility o f the budget) and the "regies d'avance". The later enables the treasury to authorize the execution o f expenditures byproviding a cash advance to Credit manager. -40- (e) Recommendations 0 Prepare 2005 Priority Action Plan. Government should develop the action plan for the year 2005 which should specify Government's priorities in the area of public finance. The MEFB should report to Cabinet about the implementationo f the plan on a regular basis (e.g. every 3 months). 0 Restrict application of Art. 20. Government should limit the application o f Art. 20 o f the new Organic Finance Law to the exceptional cases stipulated in the law. In case o f deviations from the original budget it should follow regularly procedures and submit a LFR. 0 Focus on budget execution rates. Government should systematically track budget execution rates, inparticular inpriority sectors. The MEFB should inform Cabinet on a regular basis (e.g. every 3 months) about the execution o f the budget and potential bottlenecks to enable policy interventions at an early stage. 0 Strengthen cash management. Government should intensify efforts to consolidated accounts on a monthly basis. The line ministries, in close collaboration with the MEFB, should specify their cash requirements which should be consolidated inthe cash management plan at the level of the MEFB. o Extend the commitmentperiod. Government should ensure that the Credit Manager who are instrumental for the commitment o f expenditures are nominated as o f 1'' January of the fiscal year. The commitment period should be extended to the end o f November o f the fiscal year. (0 ImprovingInternalandExternalControls Issuesand current practices 83. Implementation of reforms ongoing. Government has begun to revise the legal framework for the control institutions and to implement capacity buildingmeasures (see Annexes 7 and 8 for more details): it has, in particular, strengthened the operational efficiency o f the internal control o f the Treasury (Brigade de TrCsor) and o f the Auditor General. In 2004, the MEFB also established an internal control mechanism at the level o f the Ministry o f Finance (Inspection GCnCrale des Finances, IGF). The creation o f the IGF, however, has left the competences o f the General Internal Auditor (IGE)36unchanged -resulting in overlapping responsibilities o f IGF and IGE. Government intends to address this issue in the context of the development o f the annual work plans o f the two institutions. It would, however, be useful to amend the legal instrument governing the operations o f the IGE. 84. Since 2001, tracking surveys, conducted by INSTAT and funded by HIPC resources, are used to trace expenditure and to assess satisfaction o f beneficiaries o f public services. Though still limited to expenditure o f the health and education sectors these surveys are important instrument to measure the impact o f public spending intwo key PRSP sectors. Recommendation 0 Clearly define the relationship between IGF and IGE. Government should clarify the roles and responsibilities of IGF and IGE. To this end, it should amend the legal instrument for the IGE. Government should also review the efficiency o f the internal control system on a regular basis (e.g. yearly). 36 IGEis attached to the Presidency -41- (g) StrengtheningPublic Procurement Issuesand CurrentPractices 85. Public procurement represents 9-1Opercent of GDP. A Country Procurement Assessment3' was finalized in June 2003. The CPAR report highlights that better procurement practices and institutional arrangements could save 20-35 percent o f the procurement cost. In addition, the acceleration o fprocurement procedures could improve the absorptive capacity o f spending agencies. 86. Weaknesses of the Malagasy Procurement System. The main weaknesses o f the Malagasy procurement system are documented in the CPAR, and include: (i) an outdated regular framework that requires modernization and clarification in a number o f areas (i.e. coverage o f the procurement code, thresholds), (ii)a lack o f standardized bidding and contractual documents and related manuals to facilitate the procurement process, (iii) a weak institutional framework combined with low operational efficiency o f tender boards and lack o f service standards (iv) low transparency and inefficient control mechanisms; and (v) limited capacity inthe area o f public procurement inthe administration. 87. Government reforms under way. In the aftermath o f the political crisis in 2002, the Government began to initiate transitory reforms targeting procurement procedures with the objective to accelerate budget execution duringthe implementationo f Government's reconstructionprogram. These regulations foresaw an increase o f the threshold for certain goods and servicesJ8, a reduction o f delays at the level o f the tender board^.'^ In addition, "one-stop shops" ("guichet unique") were set up to accelerate the clearance process. These reforms have been successful although they have not fully resolvedthe more fundamental weaknesses o f the public procurement system.@ 88. The main instrument to fundamentally reformthe system i s the new procurement code which has recently been approved by Parliament. The new code will pave the way for significant procedural and institutional changes by transforming the central tender board into a Procurement Oversight Institution, limiting its ex ante control and advancing its ex post auditing function. Additional work i s ongoing to develop standardized biddingdocuments and proceduralmanuals, which should be available in early 2005. The key challenge in 2005 will be to institutionalize these new arrangements, in particular to ensure that the new regulations are consistently applied by the public administration. Recommendation A comprehensive training and capacity building program is needed. A comprehensive training and capacity buildingprogram should be prepared in close collaboration with relevant training institutions to ensure that line ministries adequately and consistently apply the new regulations. If necessary, local training institutions should be strengthened to address potential supply side problems. The Government's budget should set aside sufficient resources for the demand and supply side o f training and capacity buildingactivities. 37Madagascar, Country Procurement Assessment Review (CPAR), June 2003. 38The threshold for consultations o fprices for furniture and services were raised from 20 to 30 million MGF and those for public work from 50 to 100million MGF. 39Issuedby the ((circulaireD n0108/PM/SGG inOctober 2002. 40Approved as a decree, the new regulations cannot overrule the existing procurement code. The administration and private operators seem to be not well informed about the new regulations and continue to apply the existing code procedures. Several o f the measures conflict with other rules (e.g., the new regulations allow for the bidding process to start before line ministries are notified about their credits, while such practice is not permitted under other rules). -42- (h) ImprovingFinancialReporting Issuesand CurrentPractices 89. Financial reporting dejkient. A key prerequisite for a modem budget execution system i s reliable and timely financial reporting. Madagascar's reporting system is, however, flawed by lack o f coverage, timeliness and reliability o f the execution data. Key information i s not available, i.e. execution data on investment projects financed by donors, activities o f local Governments and a number o f public institutions. Financial statements fiom the sub-national levels o f Government are sent with delays to the central Govemment level - as a result budget execution data is outdated and inconsistent. There i s no systematic reconciliation process between the Treasury and the Central Bank. Many public entities have accounts in commercial banks; the account information i s not known by the Treasury. 90. Budget reporting is not done by all line ministries. While budget execution reports are prepared by several key ministries (Education, Health, Agriculture, Justice, Water and Forest and Public Works), this has not been extended to all ministries. The education and health ministries have produced regular budget execution reports since 2000, which are available with a delay o f two or three months on a tri-semester basis. In 2003, additional budget execution reports were prepared by four additional priority ministries (Agriculture, Justice, Water and Forest and Public Work) to fulfill a Government commitment inthe HIPC context. 91. Treasuryperformance requires improvement. The Treasury also faces difficulties inclosing the annual accounts in line with the provisions o f the Organic Finance Law. Govemment reforms, however, have yielded some positive results.41 The final accounts for 2002 and 2003 have been submitted to the Auditor General. The Auditor General faces capacity problems in validating the Govemment accounts on time. As a result the audit report (((L o i de Reglement D)for 1996 Law was presentedto the Parliament only in2000.42The law for 1997 was adoptedby Parliament in2004. 92. Poverty related expenditure should be systematically tracked. The introduction o f a new budget classification system as well as the Tableau de Bord (TDB) o f MEFB improved the capability to track expenditures from the commitment to the payment stage. However, the range o f possibilities to define poverty-related expenditures has not yet been fully expl0ited.4~. Quarterly reports exist for certain priority spending ministries in the area o f poverty reduction, but do not allow a functional analysis o f poverty-reducing expenditure although the relevant classification exists since 2000.44N o code i s provided inthe budget and accounting nomenclature to specify poverty-reducing expenditure by budget line - only the tracking o f HIPC expenditure i s clearly identified inthe budget. 41 Starting in 2001, the centralization procedures and consolidation o f all balances o f the Principal Treasurers by the ACCT have been improved through the formation o f a consolidation and audit unit, allowing for the establishment o f general balance sheet statements and o f reliable opening and closing accounting balances in a timely manner. 42The regulations require that the draft budget review law for the prior year be submitted to Parliament no later than November 15 o f each year. It has not been possible to meet these deadlines owing to human, physical, and organizational resource constraints. 43World Bank (2003) "Madagascar Tracking Poverty-Related Spending Assessment and Action Plan"; IMF (2004) " Madagascar Heavily IndebtedPoor Countries Assessment and Action Plan" 44A study by Cornel1 University and INSTAT found that only half o f the poverty related expenditures inhealth and education sectors can be considered to be poverty reducing. Glick, Peter and Mamisoa Razakamanantsoa (2001), Services d'kducation et de santk a Madagascar: l'utilisation et les dktenninants de la demande. Antananarivo, INSTAT. -43- 93. Government reforms under way. To improve the overall efficiency o f the reporting system Government has launched several reforms: First, a computerized financial management system was developed at a pilot site inTamatave. The system, which has been operational since 2002, integrates the different stages o f the budget cycle to improve availability and comprehensiveness o f financial information as well as to enhance transparency and accountability o f public finance operations. The Government intends to deploy the system to additional key treasuries. Inthe long run it i s envisaged to implement an IFMIS which follows international standards. Second, Government commissioned an independent audit o f the which aims to improve the operational efficiency and transparency o f treasury operations. The implementation o f the recommendations o f the audit i s ongoing. Third, MEFB intends to develop user manuals to clarify accounting and reporting procedures. This is particularly relevant for the line ministries which lack basic knowledge and capacity inthis area. Recommendations 0 Continue treasury reforms. The Government should continue to improve the operational effectiveness o f the Treasury. While the modernization o f the Treasury and the introduction o f a comprehensive integrated financial management system may resolve existing problems inthe long run, it will be important to identify concrete actions to be implemented inthe short run.Inthis context, the independent audit o f the Treasury has proposed important changes in the procedural and institutional set-up o f the (central and provincial) treasuries - these recommendations should be implemented without delay. To ensure full accountability o f the Treasury for the successful implementation o f the reforms benchmarks should be introduced which - over time - aim at bringing treasury operations to the intended service standards (e.g. monthly accounts produced within two weeks after the closing o f the previous month; deviation between treasury accounts and Central Bank accounts reduced by x percent). 0 Improve tracking of poverty related expenditure. Pro-poor expenditures should be clearly identified in the budget. To better target public expenditures it could be useful to classify budgetary expenditures according to their level o f relevance in reducing poverty, e.g. "high poverty reduction relevance", "medium poverty reduction relevance" and "low poverty reduction relevance". These expenditures should be systematically tracked, relevant information should feed into the progress report for the implementation o f the PRSP. 0 Establish an integrated reporting system. The preparation o fjoint budgetary executionreports by the six key ministries since 2003 presents an important achievement bythe Government towards the establishment o f an effective and comprehensive reporting system at the level of the line ministries. These reporting requirements should be reviewed and streamlined with the objective o f establishing an integrated reporting system that can serve different purposes (e.g. PRSP, business plan implementation, budget execution). Such a system should also be capable o f producing monthly status reports. These reports should be evaluated on a regular basis; a consolidated report should be sent to Cabinet on a quarterly basis to substantiate discussions about the implementation o f the Government program. (i) Transformingthe Budgetinto a ManagementInstrument Issuesand CurrentPractices 94. Using the budget to monitor the implementation of policies and programs. The budget should serve multiple purposes - it outlines the strategic priorities o f the Government, it aims at 45This is the second audit o f the treasury since 1997. The first one financed by a World Bank project failed because o f disagreements about the recommendations o f the consultants. -44- ensuring macro-economic stability, it serves as the legislative control over the executive, and it i s a management instrument for the Government. Most modem administrations nowadays use the budget to monitor performance and to measure impact (see box 11.8). If Madagascar i s to succeed with its ambitious plans to transform the country and its economy it will over time need to strengthen mechanisms to adequately assess and monitor the effectiveness o f policy implementation. Box 11-8 Performanceor results-basedmanagement According to a recent Organization for Economic Cooperation and Development (OECD) review, "Zmproved performance of the public sector is a central factor in maintaining welfare of individuals and the competitiveness of the economy. Performance management is the key aspect ofpublic sector reforms of many OECD Member countries." Results based management implies a management culture that i s fact based, results oriented, transparent and accountable. A results based approach requires that the traditional administrative system focused on budget, human resources, and auditing be augmented by a feedback loop on the outcomes and consequences o f Government actions. Over the last two-plus decades many OECD countries -- including Australia, N e w Zealand, Canada, the UnitedKingdom, and the United States - have made progress inimplementing new practices to better manage and account for the results being acheved with public expenditures. Similar efforts have been started inmany developing countries --such as the Philippines, Andra Pradesh in India, Brazil, Tanzania, and others. These experiences provide useful lessons on how to strengthen or introduce a results orientation into Government. 95* Need to look differently at budget. To transform the budget into a management instrument a change in the budget philosophy i s necessary such as from line item budgeting towards program budgeting. The key reference point for program budgeting i s a Government program like the PRSP (or a series o f sector programs). The budget i s classified in terms o f programs and their objectives rather than organizational lines. The attainment o f the objectives i s measured on the basis o f outputs and outcomes which are explicitly mentioned inthe budget. There are various degrees o f this model which reach from a sophisticated budget system that intends to measure efficiency at the output level to a more "simplistic" system that links expenditures to policy objectives. The key advantage o f such a budget system i s the focus on results which allows Government to better track policy implementation. 96. Focus on monitoring and evaluation. The discussion about the introduction o f a more advanced budget system often goes hand-in-hand with the need to strengthen the existing monitoring and evaluation framework o f Government. Such a framework exists inprinciple in every administration but is often deficient (see box 11.9). This situation can also be observed inMadagascar. At the central Government level Government has begun to address these issues with the introduction o f program budgets ("budget program") and the preparationo f work plan agreements ("business plans") by the line ministries for the PRSP implementation. To monitor the implementation of policies and programs the Government has also developed a matrix, which defines specific results for all ministries and departments for the year 2005 ("la Politique Generale de 1'Etatpour 2005"). It i s envisaged to track the implementation o f this matrix on a regular basis as well as to discuss implementation progress and bottlenecks at the Cabinet level. -45- Box 11-9 Commonproblemsof monitoring& evaluationsystems The M&E system often deficient due to several factors : (i) current planning processes do not or not the sufficiently involve the definition o f results information nor how this information will be used to informpolicy, program or management decisions. (ii) There are no systematic feedback loops on effectiveness o f programs and on the achievement o f objectives. A wide range o f data i s collected inministries and departments, however this information i s not used in a systematic manner to inform managers and policy makes. (iii)Monitoring & evaluation mainly focuses o n assessing the physicalperformance of (donor funded) projects. These mechanisms do not encourage a systematic analysis o f common problems across Government. (iv) The public service culture i s not performance and results focused; most institutions do not have service standards to measure the efficiency o f the delivery o f public services. (v) Implication o f civil society in assessing the performance o f the public sector i s weak; users o f public services have very little opportunity to provide feedback to the public administration. Source: The World Bank 97. Developing a more consistent results-based framework. Further work needs to be done to integrate the different activities into a consistent results-based management framework. Though, in theory, the business plans should feed into the preparation o f the program budgets but coordination o f and methodology for the two exercises remain unclear. Many business plans lack detailed costing and fail to adequately prioritize policy objectives/actions. To date, only a few sectors (e.g. education and transport) have advanced fully-costed sectorial programs. Inaddition, capacity at the line ministry level to develop and execute program budgets i s limited. 98. The need to better consolidate and use information. At the level o f the sector ministries work programs and achievements have been subjected to regular reviews, mainly in the context o f evaluating the progress o f PRSP implementation. Through the reporting system set up to monitor PRSP implementation a vast amount o f information i s collected. However, line ministries report that they are overwhelmed with regular reporting requirements. This information i s consolidated and submitted to decision-makers within Government. In essence, however, the information generated by the various reports i s not systematically used by the technical and political levels to adjust policies and to link policy priorities with public resources. It i s also unclear how findings o f the evaluation impact the resource allocationprocess. Recommendations 0 Strengthen the monitoring and evaluation system. The institutional and procedural arrangements of the coordination and consolidation o f policy-related information should be reviewed. With technical assistance the Government should further refine the existing PRSP monitoring system. To ensure that the information generatedat the different levels o f Government i s effectively usedit will be important to establish a well-functioning and well-equipped mechanism at the centre o f Government that can serve as a link to Cabinet and i s capable o f scrutinizing existing information for potential follow-up. Sufficient resources for training and capacity building should be set aside to develop and institutionalise such a mechanism, also at the level o f line ministries. 0 Business plans developed by line ministries should be further refined. They should in particular define more realistic outcomes and the cost implications o f proposed activities. A progress report, which should be presented to Cabinet (e.g. every three months), should systematically evaluate implementation against the proposed targets. -46- 3) EFFICIENT OFHUMANRESOURCES USE 99. Poorly functioning public sector institutions and weak governance are major constraints to growth and equitable development in numerous countries. Consequently, the Madagascar PRSP puts the modernization o f the State and the promotion o f good governance at the center o f the Government's policy agenda. While the PRSP stresses the need to comprehensively reform the public sector previous reform initiatives (at the beginning o f the 90s and at the end o f 1999) by-and-large failed to produce tangible results. Surveys46 and studies4' confirm that the public sector i s flawed by governance problems, inparticular corruption, inefficienciesand mismanagement. 100. The Malagasy civil service accounts for about 146,000 agents in 2004 and represents 0.8 percent o f the total population. Compared to other countries inthe region this number i s relatively low (see table 11.11).Duringthe period 1997-2004, the civil service grew moderately by around 14 percent with an average annual growthrate o f 2.7 percent. Table 11-10: Civil servant ratefor 100 inhabitants insome African countries Civil servant/100 Country habitants Sub Saharan African 2.0 Zanzibar 2.7 Cameroun 1.8 Benin 1.4 Gambia 1.o Madagascar 0.8 M a l i 0.8 Niger 0.5 Source: World Bank 101. Civil service composition requires change. Chart 11.6 below outlines the distribution o f civil servants among different sectors. Education employs the largest share o f civil servants (around 44 percent), after public security (15 percent) and administration (12 percent). Despite the highnumber o f civil servants inthe education sector, analysis o f this sector suggests that the number o f teachers i s still not adequate (see education chapter for more details). Furthermore, the Malagasy military (including police and gendarmerie) accounts for over 20 percent o f the total civil service. This percentage seems fairly excessive for a country without any major external or internal threat. It i s unclear to what extend this distribution of civil servants across sectors contributes to achieving the goals and objectives o f the PRSP. 46For example the annual Anti-comption Perception Index issuedby Transparency Intemational 47 See among other reports ((Etude sur 1'Etat des Lieux du Regime Actuel de RtmunCration du Personnel de 1'Etat))Vol I& I1Cabinet R conseil Octobre 1998,and ((Etude sur 1'Etat des Lieux du RCgime Disciplinaire et de la Deontologie duPersonnel de 1'Etat)) Novembre 1998 OSIPD. Both studies were financed by the World Bank -47- Chart 11-6: Distributionof civilservantsby sector in 2004' (in% of total civil servants) Admin Defense Pub Sec Energy Agri Env GrOthr Educ Health SocOthr Source: WorldBank About 0.02percent of civil servants cannot be classified. 102. As in many other developing countries the Malagasy civil service is characterized by a relatively highnumber o f support staff, around 60 percent of the civil service in2004 (see chart II.7.b). At the same time, the intermediate level staff, including technical specialists who typically form the backbone o f a good civil service, account for only 30 percent o f the total civil service. The management level i s rather small with less than 8 percent o f civil service agents. This structure has not changed over the past seven years 1997-2004 (see chart II.7.a). It in part explains existing bottlenecks in delivering public services - not enough qualified personnel i s available to deal with the multiple tasks specified in the PRSP. -48- Chart 11-7 a b: a Evolution and distribution of the Malagasy civil service by functional level 1997-2004 - b Qualification of Civil servants by functional level a Evolution and distribution of the Malagasy civil service by functional level 1997-2004 (in% oftotal civil servants) IIntermate I 1997 1998 1999 Zoo0 2001 2002 2w3 xx)4 I Source: World Bank b Qualification of Civil servants by functional level (in% oftotal civil servants) lntermlevel Suppoct Staff Source: World Bank 103. Civil service highly urbanized. While urban centers account for only 25 percent o f population, they have in 2004 about 40 percent of public employees (see chart 11.8). The number of civil servants i s 2 for 100 inhabitants in urban centers whereas this ratio i s 0.75 for rural areas. This distortion undermines the vision o f the PRSP o f an administration which i s close to the people -49- ("administration de proximitk"). It constitutes another explanation for the existing service delivery problems -not enough civil servants are allocated to front-line services. Chart11-8: Geographicaldistributionof civilservice agents 1997-2004 (in% oftotal civil servants) 180.0 160.0 140.0 120.0 100.0 Total Urban Centers 80.0 m Secondary Urban Centers 60.0 40.0 20.0 0.0 I 1997 1998 1999 2000 2001 2002 2003 2004 I Source: WorldBank 104. Salary compressionrateJawed. Adequate remuneration, inparticular for higher level staff, poses a serious problem: The civil service salary scale favors support staff whose salary appears to be competitive with the private sector. Despite efforts undertaken since 1997 to reform the salary scale the compression ratio between the lowest and the highest salary i s only 6.7 in2004, while it usedto be 11.6 in the 1960s. A somewhat "adequate" compression rate would be in the range o f L o w remuneration o f technical and management staff i s one reason why Government i s not able to attract enough qualified personnel inkey positions. 105. Performance orientation low. There i s little performance orientation in the public administration. Job descriptions that typically outline objectives and performance expectations are lacking or outdated. Service standards (i.e. time to deliver a service) do not exist for most parts of the public administration. Administrative decision-making i s extremely hierarchical. Discretionary authority and informal behavior are high and key sources for mismanagement and corruption. General work ethics in the public sector are low; there i s widespread absenteeism, which remains unsanctioned. The morale in the public sector i s low, mainly because o f low remuneration and lacking career prospects4'. 106. Government reforms under way. The Government i s trying to address the low performance through a combination o f measures: Since 2003 all ministries are required to develop and submit business plans on an annual basis. It i s also envisaged to introduce performance contracts for the key managers in public institutions. In parallel, the Govemment i s changing administrative structures and processes in important ministries (e.g. transport and education) to improve overall organizational effectiveness. These reforms have already yielded some positive results, in particular in the education and transport sectors. The Govemment i s also evaluating the feasibility o f a comprehensive administrative reformprogram that i s aimed at changingthe existing institutional and procedural set-up. 48 See findings o f the following report: "Etude sur 1'Etat des Lieux du RCgime Actuel de R6munCration du Personnel de 1'Etat))Vol I& I1Cabinet R conseil Octobre 1998 49Civil servants rapidly reach the promotion limit, on average after 22 years, whereas a civil service career may last intotal 42 years untilretirement. -50- 107. Payroll system modernized. A new payroll system has been operational since 2001. The system i s managed by the Payroll Directorate o f the Ministry o f Finance (Direction de la Solde - DGS)." The system integrates the various components o f payroll management, which were previously split between different institutions. It allows Government to monitor earnings o f civil servants, to manage pension benefits, and to make projections o f future payment obligations. The present system has, however, three major shortcomings: (i)the database i s not comprehensive; neither military (representing 20 percent o f the civil service agents) and nor contractual agents paid under the non salary recurrent budget are regi~tered.~~Regular (ii) updatesofthepayrolldata,whicharedonemanually,take a long time (about a year) and follow complex administrative p r o c e d ~ r e sTo~address these problems, . ~ a more advanced version o f the system i s under preparation. This new system will also enable forecasting o f the wage bill. 108. Prioritization of human resources needed. During the budget preparation process no systematic review o f the staffing allocations o f the sector ministries takes place. Staff and payroll estimates are based on the staffing plans prepared by the line ministries in collaboration with the Ministryfor the Civil Service. Itis unclear how these plans are developed since atthe level ofthe sector ministries neither job descriptions nor a breakdown o f staff against goals and objectives o f the respective ministry or department exist. In a modem civil service ministries and departments would be requested to specify their goals and objectives for every organizational level and to allocate existing staff to identified tasks or objectives. The ministries and departments would also develop job descriptions for every "poste budgktaire". This information can then serve as the basis for staff and payroll estimates that would be discussed with the MEFB and validated by Cabinet. If certain tasks become obsolete or selected objectives are considered low priority the budget discussions should lead to an adequate staff reduction and to a reallocation o f existing staff to high-priority tasks. This would include the transfer to staff to other ministries and departments. Staff that cannot be removed immediately would be earmarked for future reallocation. Recommendations o Increasing the number of staff allocated to theperiphery. As resources should be realigned with policy priorities Government should increase the number o f civil servants allocated to deconcentrated line agencies and facilities. Government should consider introducing mandatory assignments outside the big urban centers over a period of one or two years for all civil servants as an important prerequisite for further career development. 0 Focus on reforming specific Government services. While substantial improvements o f efficiency, quality and performance orientation o f the public service are important for the successful implementation o f Madagascar's reform program the Government should learn from past failures and experiences in other countries in designing and implementing its reform strategy. These experiences overwhelmingly indicate that comprehensive reforms o f the civil service are extremely difficult to implement and sustain - many o f these reforms have proven ineffective in other countries. A more promising approach would focus on interventions with highimpact. These interventions should ideally target specific public services with a large public-private interface (such as primary education, health centers, customs, land titling and the lower courts). Objective would be to improve service delivery and j0 This systemreplacedthe previous one at INSTAT whose finction was limited to issuing the payroll statements. 5 1An additional complication is that the 1999 census o f civil servants found 7500 entries which could not be identified. These are not necessary ghost workers; in many cases the respective ministries engaged them without authorization and didnot register them with the payroll system. 52 The recruitment o f 4500 teachers in 2001, as part o f the HIPC savings, i s a good example - 73 teacher are still missingfromthe list. -51- produce visible results in the short term. While these reforms cannot substitute a long-term vision for the reform o f the civil service they keep the reform program focused on tangible results and, thus establish muchneeded credibility. Enforcing accountability. Absenteeism and informal behavior that i s not inline with the rules and regulations o f the public administration should be systematically sanctioned to improve overall work ethics. To assess public sector performance Government should consider the introduction o f citizen report cards that have positively impacted public sector performance in other countries (see box 11.10). Box 11-10 Citizens reportcards - a tool to asses public sector performance T o give voice to the recipients o f public services and to improve public sector performance citizen report cards were developed by the Public Affairs Centre in Bangalore, India. Citizens are asked to rate service access and quality and to report on corruption and general grievances about public services. Citizens' report cards have spread to cities in the Philippines, Ukraine, and on a pilot basis, Vietnam. The results have stimulated considerable media, bureaucratic, and political attention and acknowledgement of their contributionto service improvements. Report cards seem to have had a more direct influence on the heads and senior managers o f the municipal and utility agencies responsible for services, as inBangalore. The highvisibility o f report cards inthe press and civic forums turns them into league tables o f the efficacy o f municipal agencies. The reputational competition arising from the report cards i s enhanced by joint agency meetings discussing the report cards with attendance by prominent social andpolitical leaders and citizens. Source: The World Bank o Address civil service salary problems. To address difficulties in attracting qualified staff at technical and managerial levels the Government should consider the introduction o f a special scheme for key positions in the administration. These positions would receive a comparatively higher salary than the rest o f the civil service. In this context, Government can build on previous reforms which attempted to introduce the concept o f "postes clCs" into the public administration. Alternatively, the Government could recruit much needed personnel on a contractual basis without integrating them into the civil service. Under this model positions would be filled on a term and not on a permanent basis. This would increasethe flexibility o f Govemment to adjust staffing inaccordance to policy priorities. It will be important to carefully evaluate these options and determine their macro-economic implications. As an interim step Government shouldbegin to decompress the existing salary scale infavour o fhigher level technical and managerial positions inthe civil service. 0 Strengthening estublishment control. Government should ensure that all public sector employees are integrated into the payroll system; the widespread practice o f hiring o f public sector employees without approval by the Ministry o f Finance and registration under the payroll system should be discontinued. The payroll system should be subjected to regular reviews - to this end Govemment should randomly verify the payroll data in line ministries on a regular basis. Sufficient resources shouldbe set aside for these controls. 0 Subject human resources to budgetarypriorities. Line ministries should be required to prepare and regularly update recruitment and staffing plans as part o f the budget preparation and execution process. These plans should be subjected to Govemment priorities inseparate manpower hearings with the MEFBand the Ministryo f Civil Service. The final staffing plans shouldbe submittedto Cabinet by MEFB with a substantiation that the allocation o f human resources is in line with the goals and objectives o f the Government. -52- 111. USE OF RESOURCESINTHE EDUCATIONSECTOR 109. This chapter will analyze the use o f public resources in the education sector. It will give an overview on key policy issues and assess Madagascar's past expenditure performance inthe education sector. Additionally, it will focus on budget management issues, the organization o f the sector and community participation in service delivery. 1) KEYPOLICYISSUES 110. To assure that the entire population has access to quality education and to eliminate the negative effects o f the political crisis o f 2001-2002, the new Government committed itself to the "Education For All (EFA)" initiatives3and formulated a new strategy to reform the education system.54 This strategy is supported by the international community which has pledged to help Government reach its education goals. As a first step, the Government has put in place a set o f policies such as the elimination o f school fees in primary schools at the beginning o f the school year 2002-2003. This change inpolicy, along side with the distribution o f textbooks and school kitsto students and teached5, resulted ina dramatic increase inenrollment. 111. This large increase in enrollment, however, does not assure that the country can attain the EFA goal o f a 100percent primary completion rate. The education authorities need to make sure that those children that are still out o f school enroll, and that those already enrolled complete the five years o f primary. This will present many challenges to policymakers which are summarized below: 0 Coverage, accessfor poor and in remote areas. Madagascar has a large network o f schools. Today, there are about 19,000 primary schools, more than 80 percent o f which are public. These schools serve a population o f about 3 million children. Inremote areas many communities have created (built and financed) their primary schools, and are paying for the community teacher (FRAMteachers). Most of the out-of-school primary-age children come from the poorest families, and live in rural isolated communities. o Enrolment and survival. Enrolment inprimary education and new intakes in first grade have significantly increased during the past two years (see table 111.1).Administrative data from the Ministry o f Education and Scientific Research (MENRS) shows, however, that out o f 100 children who start primary school less than half completed their five years o f primary in 2003/2004. Assuming that only 82 percent56o f children in school age enroll in school it can be concluded that only about 35 percent complete the five years cycle o f primary in2003/2004. 53 Principal focus o f EFA i s primary education. Goal i s to attain a 100 percent primary school completion rate by the year 2015. 54 See Annex 9 for a description o f the education sector. 5s Effortthat received the financial support o f the World Bank project CRESED 11. 56Based on the EPM2001. The proportion of 15 to 19 year olds that never went to school. -53- Table 111-1:Number of new entrantsinprimaryschool Number of new GrossIntakeRate* NetEnrollment Percentageof childrenof Year entrants (in %) Rate**@%) 6 years old enrolled 00/2001 422,752 102 67 61 01/2002 577,83 1 136 70 66 0212003 778,041 178 82 80 0312004 886,513 198 97 94 Source: MINESEB 2000.2001.2002.2003., 2004. Statistical vearbook , . , *Total number of new entrants infirst grade ofprimary education, regardless of age, expressed as a percentage of thepopulation at the oficial primary school-entrance age (6 years) **Calculation is based on the demographic projection of the 1993population census and the school data survey. a Efficiency and cost-effectiveness. Efficiency remains a key problem o f the education system (see table 111.2). Of those who complete primary education, around 66 percent5' repeated at least a grade and only few o f them will continue to junior secondary education (44 percent). Cost-effectiveness i s low: To produce a 5th grader the Malagasy education system spends 2.7 times more than in a system without dropouts or repeated8. Comparable African countries, with a similar public budget allocation to education (interms o f percent o f GDP) performbetter at all levels of education. Table 111-2: SurvivalRate and GrossIntakeRateinlastgrade of primary inselected Sub-SaharanAfrican Countries PUBLIC SURVIVAL GROSSINTAKE RATE EXPENDITURE GDP RATETO INLAST GRADEOF ONEDUCATION per Capita GRADE5 PRIMARY AS PERCENTAGE OFGDP (US%) Year I N % Year in % Guinea 1.9 415 199912000 84 199912000 33 Zambia 1.9 361 200012001 77 200112002 60 Chad 2.0 240 200012001 45 200112002 27 Niger 2.3 190 200012001 71 200112002 21 Madagascar 2.6 268 2000/2001 34 2001/2002 36 Gambia 2.7 257 1998/1999 70 199912000 53 Rwanda 2.8 295 200012001 40 200112002 36 M a l i 2.8 321 200012001 84 200112002 31 Cameroon 3.2 699 199811999 81 2001/2002 56 Senegal 3.2 629 200012001 68 200112002 48 Benin 3.3 443 199912000 84 200112002 46 Source :WorldBank /SIMA - UNESCO/EducationDatabase a Regional disparities. Average enrollment and survival rates also vary significantly across regions and levels o f income. In 2001, enrollment rates in primary education in Antananarivo were almost twice as highas enrollment rates inToliara, where the rate was below 40 percent (see table 111.3) More strikingwas the difference between rural and urban areas5'. While enrollment rates inurban areas were above 70 percent in all provinces, enrollment rates in rural areas were lower than 50 percent. *'MENRS 2004, Statistical yearbook 58The Ministry o f Education is inthe process of finalizing a study o n monitoring student learning achievements. The study will also provide information on how budgetary resources affect the learning level o f the students. 59EPM2001 -54- Given that the majority o f the poor live in rural areas, this pattern shows an unequal distribution of the benefits of primary school. Table 111-3: Net enrollment acrossFaritany 2001 [in%] Antananarivo Fianarantsoa Toamasina Mahajanga Toliara Antsiranana Avg 2000-2001 Pre- 12 primary 28 6 10 7 12 4 Primary 77 63 62 51 38 64 60 CollBge 22 8 7 10 7 7 11 High 3 Sciool 10 2 2 1 1 1 Source: EPM 2001 0 Equity. There are also large income variations in primary and, in particular, secondary enrollment. Even in a period of economic growth, between the years 1999 and 2001, the distribution of benefits from education became more unequal. During that period, the enrollment rates o f children belonging to the poorest levels o f the income distribution decreased while the rates of children belonging to the highest levels increased (see chart 111.1below). Chart 111-1:Net Enrollment across quintiles in rural areas - 1999 -2001 Rimary net enrollment: rural areas Secondary net enrollment:rurd areas 90.0 , 35.0 7 80.0 70.0 30.0 60.0 25.0 50.0 20.0 40.0 30.0 15.0 20.0 x1.O 10.0 5.0 0.0 0.0 Source :Romani, 2003 -EPM 1999-2001 2) TRENDS EDUCATIONEXPENDITURES IN 112. Trends in public sector resource allocation. The Malagasy policymakers have increasingly focused on improving access to primary education. This new emphasis has been reflected in higher budget allocations to the education sector, which have significantly increased inthe past years. Over the 1997-2004 period, allocations to the education sector improved by an average annual rate of 23 percent, compared to an average annual increase of 14 percent for the non-interest expenditure envelope. The negative growth in2003 was due to a decrease inthe budget allocated to tertiary education and research -55- - the budget allocated to primary education increasedby 12percent inthe same year. The temporary cut inallocationwas more than compensatedby a 40 percent allocation growth rate of the sector for 2004. While in 1997the share of the budget allocatedto education was about 2 percent of GDP, the allocation increasedto nearly 4 percent of GDP in2004 (see table 111.4). Table 111-4: Madagascar- Education Sector: CentralGovernmentallocation and expenditureandprivate sector spending (selectedyears between1997 and 2004)60 Per Avp (in %) 1997- 2003- 1997 1999 2001 2002 2003 2004 2002 2004 CentralGovernmentBudget Allocation Total budget allocation (MGF bill) 341.6 573.9 865.3 1060.8 1004.0 1407.0 679.7 1205.5 Total allocation (% of GDP) 1.9 2.5 2.9 3.5 3.0 3.6 2.7 3.3 Total allocation (% o f total budget) 10.5 11.7 12.5 14.4 15.2 18.4 12.4 16.8 Annual growth in total allocation 20.2 14.0 22.6 -5.3 40.1 25.8 17.4 Allocation to Primary (% o f total educationbudget) 20.3 27.5 18.6 35.5 42.0 33.3 25.4 37.7 Expenditure Total expenditures (MGF bill) 355.1 645.0 676.2 673.8 973.9 578.6 973.9 Total expenditures (% of GDP) 2.0 2.8 2.3 2.2 2.9 2.3 2.9 Total expenditures (% o f total budget) 103.9 112.4 78.2 63.5 97.0 90.6 97.0 PrivateSector Private spending (inbillion FMG) 415.0 Private mending(% of GDP) 1.4 Memo Real Growth intotal allocation 4.8 4.8 -7.5 15.6 12.9 2.9 14.3 Real allocation 341.6 504.3 656.6 708.6 681.4 545.4 681.4 Real expenditure 355.1 566.8 513.2 450.1 660.9 475.8 660.9 Source: MEFB, EPM 2001, IMF estimate 6o The total allocated budget and actual expenditures for education in this table differ from to those shown in the public finance chapter. Data in this table refers to education expendituresin the whole budget under the functional classification, and hence include education related expenditures made by ministries other than MINSEB and MESRES. Table 111.5 includes only budgetaryallocationto MINESEBandMESRES. 56 113. This allocation is much closer to the Sub-Saharan Afixa average which was around 3 percent o f GDP in2000. However, countries that have less GDP per capita than Madagascar were spending more on their education sector (see chart 111.2). The higher budget allocation to the education sector in Madagascar is mostly attributed to higher allocations for primary education. These allocations increased from 20 percent o f the total education budget in 1997to 33 percent in 2004. Overall, these figures underpin the increasing priority the Government has attached to the sector -notably to primary education. Chart 111-2: Public expenditure ineducationas percentageof GDP 2000 - selectedSub-Saharan African Countries- Ethiopia (123) Togo (320) Ghana (428) Malawi (157) Comros (433) .-mP h Burundi (143) +a m Benin (442) 0 Mali(321) P ij FWanda (294) n 0 Gambia (256) E. Eritrea (149) +a F Madagascar (267) a E Uganda(359) 0 0 Niger (189) Guinea-Bissau (140) Chad (240) Zambia (360) CentralPhfrican Republic (273) Guinea (415) 0 1 2 3 4 5 6 Source: MadagascarMENRS; Statisticalyearbook 2001/2002-2002/2003; UNESCO Note: For the selected countries, the GDPper capita is less than US$500. 114. Trends inprivate spending. Private expenditure ineducation i s an important source o f financing for the sector. In 2001, spending by private households represented about MGF 415 billion, or close to 1.4 percent o f GDP.61Nearly half o f these expenditures were spent on primary education. 3) BUDGETMANAGEMENT ISSUES 115. A larger education budget, however, does not necessarily improve access to primary education to the poor. Several problems negatively impact the increase o f budgetary resources and the pro-poor distribution o f the budget: Calculationbasedonthe EPM2001. -57- 116. Higher primary education budget absorbed by increase in student enrolment. Despite strong government efforts to allocate more resources to primary education (from MGF 159billion to MGF 332 billion), expenditures per student (as a percentage o f GDP per capita) remained at the 2000 level (see table III.S).This i s due to the fact that student enrolments increased considerably as a result of the elimination o f the school fees. Budget allocations to secondary education also increased by 74 percent for the same period to MGF 169 billion. Expenditure per student for secondary education augmented by 51 percent suggesting that students in secondary education benefited by the increase o f public resources, despite the drop inthe share o f resources allocated to secondary education. Table 111-5: Primary and secondary expendituresin2000 and 2003 Growth 2000 2003 2000/03 (in%) Expenditureto Primarvand Secondarv Primary (billions MGF) 159 332 109 Secondary (billions MGF) 97 169 74 Primary (in % oftotal budget to primary & secondary) 62 66 6 Secondary (in% oftotal budgetto primary & secondary) 38 34 -10 Expenditure per Student to urimarv and Secondarv Total expenditureper primary (inMGF) 265,811 312,967 18 Total expenditureper secondary (inMGF) 316,422 478,430 51 Percentageof GDP per capita to primary and secondary Primary 15 15 Secondary 28 46 64 117. Poor performance atprimary school level. Repetition rates at primary school level in Africa (see chart 111.3). On average children 8 to 15 years old repeat a year inprimary 1.3 times - Madagascar have been around 30 percent inthe last years; one o f the highestrates inSub-Saharan only 4 percent o f children report no class repetition and 25 percent repeat two or more times during primary school. The Government has taken steps to address this problem. A new policy will automatically promote children inside the three primary school sub-cycles62. It will be important to closely monitor the impact o f these changes. Primary school i s composed by three sub-cycles: the preparatory courses or first two years of primary, the elementarycourse or third year, andthe middle courses or last two years ofprimary. -58- Chart 111-3: RepetitionRatesfor selectedcountriesin Sub-Saharan Africa in2001/2002 (in%) 40 35 30 25 20 15 10 5 0 Source: Madagascar MENRS 2004, Statistical yearbook;.UNESCO Note: For the selected countries, the GDPper capita is less than US$400 118. Inadequate allocationsfor non salary recurrent expenditures. Over the period 1997- 2004 salaries represented the largest share, on average 67 percent, o f the allocations for primary education (see table 111.6below).63 The budget for utilities in the education sector over the same period i s very low - it represented less than 1percent of the total recurrent budget in 2004. The budget allocated for goods and services i s equally small ranging from 5.4 to 13.3 percent o f the total primary education budget between 1997 and 2004.64 The drop o f spending on goods and services from 7.5 percent in 2003 to 5.4 percent in 2004 i s due to the introduction of a special cash transfer to the schools in 2004. On the whole, the indicators suggest that resources for maintenance are not adequate, which explains the poor conditions inmany schools. 63This analysis is based onbudgetary allocations after 2001. Only after this year it was possible to classify teaching expenditure as expenditure to the schools and not to the CISCOs. 64Total resources for maintenance are higher as communes also finance these costs with own resources. In addition, the funds of the "caisses ecoles" canbe used to cover maintenance costs. -59- Table 111-6: Budget allocationto primaryeducation(selectedyears between1997-2004) (in%o ftotal primary educationbudget) TOTAL 100 100 100 100 100 100 100 Current Expenditures 71.4 73.2 67.0 67.8 83.9 71.5 76.0 Salary 59.7 59.8 47.8 52.2 68.3 52.2 66.8 Goods and Services 7.9 9.8 13.3 8.8 7.5 5.4 8.5 Utilities 0.7 1.1 0.6 0.4 0.5 0.0 0.2 Transfer 3.2 2.5 5.3 6.4 7.6 13.1 0.5 Capitalexpenditures 28.6 26.8 33.0 32.2 16.1 28.5 I 24.0 Source: MENRS 119. Inadequate mix of recurrent and capital expenditures. As shown in table 111.6, the average spending, over the 1997-2004 period, on capital investment for primary education i s around 24 percent o f the total primary education budget while recurrent expenditures amount to 76 percent o f the total budget. Over the next years, an increase in recurrent spending will be necessary to (i) ensure adequate teacher training and recruitment for achieving EFA by 2015, (ii) finance and reinforce FRAM teacher salaries, and (iii)to improve district, ZAP and school service delivery. Given these needs, Government may wish to increase recurrent expenditures, either by allocatingmore funds to the sector or by prioritizing recurrent over capital expenditure. 120. Administrative overhead too high. The total administrative recurrent expenditures decreased significantly from 61 percent in 1997 to 35 percent in2003. Government's expenditure on administration, however, remain relatively high (see table 111.7). Key reason i s the relatively high number o f staff, in particular at the central level, that results in a ratio o f about 1 administrative staff to 10 teachers (see annex 10 - table 10.1). In addition, the proportion o f the salary budget allocated to administrative staff at the central level (20 percent in 2003, see table 111-7) compared to allocations for teachers (65 percent in 2003) seems to be disproportionately large. Allocations o f non salary expenditures are also in favor o f central administration (36 percent in 2003) compared to primary and secondary schools (29 percent in 2003). The share o f non salary expenditures allocated to primary schools increased after Government's decision to provide direct school grants. Transfers to others institutions (private and public institutions, teachers training institutions) are relatively important and represent more than one third o f the non salary recurrent expenditures. The Govemment intends to address the problem o f relatively highadministrative overheads bydevelopinga humanresourcesreformplan. -60- Table 111-7: Allocation of recurrent budget expenditures to administration, primary and secondary schools and transfers selected years between 1997-2003 (inpercent) 1997 2001 2003 Non salary `on salary Non salary recurrent ecurrent recurrent Salary expenditures Total Salary yenditures Total Salary expenditures Total TOTAL 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Administration 62.1 55.5 61.2 62.6 73.1 65.1 34.5 36.2 34.8 Central 12.2 33.5 15.3 13.0 24.2 15.6 20.4 21.4 20.5 DIRESEB 3.7 11.8 4.9 4.2 9.8 5.5 5.8 9.2 6.3 CISCO 46.1 10.2 41.0 45.5 39.1 44.0 8.4 5.6 7.9 Primary schools 25.1 9.2 22.8 24.8 14.7 22.4 44.0 25.2 40.9 Secondary school! 12.5 7.1 11.7 12.4 4.0 10.4 21.3 3.4 18.4 Transfers 0.3 28.2 4.3 0.2 8.3 2.1 0.2 35.1 5.9 Source: ME1`B 121. Budgetplanning lacks realistic estimates. Until2003 budget planning inthe Education Ministry was based on the allocations for previous budget years. In 2004, the Government introduced program budgets for all ministries including education. For the education sector, budget preparation i s basedon cost estimates o f the PRSP education business plan and the "EFA" plan. Although the Ministry o f Education i s the most advanced in terms o f program costing compared to other ministries, both documents are not well harmonized and lack prioritisation o f policy actions in line with realistic cost estimates. At the district level no clear guidelines exist to determine within-district allocations; qualitative evidence fiom the year 2003 PET study suggests that planning criteria (e.g. regarding the distribution o f supplies to schools) are not consistent and differ quite strongly between district^.^^ 122. Poor performance in investment expenditures. The chart 111.4 shows that during the 1997-1999 period the education budget was fully implementedreachingan average execution rate o f more than 100 percent. Between 2000 and 2002 the execution rate declined significantly with an average annual rate o f 60 percent. This i s mainly attributed to the low performance o f the investment expenditures which achieved only a 50 percent execution rate. The year 2003 saw a substantial improvement in the total execution rate for the budget o f the education sector, which reached 97 percent execution rate, and was well above the execution rate for the national budget (88 percent). 65 See annex 10, table 10.2 -61- Chart 111-4: Madagascar Executionrate of the Educationbudget 1997-2003 - (inpercent) 160 140 120 100 80 60 40 20 0 II 1997 1998 1999 2000 2001 2002 2003 --tExecrateoftotaleducexp-c-Execrateofcurrentexp Exec rate of capital exp Source :MEFB 123. Madagascar's investment budget in the education sector i s financed by a variety o f sources, including general revenues and external loans. External resources represents about 70 percent o f the education investment budget. The chart 111.5 shows that execution o f external and internal resources decreased during the 2000-2002 period. The decrease o f the education investment budget execution was mainly due to poor external resources execution; internal resources execution depends mostly on external resources disbursement. While investments funded by own revenues attained a high execution rate, the execution rate o f external assistance was low (7 percent in 2000 and 6 percent in 2001). Several factors contributed to the low execution o f the externally funded investment budget: (i) poor procurement and management capacity in the ministry o f education, (ii) poor capacity in planning the investment budget (iii) complicatedmanagement and disbursement procedures regarding donor funded projects, and (iv) weak coordination between MENRS and MEFB in preparing the investment budget. The situation improved significantly in 2003 with an increase o f the execution rate o f the externally funded investment budget by 152 percent 66 66 The execution ratei s largely over 100%because disbursement was greater than the allocated amount. -62- Chart 111-5: Executionrateof investmentexpenditures per source of funding 1997-2003 (inDercent) I +-Total investment - Internalresources External resources Source :MEFB 124. Regressive distribution of public resources. Public expenditure in primary education benefit the richer households proportionally more than the poorer ones. A 2001 benefit incidence analysis of public expenditures to primary, college, and high schools in the year 2001 confirms this trend, which constitutes a strong contradictionto Government's objectives to improve access of the poor to primary education6'. As shown in chart 3.4 children belonging to the poorest 20 percent of households receive less than 20 percent o f public expenditures. Similarly, children living in the poorest 40 percent of households also receive less than 40 percent of the expenditures (see table 111.8 and chart 111.6). 67EPM 2001 63 Table 111-8:Percentageof total public expenditure onprimary and secondaryschool across income quintiles Income Quintiles I I1 I11 IV V Elementary 14.0 17.8 22.1 30.6 18.5 College 2.6 7.1 14.2 29.8 48.3 High School 0.3 1.6 2.9 19.8 78.0 Source: MEFB. EPM 2001 Chart 111-6: Benefitincidenceof public expenditurein education2001 1.oo 0.80 0.60 0.40 0.20 0.00 0.00 0.20 0.40 0.60 0.80 1.00 +EPP +CEG t-Lycee 45 degree line Source: Calculations based on expenditure datafrom the MEFB and income quintilefrom the EPM 2001 - EPP: Primary Public School - CEG : Public lower secondary school - Lycke : Public upper secondary school 125. The regressive distribution i s largely due to the unequal distribution of Government financed teachers (as opposed to those teachers financed directly by local communities) across the country. The table 111.9 below shows the distribution ofprimary teachers paidby Government across four zones for the school year 2002-03 and 2003-04. The zone classification i s based on the accessibility and the remoteness o f the areas. The data underlines Government's difficulties to hire and maintain teacher in remote areas. Publicly paid teachers are much less allocated to zones 2 and 3 where access i s difficult. To address this problem, Government recruited under the HIPC program 3,427 teachers on a contractual basis in 2001 for public primary schools. Most o f them (82 percent) were assigned to remote areas. In addition, new financial incentives have been implemented for teachers to serve in rural public primary schools.68 Overall, the studentlpublic teacher ratio remained highwith 57 and 59 inremote areas (zone 2 and 3) in 2003-2004. This i s mainly due to strong increase in student enrollment inboth zones between the school years 2002- 03 and 2003-04. Inearly 2003, MENRSrevisedthe rates o f hardship allowances on the basis o f4 zones. Inaddition, for the first year (2004) o f the EFA plan implementation, the construction o f 110 houses for teachers inremote areas i s under way. -64- Table 111-9: Distribution of primary teachers across zones 2002/03 and 2003/04l 2002-2003 Parents Ratio of Number Publicly Ratio of association pupils to of paid pupils to paid all ZONE' students teachers teachers teachers teachers 0 140,760 3,360 42 164 39.9 1 628,737 11,013 57 1,838 48.9 2 228,287 9,570 24 2,015 19.7 3 876,559 14,567 60 3,090 49.6 TOTAL 2,274,443 38,509 59 7,107 49.9 2002-2003 Parents Ratio of Number Publicly Ratio of association pupils to of paid pupils to paid all ZONE' students teachers teachers teachers teachers 0 128,909 2,612 49 137 46.9 1 588,455 11,291 52 2,519 42.6 2 707,569 12,500 57 3,338 44.7 3 1,143,778 19,476 59 6,191 44.6 TOTAL 2,568,711 45,879 56 12,185 44.2 126. Other distributional issues in education. Two additional distribution issues are on Government's agenda: First, Madagascar is seeking to arrive at an acceptable formula to channel budget allocations to education districts (CISCOs) throughout the country. At present allocations are based on the previous year's budget only. It would be desirable to allocate on the basis o f need, as determined by the number o f students or the conditions o f the school infrastructure. In addition, there are concerns about the distribution o f resources to the various schools within a CISCO district. It seems that the poorest Fivondronanas in each Faritany receive fewer resources than the richer ones (see annex 10, table 10.3). The development o f more objective criteria for the allocation o f funds by the CISCOs would ensure that all Fivondronanas in each Faritany receive adequate funding. 0 Second, public and private schools in rural areas receive the same per student subsidy regardless o f whether or not they offer services to the poor. The new policy to waive school fees was extended to all private schools in rural areas which now receive a direct subsidy from the Government to cover the cost o f all student fees. The basic assumption was that many private schools offer services to poor people, and that some o f these schools are in isolated areas where no other schools are present. At the moment, there i s no data available to assess enrollment o f the poor inthese schools after the fees were waived. Government should verify the initial assumption to ensure that resources reachpredominantly the poor. -65- 4) SECTOR ORGANIZATION 127. Even if policymakers are committed to make education services "work for the poor" their success will depend on how effective their commitment i s "passed-on" to service providers. The Government has made significant efforts to strengthen the institutional framework for the delivery o f services and to clarify the various roles and responsibilities. Following a series o f institutional reforms that started in 1995, the education ministry i s today the most deconcentrated Government service in Madagascar. More than 60 percent o f the ministry's resources are allocated to deconcentrated line agencies and facilities. There are three levels o f deconcentration o f the education ministry - the provincial administration (DIRESEBs), CISCOs at the district level, and Administrative and Pedagogical Zones (ZAPs) at the commune level (see annex 9 for more details). 128. No clear division of responsibilities between thepolicymaker andprovider functions. According to the institutional framework set up by the Government, the CISCOs at the district level are mainly in charge o f the provision o f primary education services. The main functions o f the central level - the ministry o f Education including deconcentrated services at the provincial level (DIRESEBs) - i s to regulate, supervise, and design policies for the sector. In the past, however, the central level retained major delivery hnctions: it directly managed teachers and controlled all investments for the schools. CISCOs had little leverage to influence school performance. Since 2003, the Ministry o f Education has began to strengthen the roles and responsibilities o f the CISCOs: 3,500 teacher positions were created and distributed by the CISCO according to school needs. A new statistical system was. put in place to improve the distribution o f administrative positions among CICSOs and the development o f teacher recruitment criteria for CISCOs. Finally, CISCOs have played an important role in the planning and construction o f schools inthe context o f the EFA strategy implementation. 129. The division o f responsibilities across levels inthe ministryor between different public agencies outside the ministry, however, continues to be flawed. The division o f responsibilities between CISCOs, FAFs (community-government organisations), and communes regarding non- recurrent expenditures at the school level remains unclear. While ZAPs carry out important functions on behalf o f the CISCO their formal responsibilities are not clearly defined. The CISCOs are in charge o f managing the non-recurrent schools budget, which includes resources for operations and maintenance; in parallel the communes are in charge o f school maintenance. The FAFs add to the confusion as they also manage a small school budget for maintenance. 130. CISCOperformance has improved. Inthe past budget executionrates o f the CISCOs have been below the rates o f the central ministry and o f the DIRESEBs (see table 111.10). Between 1997 and 2001 the average execution rate o f the CISCO'S was about 85 percent in comparison to the MENRS which reached an average execution rate o f 92 percent over the same period. The low rates at the CISCO-level reflected capacity problems o f the district administrations to adequately perform theirs functions. To address performance problems o f the CISCOs the Government launched (in 2003) measures to improve capacity o f the district level staff such as: (i)under-performing directors o f CISCOs were replaced; (ii)new equipment (including computers, and transportation equipment) was distributed to the districts; and (iii) training inpublic finance and procurement aimed at strengthening efficiency and accountability at the CISCO level has been provided. As a consequence, the execution performance across all administrative levels improved significantly in 2003 reaching an overall execution rate o f non salary recurrent expenditures o f nearly 100 percent. The slightly lower execution performance o f -66- the CISCOs in2003 (around 91 percent) i s mainly due to (i) ongoing organisational changes, and (ii) complex budget execution procedures that challenge the existing capacity at the district level. Table 111-10: Executionrates (inpercent) of nonsalary recurrent expendituresby educationlevel (selected years between 1997-2003) Avg 1997- 1997 1999 2001 2002 2003 2001 TOTAL 93 96 90 60 97 90 MENRS 93 96 76 53 98 92 DIRESEB 94 97 96 45 98 90 CISCO 78 94 95 52 91 85 High School (Lycee) 95 95 94 52 93 89 Middle School (CEG) 93 95 98 54 93 89 Primary School (EPP) 97 97 100 55 93 91 Pre-school (EN1) 92 79 90 60 87 83 Transfers* 98 99 94 84 100 97 Source :MEFB * Transfersare managed by Central Ministry 131. Direct resource leakages. Three expenditure tracking surveys confirmed that resources do not always arrive at the intended re~ipient.~'Almost 75 percent o f the schools reported that they had not received equipment and funds that were allegedly sent to them by the CISCOs. The average gap between expenditures reported by the CISCO and those received by the schools amounts to 16 percent7'. The 2003 survey found leakages o f in-kindtransfers in 55 percent o f the schools.71These rates, however, differed largely between the provinces (up to 87 percent and 77 percent o f schools with leakages inFianarantsoa and Toliara and no leakages at all inthe provinceofToamasina). 132. Theproblem of overpricing. There i s evidence that the CISCOs pay more for supplies than the market price. The 2003 survey found that for school supplies such as pens, notebooks, and glue the CISCOs paid 20 percent, 37 percent, and 82 percent more than the market price. Similarly, another study72found large price gaps in the invoices paid by MINESEB for certain articles. The gap came on average to about 50 percent. In the case o f technical material and equipment, which was mostly procured at the central level the gap increased to about 100 percent. 69 The first two were carried out in 1999 and 2001 by the National Statistics Institute INSTAThfADIO, with financial and technical support o f the EU (INSTATProject Madio 2003). The third one was commissioned by the World Bank inM a y 2003. The INSTAT/MADIO survey covered 143 schools and 24 o f the 111 school districts. The Bank's tracking survey covered 185 schools and 24 districts 70 Five percent o f the schools surveyed reported receiving only a quarter o f the amounts reported by the CISCO. 71 These discrepancies included both differences in the number and type o f supply reported, i.e. in 28 percent o f the schools the amount reported by the school and the CISCO i s different, in the rest o f the schools the reported material was different. 72 INSTAThfADIO, 1999,2001 -67- 133. The problem of bureaucracy and oversight. CISCOs are subjected to complex and bureaucratic accounting and procurement proced~res.~~These requirements reduce the time available for the district's core functions: technical supervision and planning. Downward financial accountability from the CISCO to the facility i s almost non-existent. The head o f the CISCO handles almost all procurement and financial management issues o f the school's non- salary recurrent expenditure. School facility staff have no oversight over appropriate prices and payment procedures usedby the CISCO. 134. Management information not readily available. Every year MINESEB collects standard information on all primary schools (both public and private), teachers and students inthe country. This information i s vital to monitor the performance o f the schools and o f the education system in general. It allows policymakers to assess progress in reaching the education sector goals, and it provides necessary information to design or adjust education policies. Important data on schools and teachers, however, are not collected; in particular data on teacher absenteeism. In for instance, was based on data from the 2001-2002 school years - information about the large addition, data i s not produced in a timely manner to feed into policy decisions The budget 2004, increase inenrolment was missing. This resulted ininsufficient allocations o f budgetary resources for teachers and classrooms that were neededto accommodate this new influx o f students. 135. Private schools perform better than public schools. The public sector i s the major provider of primary education inMadagascar. Around 77 percent o fthe schools were managedby the public sector with around 80 percent o f the students in the school year 2002-2003. The majority o f the private schools are runby religious group^.'^ Some indicators suggest that private primary schools are better managed than public ones. As shown in table 111.11 below, private schools have a lower studendteacher ratio (on average 34) than public ones (on average 55) during the last four school years 2000/01 to 2003/2004. Similarly, repetition rates in private schools are lower than inpublic schools. Inthe present school year 2003-2004, the percentage o f repeaters inpublic primary schools was close to 33 percent while that o f private schools was only 16 percent.75A household survey is currently under way to help identify the main factors o f student performance inpublic and private schools. Table 111-11: Studentkeacherratio inprimaryschool acrossyears 2000101 2001102 2002103 200312004* PublicPrivate Total Public Private Total Public Private Total Public Private Total 53 32 47 52 35 47 59 35 52 56 36 50 Source: MERS 2004. Report FER `These results are based on data from 106 CISCOs out of 111. 73 See for more information annex 9 74 See annex 10, table 10.4 75 Data from MINESEB. 2004. FER survey report. Performance differences between public and private schools could be explained by the different income levels o f students - students from high income families mainly attend private schools. These schools could therefore benefit from private contributions and stronger involvement o f parents inthe management o f the schools. -68- 5) COMMUNITY PARTICIPATION 136. International experience indicates that effective involvement o f communities or parental associations (FRAM) can improve the delivery o f education services. The FRAMs have been operational for a number o f years and have played an important role in the schools in Madagascar. One o f their most important activities i s the hiring o f (non civil service) teachers. The FRAMs also pay the salary o f these additional teachers (or supplement CISCOs' payments), procure supplies to the schools, organize maintenance work, provide food to students and accommodation for teachers. All these activities are mostly financed through parental contributionsto the FRAM. 137. In2002, the Government decidedto directly transfer school resources to newly created community-Government organization (FAFs). These organizations were supposed to bring together the parents, the headmaster o f the school, teachers, students, and community authorities. It was assumed that the transfer o f management responsibilities to the FAFs would improve the delivery o f services. In practice, the majority o f the FAFs lack community and student participation. A brief survey76also found that the quality o f the FAF dialogue remains poor - in only 5 out o f 20 schools which were interviewedthe parents were well informed o f the decisions taken by the FAFs. In contrast, the FRAMs' decisions are taken in open meetings and thus are perceivedto be more transparent. 138. In the last two years the number of teachers hired by the FRAMs increased rapidly, mainly because Government recruitment was not able to keep up with rising student enrolment ratesa7'Although these teachers made it possible for the schools to absorb the significant influx o f new students their existence raises a number o f questions concerning quality and relationship with the civil service teachers. The FRAMteachers have a different set o f incentives than those paid by the line ministry. A recent budget tracking survey7* found that the salaries o f FR4M teachers are below the minimum wage (US$ 22 per month for 7 months, compared to about US$60-70 for a public-budget paid teacher for 12 months). In addition, they do not receive benefits such as health insurance or any other social or pension scheme. FRAMteachers are often from the area where the school i s located; consequently they are more likely to stay on the job. Experience indicates that one o f the major advantages o f FRAMteachers i s less absenteeism. As they are hired and paidby the FRAMsthey are also directly accountable to them. 6) RECOMMENDATIONS 139. Over the last years the Government has significantly invested in improving access to education services to the poor. The Government's commitment i s reflected in higher budget allocations to education (from 2 percent o f GDP in 1997 to nearly 4 percent o f GDP in 2004); these additional resources focused primarily on primary education. Positive results include the considerable increase o f enrollment in primary schools over the past two years. To further improve the use o fpublic resourcesreforms should focus on the following key areas: a Availability of management information should be improved. While the ministry collects standard information about the performance o f the school system inthe country this data i s not produced in a timely manner to feed into policy decisions. Budget preparation i s typically 76Lesrelations entre le FRAM, le FAFet l'kcole, Josiane Rarivoarivelomanana, Florence Miauton- 2004 '*Francken, 77 See annex 10-table 10.5 2003 -69- based on outdated information, which poses significant problems in adequately determining budgetary needs. Review allocation of resources. Despite the increases inallocations to primary education, the per student allocations remained almost unchanged between 2000 and 2003, with 15 percent of the per capita GDP. This i s due to the significant increase in enrolments over the same period. The per capita allocations will needto be improvedto meet the targets o fthe EFAinitiative. o Allocations for recurrent non wage related expenditures are insufficient. To improve the conditions o f the schools a further increase o f allocations for maintenance and utilities should be considered. New investment projects should be only approved if they are financially sustainable (which implies that Government i s able to finance medium to long-term cost implications). Address implementation problems. To address execution problems o f externally financed investment projects existing bottlenecks should be clearly established and addressed on the basis o f a time-bound action plan. To improve the absorptive capacity o f the education administration a comprehensive training and capacity building plan should be developed that focuses predominantly on public financial management. o Strengthen the role and capacity of the CISCOs. Budgetary performance o f the CISCOs should continue to be systematically strengthened. To further improve efficiency o f the CISCOs the remaining management and implementationproblems (e.g. leakages o f in-kindtransfers) will need to be addressed. It will also be important to rethink the distribution o f budgetary resources among the districts and communes on the basis o f needs, as determinedby the number o f students or the conditions o f the school infrastructure. 0 Redistribute the teacher to remote areas. The allocation o f teachers to remote and poor areas should be increased. The poorest districts are those with the largest student/public teacher ratios. While this unequal distribution i s to some extent substituted by a significant increase in privately paid teachers the Government should develop a credible strategy to systematically deploy publicly paid teachers to remote and poor areas. First important steps have been taken to establish an incentive system in form o f salary bonuses for teacher to work inremote rural areas. This approach should be further developed. Government should also consider to hire new teachers locally and to transfer control over these teachers to the communes or the FRAMsystem. Improve management and organizational structure. A clear division o f responsibilities across the management levels o f the education sector i s needed as well as a clear division o f labor between the ministry and other agencies involved in the sector. Delivery functions such as the management o f teachers and the control o f investments should be transferred fkom the central government to the service delivery levels. In this context, the relationship between CISCOs, ZAPS,FRAMs, FAFs and communes should be clarified to ensure an effective service delivery system without duplications, overlaps and inconsistencies. 0 Reinforce effectiveness of community organizations. Government should further capitalize on the existing system o f community participation. In Madagascar, the FRAMs (parental associations) play an important (and widely acknowledged) role in school issues. They provide in-kindsupport for the schools, and are inparticularly involvedinschool maintenance. In some communities these associations hire teachers and pay their salaries. The role o f the F M s has been challenged by the creation o f FAFs (community-Government organization), which manage financial transfers to schools from the central Government ("caisse d'kcole" program). -70- The creation o f the FAFs raises questions about their relationship with the FRAMS; this relationship should be clarified to ensure effective coordination o f the support to the schools provided by the various actors. -71- IV. USE OF RESOURCESINTHE TRANSPORT SECTOR 140. This chapter takes a closer look at the Government's reform agenda in the transport sector and assesses the adequacy o f finding inthe sector. To this end, the chapter will provide an overview o f the key policy issues and focus on the sector organization and governance. It will review public expenditures and discuss issues in several sub-sectors (road, air and rail transport, as well as maritime transport and port management). 1) KEY POLICYISSUES 141. I n the past, the transport sector was state run. Historically, Madagascar's transport sector has been dominated by direct state intervention and heavy reliance on parastatals and/or Government agencies to provide services through the 1990s. With the exception o f road freight transport, Government had, directly or indirectly, managed and financed railways, aviation, airports, shipping, and ports. As a result, today's service i s poor, operating efficiency i s low, safety standards are inappropriate, maintenance has been neglected, and funding for capital investments has been difficult to mobilize". 142. Government's reform agenda. In April 2000, the Government adopted a comprehensive transport sector policy and strategy, which seeks to reverse past policies. It aims at: (i) focusing the Government's role on strategic planning, sector oversight and coordination; (ii)creating jointly controlled public-private and user-financed agencies for sub-sector management and regulatory functions; (iii) divesting operational activities to the private sector, through privatization and concessioning arrangements; (iv) developing local private sector execution capacity; and (v) rehabilitating transport infrastructure to appropriate levels. 143. Inimplementingits new vision, the Government is facing important challenges which are summarized below: 0 Poor quality services in the road sector. In 2001, the road sector accounted for 90 percent o f the domestic volume o f traffic. Road density i s low with only 0.05 km o f roads per square km8o Over the past 30 years, road quality inMadagascar has deteriorated: the country has lost about 1000 kilometers o f roads per year primarily due to poor maintenance. Today, 81 percent o f roads are inpoor condition. The share o f paved roads i s 11.6 percent, compared to an average o f 13 percent for Sub-Saharan Africa, 16 percent for low income countries worldwide, and 53 percent for lower-middle income countries.81The table N.1below specifies the degree o f accessibility o f the population o f Madagascar. It indicates that, although there are regional variations, on average, only about half o f the population i s accessible by motorized vehicle. Some o f the highest poverty rates can be found in areas that have the lowest access to motorized vehicles. ''Security standards for air transport, for example, are not conform with intemational standards as recommendedby the IntemationalCivil Aviation Organisation. This is, however, more than the Sub-SaharanAfrican average o f 0.034 km, but far low the average o f 0.138 Km for Mediterraneancountries. World Bank(2004) World Developmentindicator. -72- Table IV-1: Access of population by vehicle type (1999) and poverty rates (1999) Province Poverty Access by Population. (approx. location) rate motorized vehicle' Antsiranana (North) 72.6 30 9 Mahajanga (Northwest) 76.0 16 10 Toliara (West & South) 70.7 20 14 Antananarive (Center) 61.7 76 29 Fianarantsoa (Southeast) 81.1 45 20 Toamasina (East) 71.6 80 17 Total 71.3 52 100 Correlation Coefficient -0.47 between poverty rate and access Source: Government of Madagascar and World Bank estimates IShare of population that has access by motorized vehicle. 0 Air transport system in urgent need. Madagascar has 4 controlled airports, 18 airports with flight information services, and 34 simple airports. Domestic traffic came to 500,000 passengers in 2001. International air travel, long haul and regional, came to about 368,000 passengers in 2001, compared to 260,000 in 1997. The increase i s largely owned to the development o f the tourism sector, but also to foreign investors and took place despite the relatively high travel costs to Madagascar. Most airports require urgent rehabilitation to bring infrastructure to international standards, in particular with regard to safety and security equipment, telecommunications, radio navigation, expansion o f terminals and runways, road access improvements. 0 Port services inadequate. Of the country's 17 ports, only 4 are deep enough to permit unloading from large vessels, while the others require unloading offshore and transport by small boat. Cargo handling technology in the country's main port, Toamasina, i s outdated: the port unloads containers at a rate o f 8 per hour compared to a rate o f 25 per hour in neighboring Reunion and 30 per hour in neighboring Mauritius. On average, a vessel spends 3 days in Toamasina compared to 8 hours inMauritius. 0 Mixed picture concerning transportation spending. The Malagasy economy spends relatively more than other poor countries on transportation: the transport sector absorbed 15 percent o f GDP in 2001 in Madagascar compared to an average o f 6.4 percent o f GDP spent by low income countries over the years 1995-2002 (see table IV.2). At the same time, Government spending for the transportation sector inMadagascar i s lower than comparative spending inother low income countries (on average, 1.5 percent o f GDP in Madagascar during 1997-2003 compared to 5.7 percent o f GDP for other low income countries during 1995-2002). This indicates that relative to other countries the private sector i s spending more on transportation than the public sector. In Madagascar, since the 1990s, the private sector has owned and operated public transportation in the road and rail sectors, which account for the bulk o f the country's transportation services. -73- Table IV-2: The Size of the Transport Sector Public and private sector spending on Public Sector spendingon transportation transportation (value added as % of GDP) (as YOof GDP) Madagascar Low Inc. Madagascar Low Inc. 14.7l 6.42 1.22 5.73 Source: MEFB, World Bank '2001, 1997-2003,'1995-2002 ' Madagascar's transportationgoals ambitious. The PRSP sets transportation as a critical pillar o f Government's strategy to alleviate poverty for the 2003-2006 period. It aims to (i) restructure the RoadMaintenance Fund, (ii) disengage Government from the operation o f the country's airports, (iii) turn over the rail system (two lines) to concessionaires, complete rehabilitation and reach full operation (iv) invest heavily in the four major ports. For the road sub-sector, it i s envisaged to complete rehabilitation and/or construction o f 16,900 km o f primary and secondary roads and 8,000 km o f rural access roads over the 2003-2006 year period. These objectives should be compared to the average rate o f road construction and maintenance over in 2000 and 2001, which was about 500 kmper year. Investment needs enormous. The PRSP assumes for the 2003-2006 period investments o f about MGF 7 trillion, or close to US$1 billion (with the exchange rate used in the PRSP). This represents about 45 percent o f the total investment needs identified in the PRSP. 2) IMPROVING SECTOR ORGANIZATION AND GOVERNANCE 144. Poor sector organization until 2001. Until 2001 the transport sector had been managed by three separate ministries: Public Works (roads); Regional Planning (urban), and Transport & Meteorology (other infrastructures and services). Each ministry was responsible for the management and regulation o f the transport sub-sectors it was governing. Public enterprises were incharge o f the provision o f services. Lack o f coordination, outdated regulatory frameworks and operating standards as well as minimal involvement o f the private sector and civil society resulted ina continuous degradation o f services. 145. The road maintenance fund (MF). In 1999, Madagascar created a RMF as a response to the poor quality o f maintenance o f the country's road network. This first generation RMF was managed by a secretariat and governedby a council (conseil d'orientation) composed of 4 members o f road user groups, 3 staff o f the ministries, and representatives o f local Governments. The sources o f funding o f the RMF are a tax on oil products (Tuxes sur les Produits Pbroliers, TPP) and fees on petroleumproduct through road users fees (Redevancesdes Usagers de la Route, RUR).The budget o f the RMF i s negotiated annually. Fundinghas been a key problem, in particular Government transfers to the RMF. In addition, organizational and technical problems (i.e. mismanagement, corruption) continue to hamper the efficiency o f the fund. -74- 146. Comprehensive reforms to adjust organizations and governance are under way. They consist of: Revised roles and responsibilitiesfor the center. To improve efficiency o f the public administration, the Government merged the aforementioned three ministries in 2002. The role o f the newly created Ministry o f Transport and Regional Planning i s to design transport sector policies, to plan and coordinate the activities o f the transport sector, to make budget allocation and decisions according to national development and poverty alleviation priorities. Autonomous agencies assume key functions. Under the revised administrative framework, regulatory functions will be managed by parastatal agencies, which are spin-offs from old ministerial departments or composed from existing agencies. Table IV.3 provides an overview about the new public agencies, including the various sources o f funding. -75- Table IV-3: Governance features of new public agencies Name (date of Governance Funding Hiring Descriptionlfeatures operation) ~ Civil Aviation o f Reportsto the Aeronauticalfees. Staffunder private Createdin2000, incharge of Madagascar (ACM) VPM. sector status and implementingGovemment'sair [createdin 1999; salaries. transport policy, overseeingthe Dperationalsince Boardconsistsof 7 sector, as well as drafting and 2000) members, of which enforcingtechnicaland 3 are fromprivate economic regulations. sector. Recently restructured. Resp. for managementof airports, regulationof air transport, andnavigation safety. Ongoing functional audit; annualreports for 2001 and 2002not approved RoadAuthority Autonomous To be determined Beginningyear Usedto manageall national (plannedcreationand agency; (studies in 2005 andprovincial roads, local operation 2005) deconcentrated progress). roadsat requestof offices communities;Functionsinclude planningofmaintenance, procurement,managementof constructionprojects. Surface Transport Board:transport To be determined Beginningyear Responsibilityfor vehicle Agency (plannedto operators and (studiesin 2005 licensingand control, public bepresentedto the Govemment progress). transport regulationsandroad Parliamentduring its safety, oversightof railway present session) concessions. Vehicle controlto be contracted out to private sector Port, Maritime and Reportsto the Fees on imported To be determined Responsibilityto manageports, River Agency VPM. goods, and (studies in navigation. (AFMP) (createdin Boardconsists of 8 concession fees, progress). 2000; not yet members,of which but not functional operationaldue to 4 are from the Yet fundingproblems) administration,and 4, from private sector RoadManagement User controlled. RoadUser Competitive Fund(RMF, Boardis private Charge; budget process;private 199912003) sector andusers' transfers sector salaries representatives Source :WorldBa, Devolution of responsibilitiesto local authorities. The Government plans to gradually devolve some activities inthe areas o f road maintenance, airport management, port operation and management to local authorities. 0 Disengagingfrom productive activities. The Government has already begunto withdraw from certain activities to increase the role o f the private sector: Air Madagascar has been put under private sector management on the basis o f a management contract; similar contracts are planned for Northern Railway, the main port (Tamatave) and the main airport (Ivato). P Revising theRMF. Experience inMadagascar has shown that the existence o f a RMF did not solve existing road maintenance problems as the fund i s dependent on the allocation o f budgetary resources by the Government. The Government intends to set up a second generation RMF. It is envisaged to convert the RMF into an autonomous agency which will be funded -76- mainly by an increase o f the road user fees, which are collected by oil companies and transferred directly to the RMF. Despite these intended changes, the sustainability of the RMF resources remains questionable in the long run: as the costs o f road maintenance are increasing, especially due to inflation, the level o f user fees will be insufficients2to cover the needs o f maintenance. The key challenge for the Government will be to find extra resources, especially by collecting heavy vehicle fees. To this end, lessons could be drawn from the experiences in other countries (see below box IV. ly3 Box IV-1 Criticalingredientsfor a road maintenancefund 1 Road users, stakeholders, and the general public must be well-informed about proposed reforms, and must understand the need for a RMF and dedicated road user charges. 1 All RMF revenue should be derived from user charges related to the benefits gained from roaduse and access. The revenue will likely come from levies on consumables, mainly fuel; annual vehicles license fees; supplementary heavy vehicle fees; fines for overloading vehicles and international transit fees (where appropriate). 1 Road users and stakeholders should have a strong hand in the RMF and the administration o f expenditure from road user charges. A clear separation should be maintained between purchaser and supplier functions, between finding and executing maintenance. The RMFprovides finds for road maintenance o n a sustained basis; it does not undertake implementation or procurement, although it does set rules for implementation and procurement to ensure that road user funds are efficiently and effectively spent by road agencies on implementation. 1 The RMF must be able to hold road agencies accountable for the funds provided to them. As a consequence, the RMF should have fill legal powers to set responsibilities, establish performance targets, and introduce sound business practices. 1 Full information on the planned and actual expenditure o f RMF revenues should be openly and promptly available to the public. Source: Asian Development Bank (2003) Road Funds and Road Maintenance - an Asian perspective 3) CRITICAL PUBLIC EXPENDITURE ISSUES 147. Higher allocations to the road sector are in line with PRSPpriorities. Table 4.4 below specifies expenditures for the period 1997-2004. It illustrates: (i) a drop in allocation to the transport sector after the 2002 crisis but an increase by 34 percent in 2004, (ii) the high share o f transport sector expenditures absorbed by the road sector in 2004; and (iii) the importance o f external assistance for the transport sector. 82 It i s estimated that user fees o f at leastlOUS cents per oil products liter are necessary, whereas the actual collected user fees are now 4 and 7 U S cents per liter, respectively for petrol and fuel. 83See for example: Gwilliam, Kennethand Ajay Kumar (2002) Road Funds RevisitedA Preliminary appraisal of the effectiveness of second generation road funds, World Bank; Heggie Isn G. (1999) Commerically managedroad funds: managing roads like a business, not like a bureaucracy, World Bank; De Richecour, Anne Balcerac and Ian G. Heggie (1995) African Road Funds. What Works and Why, World Bank. -77- Table IV-4: Madagascar's past allocationsto the transport sector (selected years between 1997-2004) 1997 1999 2001 2002 2003 2004 TOTAL ALLOCATIONS TO TRANSPORTATION Total VPM budget (MGFbill) 470.7 553.8 1124.0 1206.5 927.4 1126.1 o/w Transport budget allocation (MFGbill) 355.6 366.7 903.3 976.9 716.8 960.3 Total transport allocation (% of GDP) 2.0 1.6 3.0 3.3 2.1 2.4 Total transport allocation (% o f total budget) 10.9 7.5 13.0 13.2 10.8 12.6 Roads budget allocation (MFGbill) 288.5 294.3 809.4 852.5 627.1 855.5 Roads budget allocation (% o f total transport budget) 81.1 80.3 89.6 87.3 87.5 89.1 Real growth in total transport allocation Recurrent (all sub sectors) as % of transport budget 2.6 2.9 1.1 1.6 3.4 3.3 Salaries 1.2 1.4 0.7 0.8 1.5 1.8 Other 1.4 1.5 0.4 0.8 1.9 1.5 Capital (all sub sectors) as YO of transport budget 97.4 97.1 98.9 97.7 96.6 96.7 Local fin. 20.6 27.7 45.0 37.3 34.2 24.0 Donor 61.5 53.6 45.1 50.5 55.9 67.2 Other 0.0 0.6 0.0 2.5 0.0 0.0 - Memo Total Transport allocation.(% of total to VPM) 75.5 66.2 80.4 81.0 77.3 85.3 Source: Government of Madagascar and WorldBank 148. Low budget execution rates. Table IV.5 outlines the gap between budgetary allocations and actual expenditures for the transport sector during the 1997-2003 period. During this period, on average only 71 percent o fthe transport sector budgets were implemented. This is mainly due to low execution rates o f the investment budget that deteriorated from an execution rate o f 96 percent in 1997 to 89 percent in2003. Some o f the sub sectors (rail, maritime and ship) only achieved a 44 percent execution rate o f their capital budget over the 1997-2003 period. Execution performance o f the road sub sector, which was very low between 2000 and 2001 with an average execution rate o f 56 percent, increased significantly in the year 2003 with an execution rate o f 93 percent. Over the period 1997-2001, domestically financed programs performed better than donor financed projects (85 percent versus 74 percent execution rate). In 2003, however, this trend changed: foreign aid programs achieved a higher execution rate (104 percent) than government financed programs (98 percent). -78- Table IV-5: Madagascar - execution rate of the transport budget (selected years between1997-2003) Avg.' 1997 1999 2001 2002 2003 97-03 TOTAL TRANSPORTATIONBUDGET 95.3 91.9 50.8 21.0 89.2 71.4 Recurrent 86.5 102.6 94.3 100.0 94.0 97.8 Salaries 98.2 92.0 95.9 84.6 106.3 114.1 N o n Salary 76.5 112.4 91.2 62.0 84.5 84.1 Investment 95.6 91.6 50.3 19.8 89.0 72.9 Domestics 105.1 100.7 55.4 21.8 97.9 78.0 Donors 98.8 96.6 33.3 4.6 104.3 68.2 Road Sub Sector 111.8 97.9 54.5 22.0 92.5 76.1 Recurrent 97.2 89.8 97.3 73.0 108.3 105.0 Salaries 98.4 92.7 97.5 84.6 111.9 129.2 N o n Salary 95.4 84.0 96.3 21.9 98.4 83.1 Investment 112.1 98.1 54.2 21.7 92.4 75.9 Domestics 113.3 88.8 72.3 44.5 65.0 81.8 Donors 111.6 103.2 36.2 5.4 109.2 71.6 All other sub sectors(rail, air, ship) 24.3 67.2 19.3 13.7 65.8 48.2 Recurrent 77.5 111.5 92.0 112.9 91.9 94.5 Salaries 98.0 91.0 94.2 84.6 104.8 103.8 N o n Salary 66.6 122.5 89.1 64.7 83.6 82.7 Investment 20.0 63.O 14.7 4.6 57.7 43.8 Domestics 64.5 82.8 22.6 16.1 68.1 62.9 Donors 20.7 69.9 7.3 46.6 49.8 Memo Total exDenditures (% o f GDP) _. 1.9 1.4 1.5 0.7 1.9 1.5 Source:MEFB. WorldBank Average for the available year 149. Despite the improvement in 2003, the low execution rates inthe sector are concerning. They point to difficulties in implementing the transport sector programs. Main reasons include poor procurement and management capacity inthe Ministry o f Transport and Regional Planning as well as complex management and disbursement procedures under the different donor projects. Another problem in this context i s the limited absorptive capacity o f the sector - inparticular the private sector faces difficulties inimplementingthe ambitious Government program. 150. Transport budget information is inconsistent. Budget information derived fkom the official sources are inconsistent: transport expenditure data (allocations and execution, totals and subcomponents) in the official documents from MFEB and from the Auditor General (in particular "Loi de Reglement ") differ significantly from the information made available by the sector.84.Data reconciliation i s also problematic because the Ministry o f Transport and Regional 84An important reason for these discrepancies is that the data o f the Ministry o f Finance does not hlly capture all information about extemal financing In particular grants und other expenditures directly made available to the sector. -79- Planning in its present form was created only in 2002. Under the new ministry three formerly independent departments were consolidated (public works, regional planning, transport & meteorology); each with its own data and reporting system. Since the merger, the new ministry has struggled (with assistance from the World Bank) to integrate and consolidate available information. Another problem i s that reliable data (allocation and execution data, recurrent and capital, by financing source) for sub-sectors other than the road sub-sector are difficult to obtain. 151. Affordability and Realism. Table N . 6 below presents financing needs o f the transport sector as a whole. Madagascar expects to allocate on average 2.7 percent o f GDP in the period 2004-2006 to the transport sector, o f which more than 90 percent will be allocated to the road sector. Against the background o f the past performance in the sector these funding needs pose serious questions concerning affordability and implementation realism. For comparison, in the 1997-2003 period, the Government had budgeted transport expenditures o f about 2.3. percent o f GDP and executed about 1.5 percent o f GDP. The current PRSP program proposes to spend nearly two times more o f what was implementedinthe previous seven years. Table IV-6: PRSP Expenditure Forecasts 2004-2006 (as a % of GDP) Budget PRSP PRSP 2004 2005 2006 Recurrent 0.10 0.04 0.03 Salary 0.09 0.03 0.03 Non-Salary 0.01 0.01 0.01 Investment 2.21 2.92 2.55 Roads 2.11 2.82 2.43 Ports &Air 0.16 0.10 0.12 Rail 0.00 0.05 0.02 Regulation 0.02 0.16 0.17 TOTAL 2.4 3.0 2.6 Source:PRSP 152. Contingent liabilities and debt. Contingent liabilities (i.e. pension contributions to the CNAPS from public enterprises in financial troubles such as RNCFM, S M T M or CMN) pose significant risk for the budget (see table N.7). In addition, the public enterprises in the transport sector have accumulated substantial debts (to banks, other creditors, and the Government). By October 2004, the total outstanding debt o f the transport sector parastatals came to approximately MGF 384 billion or US$37 million equivalent intoday's exchange rates. Madagascar has taken several measures to address the issue o f rising public enterprise debt, i.e. increased control o f transfers o f public funds and spending, and financial audits. The Government has also taken over some public enterprise debts, such as for Air Madagascar and for RNCFM. It will, however, be important for the Government to clearly establish liabilities and debt with the objective to develop a strategy to deal with these risks (i.e. a debt management plan). -80- Table IV-7: PublicEnterprisesin the Transport Sector Date No.of Type Last Nameof agency Established Employees of PE Audit DebtOutstanding Date Air Madagascar 1961 1401 S.A. 2003 MGF 85.3 billion (US$8.3 million) Sept. 2004 Civil Aviation 1999 280 EPIC Feb. 2003 MGF 3.6 billion (US$347,037.5) Authority (ACM) +$79,750 Sept. 2004 Airport Management 1990 340 S.A. 2o04 MGF 62.2 million (US$6,017) (ADEMA) Oct. 2004 MADARAIL 2000 884 S.A. June20O4 WB, EBI: MGF 137.7 billion (US$ 13.4 million); Others: MGF 22.3 billion (US$2.2 million) Sept. 2004 Road Management 1998 30 EPIC MGF 30.5 billion (US$2.7 Fund(FER) million) June 2004 SEPT (ToamasinaPort) 1976 1,732 SA July2004 MGF 42.2 billion (US$4.1 million) Oct. 2004 Source: respective agencies 4) PRINCIPLE ISSUES OF SUB-SECTORS (a) Roads 153. Policy framework. The road sector i s central to Madagascar's PRSP as it i s an important element to increase rural incomes, by facilitating access to markets of rural households. As such, expanding the national highways and regional roads, and maintaining them, i s critical. Table IV.8 shows the road conditions of Madagascar in 2002. It indicates that 81 percent o f the roads are inpoor condition. Usually those roads are only seasonably practicable, and some not at all, causing very high costs to users, interrupting economic exchanges, and isolating towns and villages from the rest o f the country for up to eight months a year. -81- Table IV-8: Road Network Quality in 2002 and Objectives for 2008 Share o f road categ. TOTAL Compo Goomair Poor (Km) -sition (b) (Km) GoodFair Poor Primaryroads 2,560 8 1,616 944 63 37 Secondary roads 4,753 15 1,319 3,434 28 72 Temporary roads 4,549 14 1,000 3,549 22 78 Provincial roads 12,250 39 1,070 11,180 9 91 Communal roads 7,500 24 850 6,650 11 89 Total2002 31,612 100 5,855 25,757 19 81 Goal:: Total2008 31,612 100 19,062 12,550 60 40 Source: Government of Madagascar and World Bank 154. Road maintenance - the heavy burden of past failure. Poor road maintenance, as discussed before, has been responsible for the gradual degradation o f the road network throughout the country, and has left millions o f Malagasy stranded inrural areas, without access to schools, hospital or markets. A World Bank performance audit from the 1980s and project implementation experience from the 1990s suggest that Madagascar consistently under-funded the transport sector.85At the same time, recent performance has been impressive: from July 2003 to June 2004, the Government implemented routine maintenance o f 8,077km, rehabilitation and periodic maintenance o f 586km o f national roads, and rehabilitation o f 2,300 kmo f rural roads. 155. Reorganization of the transport sector is ongoing. The consolidation of the three transport ministries into one was primarily to the benefit o f the road sector, as it absorbs most o f the sector fundingand accounts for the largest volume o f transport. Under the new framework the central Government will maintain responsibility for policy malung while regulation and operations (maintenance and investment) will be devolved to autonomous agencies and to the private sector. The Government i s creating two new institutions: 0 Central Road Authority (AutoritC Routikre Centrale) i s expected to be operational in 2005. The Authority will be responsible for planning, construction and maintenance o f the national road network. Responsibilities, funding, Board composition and relationships with the other institutiondstakeholders remain to be developed and specified. Q Agency for Land Transport (Agence des Transports Terrestres) will also be operational in2005. The agency will be responsible to regulate the transport ofindividuals and merchandise, the inspection o f vehicles, the issuance o f licenses; it will assure road security and regulate concessions for the rail road. Details about roles, responsibilities, funding and governance structure remain to be specified. 85Sources are (i)PerformanceAudit -Madagascar-Ports Rehabilitation project (1752-MAG 1987), Third Railway project (1694-MAG 1986), SeventhHighway project (1905-MAG 1989) - May 27, 1998 Report No. 17915. -82- 156. To improve implementation of the road program the Government is reinforcing the RMF. A proposal has been submitted by the fund's Board of Directors (COHER) to the Government in 2004 to convert the RMF into an autonomous agency. The Board proposed to phase out Government funding o f the RMF and to replace it with increased road user fees. A Government decision i s outstanding; the budget for 2005 still maintains Government contributions to the fund. This presents a major concern as Government funding has been unreliable. 157. Risks are substantial. Two main risks could undermine the implementation o f the Government's ambitious program: First, the capacity of the private sector to implement the program i s significantly below expectations (domestic capacity o f the industry i s about half the PRSP target o f US$150 million). Second, program implementation would imply spending on roads o f about US$l50 million a year, or about 5 times the 1997-2001 average and double the 2003 expenditure levels in investment expenditures alone. In addition, maintenance costs on the existing and new road system are estimated to be around 0.6 percent o f GDP per year through 2008 (see table IV.9 below). The table confims that the largest amount o f fundingi s expected to come from the fuel levy. Table IV-9: RoadMaintenance: Total requirements and sources of funding (as percent of GDP) 2004 2005 2006 2007 2008 Fundingneeds 0.6 0.6 0.6 0.6 0.6 Sources of funds 0.6 0.6 0.6 0.6 0.6 Fuellevy 0.4 0.5 0.4 0.5 0.5 Heavyvehicle license fee 0.04 0.1 0.1 0.1 0.1 LocalGovernments 0.01 0.01 0.01 0.01 0.01 EU 0.02 0.0 0.0 0.0 0.0 HIPC 0.1 0.04 0.1 0.1 0.1 Source: Government of Madagascar and WorldBank @) Air Transport 158. Policy framework. Air transport in Madagascar has been dominated by the public sector throughout the 1980s and 1990s. In 1996, the Government initiated its policy to liberalize the air transport sector. Although progress was slow, due to this new policy the number o f tourist quadrupled between 1997 and 2000, to about 178,000. Concurrently, the Government began its disengagement from directly managing the airport system. The Civil Aviation Authority (Aviation Civile de Madagascar, ACM) which was created in 2000 i s in charge o f implementing Government's air transport policy, overseeing the sector, as well as drafting and enforcing technical and economic regulations. ADEMA (a company owned mainly and jointly by the Government o f Madagascar and Atroports de Paris, ADP) has a concession agreement with Government to manage the 12 principal airports (from 1991 to 2006). There are 43 other airports or landing strips in Madagascar (currently managed by ACM) which are particularly important for access to rural areas (21 serve exclusively for access to isolated villages). Air navigation systems are managed by ASECNA. Most airports require urgent rehabilitation to bring infrastructure to international standards, in particular with regard to safety and security -83- equipment, telecommunications, radio navigation, expansion o f terminals and runways, road access improvements. 159. Key public expenditure management issues. The Government's decision to disengage from the air transportation sector poses significant challenges. The country i s grappling with two issues: (i)the process and cost o f disengagement from Air Madagascar, and (ii) provision o f the air transport services to remote locations services. Both issues have significant budgetary implications: 160. As far as Air Madagascar is concerned, Government is committed to getting the airline profitable before it considers an outright privatization. To this end, a management consulting company was hired in April 2002. As a result, Air Madagascar returned to profitability in 2003. It achieved a surplus ofMGF 69,944 billioncompared to a deficit o fMGF28,161 billion in2002 and MGF 123,665 billion in2001. Government contributions were reportedly over and above the initial commitment to Air Madagascar in the amount o f US$7 million for an increase in capitalizationand US$lO million to meet creditor needs. 161. T o improve access to remote (and unprofitable) locations the Government i s paying Air Madagascar about US$900,000 for 2004 to serve these areas (about US$700,000/ MGF 5 billion were reportedly paid in2003). Discussions are currently under way for local Govemments to take over the management o f the 43 secondary airports. This strategy could, however, backfire and leave Government with contingent liabilities if the capacity at the local level to manage the airports remains inadequate. (c) Rail Transport 162. Policy framework. Inthe 1990s railway services were an important factor for the transport infrastructure with close to 2.4 million passengers and almost 600,000 tons o f merchandise (in 1990). Due to poor management, railway services nearly collapsed in 2001. The PRSP envisages a crucial role for the rail sector, not only for its economic but also for its social benefits because it can (potentially) grant access to remote locations (with n o road access) which are inhabited by 1.5 million people. In accordance with the policy o f the Govemment the development o fthe rail network will be driven by the private sector. 163. The rail network o f Madagascar consists o f the Northern and Southern Railways. The Northern Railway with a network o f 732 km was concessioned to a private operator (MADARAIL) in July 2003. The Northern Railway connects the country's main port in Toamasina with the capital Antananarivo and Antsirabe. The line has substantial commercial potential but i s in need o f major infrastructure rehabilitation and renewal o f rolling stock. The concessionaire plans a five-year US$35 million investment program to modernize the railway and provide a competitive transport alternative to roads. At the same time, the Govemment i s required to invest EUR 4 million to reinforce environmental protection along the tracks, and also to guarantee a European InvestmentBank loan o f EUR 11million. 164. The Southem Railway with a track length o f 163 km connects Fianarantsoa with the port o f Manakara; it i s planned to be concessionedjointly with the port o f Manakara to a private operator. Due to the limited commercial viability but important social aspects o f the Southern Railway substantial investment (US$ 17 million) i s planned on the basis o f Government and donor funding(including the World Bank) as soon as the concession agreement has been signed. -84- 165. Criticalpublic expenditure management issues. Potential contingent liabilities pose a severe risk for the implementation of the Government strategy. These liabilities needto be clearly established to assess the (fiscal) implications for the Government. In addition, funding arrangements for the concessioning o f the southern railway line are complex andnon-transparent; these arrangements should be laid open and Government contribution should be fully integrated into the annual budget. (d) MaritimeTransport andPortManagement 166. Policy framework. The main Government objectives for maritime transport and port management are to allow the private sector to run commercial ventures and to ensure that social needs are met. What i s most critical in Madagascar's maritime sector i s port management and maritime signaling. Only 4 o f Madagascar's 17 ports are deep enough for direct loading and unloading from the port, while another 6 ports allow for loading and unloading through smaller vessels. As part o f the sector restructuring, a new legal framework for the regulation o f the ports and maritime transport was adoptedby Parliament inJuly 2003. The Government created a Ports, Maritime and River Agency (Agence Portuaire, Maritime et Fluviale, APMF) to manage (not operate) the country's ports and navigation systems.86The law allows individual concessioning o f the 4 large ports (Ports a Gestion Autonome), and group concessions for the 6 medium-sized ports, while the 7 small commercially nonviable ports, as well as non-port related marine signaling, will be managed directly by APMF or fully transferred to local communities. 167. Criticalpublic expenditure management issues. Similar to the air and rail transport services it will be important for the Government to clearly establish existing contingent liabilities which could undermine the implementation o f its concessioning strategy. Adequate and reliable fundingof APMF will be also be a critical public expenditure management issue, which should be addressed through the annual budget. 5) RECOMMENDATIONS 168. Over the past two years, Madagascar has made significant progress in reforming the management and organization o f the transport sector, to reverse the twenty year policy o f state management and operation. The recommendations below come more inthe form o f identification o f risk areas to the success o f the overall reform, and include suggestions for risk mitigation. As such, risks are identified inthe following areas: Affordability and realism. Over the mediumterm, the transport sector i s poised to absorb more fundingrelative to GDP than inthe 1997-2001period and a larger share o f public resources than the 1997-2001 period. It will be important to assess whether Madagascar can afford the necessary investments, in particular in view o f foreseeable domestic revenues and donor contributions. In the short term, scarce public resources are likely to be claimed not only by the road sector, but by the contingent liabilities Government i s set to face in divesting the state from productive activities. 87.Additionally, administrative constraints will also needto betackled while 86 The private sector, through the National Employers'Association (GEM-Groupement d `Entreprises de Madagascar) and professionalport and maritime associations, has been closely involved inthe work leadingto the new institutional and legal framework. The principle o f representation at parity between the public and private sectors in the APMF board has been endorsed by both parties. This is consistent with the intention that most of the operating and development costs o f the sector will be financed through port and maritime user fees. The Govemment will retain its role in dischargingfunctionspertainingto planningandoversight, safety, environmentalprotection, and security. 87Source: MFEB -85- Government's success in attracting foreign companies will be crucial to improve the implementationperformance o f the sector. 0 The problem of autonomous public agencies. Government i s seeking to devolve the responsibility o f regulation o f the road, air transport, marine, and rail sectors to autonomous agencies. Experience in other countries, however, confirm that devolution to autonomous agencies i s not a silver bullet. The track record o f autonomous agencies i s mixed at best. Often, their introduction poses significant risks to further complicate matters and to replicate the non- functional bureaucracy o f the central Government at the level o f these agencies. International experience also indicates that these institutional changes will require significant implementation time and capacity. At the beginning o f this process, the devolution might even come with added upfront responsibility for the Government because - at the initiation phase- the Government i s responsible for assuring that these agencies operate under a viable and effective framework (Le. advisory boards and councils are not weighted by bureaucrats and have substantial private sector and civil society participation; management and staff o f the agencies are appropriately qualified and have respect o f the private sector; and most importantly, that funding for the agencies i s assured). Improvement of the Road Maintenance Fund. The effectiveness o f the road maintenance fund should be reviewed- also by taking into account international experiences with such an instrument and considering the proposal o f the COHER to transform the fund into a fully autonomous agency. For the road maintenance fund to be more effective it will also be important to update and rigidly enforce adequate accounting and procurement standards. The fund should (continue to) be subjected to regular (annual) value-for-money audits to ensure that the money i s well spent. 0 The issue of capacity of local authorities The intended devolvement o f some activities in the areas of road maintenance, airport management, port operation and management to local authorities should be complemented by substantial investments in training and capacity building to enable the local communities to adequately pursue their new responsibilities. Government should ensure upfront funding o f these capacity building activities through the annual budget as part o f the overall implementationo fthe transport sector strategy. -86- V. USEOF RESOURCESINTHE ENVIRONMENTSECTOR 169. This chapter will evaluate service delivery o fthe environment sector by taking a closer look at policy issues, government priorities, the legislative framework and the effectiveness o f public expenditures. 88Furthermore, the assessment will also focus on the institutional setting and the environmental impact assessment. 1) KEYPOLICYISSUES 170. Madagascar's economy is highly dependent on natural resources. Agriculture, forestry and fishing account for 32 percent o f GDP. Fishprovide 40 percent o f animal protein in the Malagasy diet. Around 85 percent o f domestic energy comes from fuel wood and other biomass. Inrural areas, 85 percent o f households grow paddy rice. Furthermore, shrimp harvest (US$l55 million) and tourism (US$82 million) are two o f the leading eamers o f foreign exchange. 171. Natural forests are a key asset. Natural forests provide key environmental services, regulating water flow and reducing siltation, to the benefit o f downstream paddy growers and towns with riverine fresh water supplies. Natural forests also harbor Madagascar's unique biological resources, an important global public good - Madagascar i s one o f the single most important biodiversity conservation priorities for the This biodiversity has been a crucial factor inthe growtho f ecotourism on the island. 172. The country's natural assets are at risk. Loss of natural forest, primarily due to tavy (slash and bum agriculture) and the use o f fuel wood and charcoal to meet domestic energy needs, i s the principal environmental problem facing the country. One third o f the natural forest area has been destroyed by slash and burn agriculture since the 1950's and the remaining forest area i s now highly fragmented." In addition to direct losses o f biodiversity (such as tourism revenues, timber, fuel wood, and fodder) forest destruction also leads to soil degradation and erosion, flooding, and siltation of downstream infrastructure - including dams, irrigation systems and domestic water supply facilities. Intackling these environmental problems policymakers are facing three major challenges: 0 Growing population and agricultural stagnation. The combination o f an expanding population and static productivity has generated pressure for agricultural expansion through conversion o f new land to agricultural uses, which typically consists o f the most accessible forests. Rice i s by far the most important crop in Madagascar. In 1960, before the Green Revolution, the average paddy yield in Madagascar was slightly higher than in Indonesia and significantly higher than inMali. However, average paddy yield per ha inMadagascar has been at the same level for the past fifty years, with yields o f around two tons per ha, while more than tripling in Indonesia and Mali. Over the same period the Malagasy population has tripled, with growth rates over the last 20 years o f 2.8 percent, slightly higher than the Sub-Saharan Africa average. Inthis chapter environment refers to the following subsectors: forestry, water andprotected areas. Work on commercial forestry, fisheries and minerals, while obviously germane, is left to future public expenditure reviews. 89 Mittermeir et al. 1998. 90 Facts and figures on deforestation and the physical geography o f Madagascar are given inAnnex 11.1. -87- 0 Poor incentivesstemmingfrom lack of property rights. Forested lands are the property o f the state inMadagascar, and local communities typically have no claim on these resources and no formal right to use them. Inpractice the state's property rightshave never been enforced. As a result, for local communities the surrounding forest i s in effect an open access resource. With no property rights to this resource, communities have no incentives to provide better management. Negative externalities associated with tavy in the uplands. Communities in the headlands o f water basins practice slash and bum agriculture without considering the negative effects on downstreamusers. 173. Government's policy actions. The primary response within the sector has been the creation o f protected areas, which both conserve biodiversity and protect watershed environmental services. The sustainability o f the country's protected area system i s linked to the incentives faced by surrounding communities to engage in tavy or more intensive agricultural technologies. Policies dealing with agricultural intensification, devolution o f property rights, and energy supply are important to altering the incentives faced by households to engage in deforestation. Similarly, mining regulation (excluding mineral activity in certain zone) and transport policy have profound influences on the pressure on forests." These policies are addressed in various sector strategies such as the Rural Development Action Plan (PADR)92o f the Ministry o f Agriculture and Government's PRSP. 2) SECTOR STRATEGIES The Government has a well-elaborated set o f priorities regarding the environment, as set out in the National Environmental Action Plan (NEAP) and the PRSP. 174. Madagascar's NEAP. The NEAPwas adopted by the Government in 1989, and given legal authority by the National Environment Charter and the National Environmental Policy in 1990.93Implementation started in 1991 with the support o f a broad coalition o f bilateral donors, international agencies and NGOs. Focusing on linking environmental protection and economic development, the plan was designed as a fifteen year investment program divided into three five- year phases (see also annex 11.2). The first five year phase (EP1 1991-1996) sought to create a policy, regulatory and institutional framework. The second phase (EP2 1997-2002) aimed at reversing degradation and integrating environment into other sectoral policies. The third phase (EP3 2004-2009), focuses on consolidating past efforts and establishing sustainable financing mechanisms for the e n ~ i r o n m e n t . ~ ~ 175. Positive impact of the actionplan ... Fifteen years since its publication, Madagascar's NEAP i s still relevant today. It institutionalized a clear set o f priorities - establishing protected areas, rural resource management, and environmental education and training - and was equally prescient in calling for a long term investment program. Major achievements o f the NEAP include: (i)the enactment o f enabling legislation for the protection o f the country's natural resources and the promotion o f environmental management; (ii)the establishment o f environmental institutions (such as the park service ANGAP) for the implementation o f environmental activities and programs; (iii) the development and implementation o f community- 91 Recent research suggests that each 12 km (one standard deviation) reduction in distance from forest to roads increases decadal deforestation rates by 2.6 percent. 92 The Action Plan for Rural Development (PADR) provides the strategic framework for the rural sector. 93 Law 90-033, December 21, 1990 94 The Madagascar ThirdEnvironmental Programwas approved inMay 2004. -88- based approaches for natural resources management; (iv) positive impacts on resource degradation as measured in the field; and (v) the establishment o f a platform for sustained donor support and coordination for the environment inMadagascar. 176. ...but challenges remain. However, there are several areas where the NEAP could improve its track record. The application o f policies and regulations remains a challenge because o f weak institutional capacity in some agencies and serious governance problems in the forestry sector. Resources under the NEAP have been disproportionately invested in capacities at the central level (around 70 percent), while too little has focused on strengthening institutional capacity on the ground. At the same time, the operational programs o f the NEAP at the central level have spread themselves thinly, with a proliferation o f executing agencies". Furthermore, lack o f rigorous priority setting has also led to a situation in which NEAP overlaps both conservation and rural development, resulting in potential competition with the PADR. Environmental management, for example, i s one potential area o f overlap, particularly in the periphery o f protected areas where communities have lost rights to use newly protected resources and alternative income-generating opportunities may be required to offset the loss. Finally, a particular challenge i s the limited knowledge and awareness o f the Malagasy population concerning environmental issues. This indicates that the NEAP'Sadvocacy and educational roles have not been as effective as necessary. 177. Political commitment in the PRSP...The PRSP effectively integrates the goals o f the NEAP into the program for reducing poverty (see table V-1). It is centered around three main environmental objectives: i)preserving environmental quality; ii)meeting needs for forest, land and water resources; and iii)integrating environment priorities into sectoral development policies and plans. A key statement in the PRSP i s the Government's recognition that managing natural resources and the environment will be key to sustaining growth inrural areas.96The Government proposes to achieve protection o f the environment partly by establishing financing mechanisms for the protected area system, but also by creating schemes to transfer property rights to local communities. 178. ...but success is determined by EP3 implementation. The successful implementation of the EP3 is an important contribution to the PRSPeg7Monitoring and evaluation o f NEAP through EP3 i s supposed to feed into the PRSP. However, the design and implementation progress o f the EP3 has not yet been incorporated into the PRSP. Furthermore a full costing o f the environmental strategy remains to be completed to address issues o f sustainable financing. Finally, in view of the PRSP implementation progress, the JSA recommended an acceleration o f ongoing institutionalreforms such as the establishment o f a new agency for forestry management (L'Agence National de la Gestion des Eaux et Forets - ANGEF) under EP3 and the need for measures that provide adequate incentives and improve institutional arrangements o f the sector.98 95 See table V-2 below. EP3 aims to rationalize the executing agencies to increase efficiency and effectiveness. 96 It argues that at least 50 percent o f income in the Malagasy economy is directly dependent on natural resources, making the linkages between environment and development particularly direct. 97 The policy matrix o f the PRSC in support o f PRSP implementation i s a sub-set of agreed actions under NEAP and implementedthrough EP3. 98 Such measures include : i)enhancing compliance o f public sector investments with existing MECIE legislation, ii)ensuring more sustainable financing o f Madagascar's protected areas system, iii)improving management o f coastal and marine resources and iv) levying a tax that reflect the environmental and health cost. World Bank, Poverty Reduction Strategy Paper - Annual Progress Report Joint Staff Assessment, 2004 -89- Table V-1: NEAP and PRSP objectives NEAP PRSP (i) Protecting and managing the national (i)Maintaining the quality of natural resources to allow ieritage o f biodiversity, with a special sustainable economic growth, :mphasis on parks, reserves and gazetted Developing the institutions and regulatory frameworks for iatural forests, in conjunction with the protection o f the environment and nature (applying the ustainable development o f their environmental code and legal instruments, implementation of urrounding areas; management systems, setting up governance and anti- corruption systems, developing environmental management ii)Improvingthelivingconditionsofthe indicators) )opulation through the protection and nanagement o f natural resources in rural Promoting the sustainable management of natural resources .reas with emphasis on watershed (simplified development and management plans, transferring rotection, reforestation and agro- property rights to local communities, forest zoning, and orestry; in urban areas this involves implementation o f an auction system for forest harvest mproving water supply and sanitation, permits) vaste management and pollution control; Ensuring the sustainable management of coastal and marine iii)Promotingenvironmentaleducation, resources (implementing integrated management regimes , raining and communication; transferring managementrights to local communities) iv) Developing mapping and remote Ensuring permanent financing of the protected areas system iensing tools to meet the demand for through a conservation trust fund and the development of iaturalresources management; `green' taxes v) Developing environmental research (ii)Meeting the population's needs for forest, land and water ,n terrestrial, coastal and marine resources, 0 Promoting a national program of reforestation and providing xosystems; and incentivesfor the development of tree crops. vi) Establishing mechanisms for 0 Combatingbushfires. nanaging and monitoring the mvironment. 0 Restoring the ecological and economicfunctions of catchmeni basins. 0 Supporting alternatives to deforestation (improved wood fuel management and promotion o f ecologically-friendly cultivation techniques) (iii)Integrating the environmental dimension into sectoral development policies and actions, and into spatial planning. 0 Promoting environmentlfriendly public and private investments, (applying the Environmental Impact Assessments: and ensuring the adherence of sectoral activities tc environmental goals. 0 Promoting environmental education. 0 Supporting the inclusion of environmental considerations intc regional, municipal and local developmentplans. 0 Innovating on hazard management ( testing and adopting i pollution management policy) Source: NEAP. PSRP -90- 3) POLICY AND INSTITUTIONALFRAMEWORK (a) The Legislative Framework 179. The legalframework remains deficient despite thefact that Madagascar i s one o f the countries in Afnca that has a fairly advanced legislative framework (for the protection o f the environment and natural resource management) and participated actively in several intemational environmental agreementsg9. In particular EP2 supported the revision o f various legal texts reflecting reformed management practices for natural resources and environment management. However, several problematic aspects o f the framework remain: (i) an uneven implementation o f the legislative framework - implementation decrees and administrative orders are laclung for years in some cases, or not kept up to date with legislation; (ii) a lack o f horizontal consolidation o f laws and regulations, for example, water i s subject to legislationregarding forestry, energy and the water code which leads to contradictions; and (iii) a variable range o f penalties for infractions o f environmentallaws and some current penalties that may be viewed as excessive. 180. Recent efforts have improved environmental management framework. Inthe past two years, the Government passed several laws and decrees related to environmental management. An important step forward has been the clarification o f roles between the Ministry o f Environment and Forests (MinEnvEF) and the National Office o f the Environment (ONE) following a modification o f the 1999 MECIE decree (Mise en Compatabiliti les Investissements a 1'Environnement) concerning the Environmental Impact Assessment (EM). As a result, ONE now manages a one-stop-shop for the application o f MECIE legislation which has reduced transaction costs for investors. Further legislative changes comprised o f the classification o f surface water, the regulation o f emissions o f liquidwaste and water management, the creation o f 22 administrative regions as well as the management and control o f industrial pollution and environmental protection inthe mining sector. @) The InstitutionalFramework 181. Investments made under the NEAP from 1991to date have led to the establishment o f a number o f environmental institutions - executing agencies - aimed at conservation o f the country's natural resources (see table V-2). loo Following closure o f EP2, some executing agencies have been put at arms length from the MinEnvEF and have started to offer their services to a wider range of clients."' The Ministry for Environment, responsible for the development o f national environmental policy, was merged with the Ministry o f Water and Forests in 2003. The principal challenge o f the merger i s to balance conservation and exploitation as the guiding principle for naturalresources management throughout the new Ministry. 99 These agreements are: the Framework Convention on Climate Change (1999), the Vienna Convention for the Protection of the Ozone Layer (1996), the Montreal Protocol on Substances that Deplete the Ozone Layer (1996), the Law o f the Sea (2001), the Convention on Biological Diversity (1996) and the Kyoto Protocol (accession in2003). loo The Charter for the Environment (law 90-033) provided for the creation o f a national structure responsible for development o f national environmental policy. lo' TheNational Association for Environmental Actions (ANAE),for example, specialized on specific organic and low extemal input technology which allows it to raise independently operating funds through the provision o f its services. -91- Table V-2: Executing agencies for the NEAP Fonciers (DGDSF) system management Centre de Formation aux Sciences de 1'Information - chargedwith environmental training Gtographique et de 1'Environnement (CFSIGE) 182. Institutional reforms under way under EP3. The institutional structure i s complex and reflects in many ways the needs o f the funding agencies for counterparts during the implementation o f EP1and EP2. Under EP3, a project implementation unit (CELCO) was created to coordinate all program activities and to monitor and evaluate implementation progress. A key priority for EP3 i s to rationalize the institutional structure for environment and natural resource management (see annex 11.3). There are three guiding principles for the proposed institutional structure: (i)MinEnvEFshould play a policy-making and coordinating role, (ii) functionONE'S as an executing agency has already changed to an agency with policy and oversight responsibility, (iii)operational work should be carried out by specialized executing agencies (ANGAP and ANGEF), and (iv) more responsibility for environment and natural resource management should be devolved to local communities. A core ministry dealing with policy- making and coordination, combined with executing agencies which generate revenues and deliver services to the ministry in return for financial transfers, has the potential to be an effective and capable institutional model. However, a comparison with other countries shows that such an institutional setting can be inefficient if incentive and local capacity problems emerge. Devolution o f management responsibility inthe environment sector needs to be accompanied by capacity buildingand adequate funding, either through central government transfers or by raising revenues locally. 4) CRITICALPUBLICEXPENDITURE ISSUES Sectoral and Economic Allocation 183, Mixed implementation record. Table V-3 shows budget allocation and executiono f the environment sector over the period 1997-2004. Several observations can be made: 0 While environmental expenditures increased between 1997 and 2001 with an annual average growth rate o f 20 percent, public spending in the environment sector as a share o f the -92- national budget dropped significantly since 2001 from 4.6 percent to 2 percent o f the total budget. These cuts are mainly due to the phasing out o f EP2 (2002/03). To compensate, Government increased slightly its domestic resources from 1.7 percent in 2001 to 2.3 percent o f the total budget in2003. 0 There i s a significant variance between budget allocation and actual expenditures reflected in annual execution rates, which are far above 100percent between 1997 and 2001. This can be partially explained by the availability o f grants that are executed independently from counterpart funds and not budgeted.The weak implementation performance in 2003 i s related to the closing o f EP2 inthe middle o f 2003. o Donor funding i s an important part o f the environment budget which i s indicated by the EP2 share of total grants, reaching a higho f 22 percent in2001. EP2 grants accounted on average for 55 percent of environment expenditures over the period 1997-2001. The international community continues to provide support under EP3 (around $132 million) over the period (2004- 2009) comparedto $124 million under EP2. -93- Table V-3: Madagascar Environment Sector: Central Government allocation and - expenditure (selected years between 1997and 2004) 1997 1999 2001 2002 2003 2004 EP2 implementation EP3 BudgetAllocation Total budget allocation (MGF bill) 99.53 99.63 173.84 242.24 161.71 172.65 Total budget allocation (as a % of total budget) 3.05 2.03 2.50 3.28 2.44 2.26 Real allocation (1997prices) 99.53 87.54 131.92 161.82 109.74 108.16 Actual Expenditures Total expenditures (MGF) 123.44 204.54 250.30 101.12 137.76 Total spending (as a % of total budget execution) 4.50 5.23 4.81 3.16 2.27 Total spending (as a % o f GDP) 0.68 0.87 0.84 0.34 0.41 Total real spending (1997 prices) 123.44 179.72 189.95 67.55 93.49 Real growth 16.60 12.86 (64.44) 38.40 Budget execution rates ' Budget execution 124.0 205.3 144.0 41.7 85.2 EP2 expenditure 136 305 114 28.2 0 Administration and non EP expenditures 93 108 216 69.5 84.1 Grants Actual EP2 grants (MGF bill) 63 113 170 5 3 EP2 grants (as a % of environment expenditure) 51.04 55.25 67.92 4.94 2.18 EP2 (as a % of total grants') 8.57 14.68 21.68 4.59 0.33 Memo Total grants2(MGF bill) 735 770 784 109 920 Source: IMF,MEFB ' Actualexp./ allocated budget 'Total national grant budget 184. Limited utilization of HIPCfunds. Less than one percent (MGF 8 billion) o f the HPC funds have been used for the environment sector over the period 2001-2004, compared with an initial commitment at the decision point o f usingup to 10 percent o f the total debt relief for this purpose. As the country reached the completion point under the enhanced HPC initiative, the debt relief will amount to US$ 61.5 million annually for the period 2004-2013. This provides an opportunity to allocate a larger share of budgetary resources for the environment and biodiversity conservationbased on an updated and fully costed environmental strategy for the 2006 revision of the PRSP. 185. Weaknesses in budget execution. The high budget execution rates (see table V-3) mask deficiencies in budget implementation as a significant share o f the resources are not budgeted and not reflected in the data. Major problems include: i)non-standardized procurement processes among the executing agencies and low ceilings for the special accounts o f these -94- agencies; ii)a lack of harmonizationbetween accounting systems - chart o f accounts, accounting software, and method of recording - used by the project coordinationunit (ONE) and the various executing agencies; iii)an absence o f an effective financial management system to track variances between actual and planned expenditures; iv) delays in the payment of contractors and suppliers because o f lack o f cash forecasting and irregular submission o f withdrawal applications to donors; v) weak financial management capacity o f some executing agencies; and vi) a lack o f authority as well as a clearly defined responsibility o f the unit in charge o f project management and control. 186. International comparison with OECD countries suggests no overspending of the sector. There are no low income country benchmarks to compare environment budget expenditures. However, OECD government expenditures on environment (excluding water and sanitation) average 2-3 percent o f GDP. In Madagascar, total allocations to the environment sector (excluding grants) averages less than 0.4 percent o f GDP. 187. An allocation largely infavor of capital expenditures. The expenditure breakdown in table V-4 reveals several interesting findings: First, expenditures have been allocated in favor o f capital spending, on average 91 percent o f total expenditures over the period 1997-2003. Only in 1999 and 2003 expenditures amounted to 21 and 15 percent o f total environment budget due to a frontloading o f capital expenditures between 1997-1998 and the reduction o f the overall budget envelope for the environment following the closure o f EP2 in2003, Second, more than 50 percent o f capital expenditures are financed by grants, reaching 69 percent in 2001 and 64 percent in 2003. The highreliance on foreign assistance raises the issue o f the long-term sustainability o f the environmental program. Third, operating costs, classified under goods and services, account to 2.4 percent o f the total environment budget in 2003. This represents a significant improvement compared to previous allocations (on average 1.1percent o f the environment budget over 1997- 2001). Environmental agencies, however, continue to underscore the insufficiency of operating costs, the need for skills upgrading and for the availability o f qualified staff in particular as donors are reluctant to finance recurrent costs. This i s a major problem since the conservation o f natural resources to a large extent depends on the availability o f adequate operating costs for surveillance, control etc. -95- Table V-4: Madagascar's recurrent and investment spendingpatternsin the environment sector over the years 1997-2003 (actual expenditures aspercent of total environment budget) Annual avg 1997 1999 2001 2002 2003 1997-2003 TOTAL 100 100 100 100 100 100 Current Expenditure 6.8 21.1 6.2 12.0 14.6 10.5 Salary 4.2 4.2 2.5 7.5 7.9 4.8 Goods & Services 1.5 0.8 1.0 2.1 2.4 1.4 Other' 1.1 16.2 2.7 2.4 4.3 4.3 CapitalExpenditure 93.2 78.9 93.8 88.0 85.4 89.5 Revenues 10.6 5.4 10.0 23.9 20.6 13.6 Borrowing 33.1 19.4 14.9 29.3 0.9 22.5 Grants 49.5 54.0 68.9 34.8 63.9 53.4 Memo Current expenditure (inMGF bill) 8.4 31.3 10.6 10.8 16.1 Capitalexpenditure (inMGF bill) 116.6 173.4 241.0 88.7 116.5 `Residualof current expenditures, salary and goods & services Source: MEFB 188. Distribution of expenditures across programs is not available. The identification o f the thematic content o f environmentalexpenditures i s hampered by data limitations. Environment expenditures (varying between 60 to 90 percent) are classified thematically as "common environmental services". A small sample o f the expenditures o f the four principal executing agencies shows that the spending patterns are broadly in line with the objectives o f the NEAP (see chart V-1 below). However, general conclusions cannot be drawn as these constituted only 32 percent o f total environment expenditures in2001, with the remaining spending being carried out by the Ministryand smaller executing agencies. -96- Chart V-1: Distribution of expendituresby environmentaltheme (2001) Soil &water management aaf- Water &forest re! 14% Source: MEFB 189. Coverage and transparency need improvement. Whereas MinEnvEF i s fully covered in the budget, budget information on other executing agencies are limited. Consequently, the Ministryis not able to adequately monitor the implementationof its sector strategy. Furthermore, donor financed projects fall under capital expenditures (though they - inreality - finance partially current expenditures), which leads to biases in assessing the composition o f environmental expenditures. To ensure better transparency, it will be important to improve the presentation o f the environment expenditures o f the budget. The recently introduced program budget could be an opportunity to revise the classification o f budget allocations among current and capital expenditures, and to bringthem inline with reality. Sustainablefundingoptions of protectedareas duringand after EP31°2 190. The needfor sustainablefunding of Madagascar's protected area. As Madagascar entered the third and final phase o f the implementation o f the NEAP in 2004, the question o f the sustainability o f the protected area system emerges as a key issue. Under EP3 ANGAP i s in charge o f managing this system.'03 Although the donor community continues to provide significant financing during EP3 implementation, a similar level o f foreign assistance by the donor community i s not guaranteed after the closure o f EP3 and more than 15 years o f supporting NEAPimplementation. lo2 Madagascar's protected area system is a global public good, protecting unique biological resources, a local public good at the watershed level, providing reduced erosion, stream flow regulation, and water purification, and an important source of tourism income, both in terms o f foreign exchange and fiscal revenues. I O 3Inthls respect, the investment cost ofANGAP includes i)enhancing the coverage ofprotected areas (10 percent), ii)managing biodiversity (55 percent) , iii)developing eco-tourism infrastructure (3 1 percent), and iv) increasing public awareness (4 percent). -97- 191. Shortfall of EP3funding. The long term sustainable funding o f the protected areas i s at risk as ANGAP could already experience financing shortfalls (based on budget projects from 2003) during EP3 implementation. The estimated capital and operating costs o f ANGAP (around MGF 53.2 billion annually) reveal a significant shortfall of funding, averaging MGF 11billion a year over the next five years (see annex 12.1). Without additional resources (either from donors or fiscal revenues) this shortfall will almost certainly entail reductions ininvestments, which may inturnhurtthe growth potentialfor ecotourism. 192. Establishing a Madagascar protected area trustfund. A key proposal in EP3 i s the creation o f a protected areas trust fund. This fund i s expected to total US$50 million by the end of EP3. Assuming a 6 percent interest rate and 10 percent operating costs per year, this endowment fundwould ensure sustainable financing ofupto MGF 15.5 billion a year (US$ 2.5 million) after EP3 ~ompletion."~ 193. Potential mobilization of revenues. Ecotourism provides an important potential in generating revenues. With an average annual 10 percent growth rate since the early 1990s, tourism i s an increasingly important economic sector that has become the third foreign exchange earner with US$ 82 million in 2001, after fisheries and vanilla. In 2001, 170,000 tourists visited Madagascar and more than half o f them (55 percent) considered themselves ecotourists. Madagascar's protected areas have established themselves among the main tourist attractions o f the island. Ten protectedareas currently contribute to the development o f tourism inMadagascar, attracting a growing number o f tourists. Five other protected areas have untapped ecotourism potential. 194. Studies suggest that foreign visitors are willing to pay (WTP) higher amounts than the fees currently charged for visits o f protected areas.lo5Conservative estimates o f potential tax and fee revenues from ecotourism assume a doubling o f park entrance fees and an increase o f hotel taxes amount annually to MGF 61.8 billion (see annex 12.2).106, 107These revenues would exceed the annual capital and operating costs o f the protected area system by MGF 8.6 billion.'0s However, Government's commitment to ensure the development o f the protected area system i s imperative to the success o f generating higher tourism rents. These rents are closely linked to the quality o f the resources that the tourists come to enjoy, as well as the facilities and services that support their tourist experience. In annex 12.1 this is assumed to be half this amount, reflecting the time required to buildup the full endowment for the trust. lo'In2001 a contingent valuation study was carried out to evaluate the WTP o f foreign tourists to visit the National Park o f Andasibe, the most visited park o f Madagascar. Results showed a WTP o f MGF 80,000 per visit without any service upgrades and MGF 100,000 with upgrades o f guide services. lo6The revenues estimates assume that hotel taxes could be raised from MGF 3000 (roughly $0.50) to MGF 30,000 (around $5) per night, still extremely modest by international standards. The hotel tax i s generally the preferred way to capture tourism rents because it is proportional to the tourist's use o f the natural asset. International experience has shown that demand for hotel rooms in quality tourist destinations is quite price-inelastic. lo'The estimates are based on a yearly constant 170,000 tourists inflow (instead o f the projected 6-8 percent annual growth by the World Tourism Organization). They also include the sharing o f 50 percent o f entrance fees with local communities. lo*This also ignores the VAT and other fiscal revenues generated by tourists as they travel around the island and purchase goods and services - a rough estimate, based o n the 20 percent VAT rate and US$ 82 million in current annual tourist expenditures, suggests VAT revenues from eco-tourists (55 percent o f the total) o f MGF 56.9 billion a year. -98- 195. External benefltsfully offset local opportunity Beyond tax and fee revenues, the protected areas system i s currently generating MGF 40.6 billion annually in external benefits to the rest o f the economy. These benefits include payments to local communities (MGF 1.4 billion), reduced siltation and flooding o f downstream rice terraces (MGF 17.6 billion), and cleaner and more regular water flows for potable water supplies to downstream towns and cities (MGF 21.6 billion).''03 "' This must be viewed against the lost opportunity cost o f local communities due to the establishment o f protected areas, amounting to MGF 37.2 billion annually. This includes revenue losses associated with natural forest conversion through tavy to produce rice (MGF 20.0 billion), non-sustainable fuelwood harvesting (MGF 7.1 billion), and non-timber forest product collection (MGF 10.1 billion)."' Overall, the external costs and benefits o f the protected areas system are roughly inbalance. However, gainers and losers are not the same households; there remains the problems o f how can beneficiaries compensate losers. A system o f payments for environmental services in these watersheds i s worth exploring to address this issue. These systems have been successfully implementedinother countries (e.g.Costa Rica) butthese countries are typically muchricher than Madagascar. 5) EFFECTIVENESSENVIRONMENTAL OF INSTITUTIONS The institutional setting for environmental management inMadagascar i s currently in flux, as the govemment began implementation o f EP3 in 2004. As noted before, a principal aim o f EP3 i s to rationalize the institutional structure for environmental management. The reorganization o f institutions will have (hopefully only) a short-run impact on their effectiveness. 196. Mixed effectiveness of the main environmental institution^."^ In 2004, the World Bank conducted an assessment o f the institutional effectiveness of the main environmental institutions (DGEF, DGE, ANGAP and ONE), drawing in part on the planning documents for EP3 (BIODEV 2003)'14. This evaluation i s subjective since reliable benchmarks against which to evaluate organizational effectiveness are missing. As a consequence, the findings are limited to two general observations: first, capacity and effectiveness of the two agencies (ANGAP and ONE) are considered to be stronger than that o f the two general directorates within MinEnvEF. Second, strategic planning i s a problem across the environmentaldomain. 109 Carret J.C., Loyer D. "Comment fiancer durablement les aires protCgBes de Madagascar D, Notes et Etudes N o 3, Agence Franqaise de DCveloppement, Novembre 2003. 'lo Analysis o f each watershed having a protected area in its headwaters shows obvious hydrological linkages between 20 PASand at least 430,000 hectares o f irrigated paddy rice fields and 17 towns with an annual 8.4 millionm3potable water consumption situated downstream. Employment opportunities in PAS and income from handicraft sales are additional benefits not accountedhere. '12Note that this MGF 37.2 billion loss inincome over 5 years i s a worst-case scenario, because it assumes that slash and burn activities in forested areas are entirely additional to existing economic activities inlocal communities. 'I3The assessment i s based on interviews with senior officials and the lessons learned inthe design o f EP3, particularly with regard to institutional design (Biodev 2003). The evaluation does not necessarily reflect the view o f all relevant stake holders, the MinEnvEFi s questioning its findings. This assessment (see annex 12.3 and 12.4 for details) evaluatedinstitutionaleffectiveness onthe basis o f selected criteria (alignment with government priorities, whether objectives are reflected in organizational structures, and whether the building blocks for effective implementation are in place). Furthermore, the assessments took a closer look at the main organizational functions o f each institution, including strategic planning, budget management, operational management, monitoring and evaluation. -99- 197. Protected area system effective. ANGAP manages 46 protected areas, covering 3 percent o f land area and 15 percent o f forested land. Protected areas include 87 percent o f the major ecosystems o f Madagascar, which compares favorably with the intemational good practice benchmark o f 75 percent. The key effectiveness indicator i s the deforestation rate within protected areas - this has fallen essentially to 0, while remaining at 1percent per year outside o f protected areas. Recent statistics indicate that forestry fire and the practice o f tavy have been reduced over the past years'15. 198. While Government has dealt with some problems, several issues o f institutional effectiveness remain. Some o f the key issues regarding the institutional effectiveness o f the environmental institutions are listedbelow intable V-5: Table V-5: Summary of some institutionalissues andreformstatus ~~~~ ~ Issues Reformstatus DGEFI 4 particular challenge is the combination o f the DGEF has not, to date, taken concrete steps DGE xganizational culture o f a production sector for the implementationo f EP3 :forestry) with a mandate for a more integrated :nvironmental management (e.g. its Petty corruption and offenses against the :ffectiveness as a manager o f conservation forest forest code are still features o f the sector. remains to be confirmed ) Legislation and regulation in the sector i s overly complex and applied only unevenly and arbitrarily. However, fee collection from logging permits has increased recently. Law 96-025 on Secured Local Management The government approved the Forest (GELOSE) aims to reduce deforestation by Management Contracts (GCF) decree, which providing for the transfer o f management transfers the management o f forests to local responsibilities for renewable natural resources communities in 2000. The GCF falls under to local communities. The operation o f GELOSE the GELOSE law but places specific emphasis requires the intervention o f an official mediator on community-based forest management. to supervise the transfer and resolve conflicts There is, however, an overlap and lack of that may arise. The complexity o f the GELOSE clarity between the GCF and GELOSE process i s responsible for the low number of programs, hectares transferred to community management ONE Average time to process environmental impact The decree 2004-167 on MECIE has assessments has been extremely long due to: eliminated the institutional overlap and made 0 a significant learning process required ONE the `one-stop shop' for EM. It also between 1997-1999 (e.g. each new sector to mandates a processing time o f 60 days from be assessedhad its particular features) and the date o f eligibility assessment, or 120 days 0 a significant overlap in the responsibilities in the case o f EIAs requiring public o f the MINENVEF and ONE with respect to consultations (See also annex 12, table 12.5) EL4since 2000-2003. ANGAP Effectiveness o f transferring entrance fees to Reforms aimed at increasing community communities (this is an issue largely because empowerment are underway - see section on ANGAP's core mandate is conservation, no1 revenue sharing with local communities. development). Source: h EnvEF, World Bank 1*5Substantial reduction in the deforestation owing to tavy i s evidenced by the most recent figures o f Conservation International based o n an anlysis o f N A S A satellite imagery. -100- 6) ANUNFUNDED MANDATE: ENVIRONMENTAL IMPACTOFPUBLIC ASSESSMENTS INVESTMENTS 199. Double standardfor public investments?Environmental impact assessments (EIAs) in Madagascar are governed by the MECIE decree. While all large-scale investments, both public and private, are subject to the decree (at least to the point o f determining whether an EIA i s required), to date there has been little application o f EIA for public sector investments. Untilmid- 2003, 72 EIAs were carried out under the decree; only two o f them for public sector investments due to a lack o f a uniform enforcement o f environmental laws. This situation presents a severe environmental risk, since large-scale public investment i s expected to average more than three times the value o f similar private investment until 2006. This also creates a negative image for the government because o f the double standard applied to private and public investments. 200. Public sector does not systematically pay the permit fee. Environmental Impact Assessments are administered by ONE, which i s financed by two types o f revenues: (i) general fees for EIAs from potential investment projects o f the public and private sectors, and (ii) permit fees (assessment and monitoring) for large investment projects. While the general fees for EIAs are used to cover the cost o f conductingthe impact assessments, the permit fees, which flow into an off-budget account o f ONE, are used to fund the operating and management costs o f ONE. Current practice suggests that all line ministries budget adequately for EIAs intheir project costs. Nevertheless, ONE faces financing problems for two reasons related to the permit fees: first, not all line ministries budget for permit fees, and second, the permit fees that are budgeted are at times not adequate. The Government i s in the process o f exploring sustainable funding options for ONE. Prior analysis, under the EP3 project, had anticipated that (i) the general fees be set at cost recovery level for the EIAs, and (ii) that total potential revenues from permit fees broadly cover the costs o f EIA management by ONE - estimated at around MGF 4billion inthe planning documents o f EP3 (see annex 12.6)'16. Under these circumstances, EIA administration in Madagascar could be placed on sustainable footing ifpayments for permits o f public investments are budgeted and paid for by all corresponding line ministries as well as decentralized layers o f Government. Altematively, ONE had suggested to create a budget line item at the MinEnvEFthat would ensure funding for management o f ONE; a Government decision on this proposal i s still outstanding. 7) REVENUE SHARING WITH LOCAL COMMUNITIES 201. Inadequate transfer of funds to communities. Since the beginning o f NEAP implementation in 1991, park entrance fees have been shared on a 50/50 basis between 27 protected areas managed by ANGAP and the neighboring communities. From 1992 to 2002, park fees generated MGF 6.3 billion o f PRDEAP'" (local communities share o f revenues from park entrance fees), 90 percent o f that amount coming from 5 protected areas: Isalo, Andasibe, Ranomafana, Amber Mountain and Ankarana. To date, only MGF 2,3 billion or 37 percent o f PRDEAP have been transferred to neighboring communities to finance 520 micro-projects: 175 social assets projects (rehabilitation o f schools, bridges, markets, for example) duringNEAP first stage (1991-1996) and 345 diversification o f livelihood projects (income generation) during NEAP second stage (1997-2002). `16Based on ONE experience, it is estimated that the total cost o f EIAs amounts to 0.75 percent of a project's investment costs. Permit fees are projected at 0.35 percent of total investment costs. Potential permit fees could rise to MGD 3.6 billion per year. Moreover, ONE estimates that additional MGF 1.2 billion will be raised annually from private sector investments. (Biodev 2003) 11'Part des Recettes pour les Droits d'EntrCe dans Aires ProtCgkes (PRDEAP). -101- 202. Complex institutional arrangements. During the second stage o f NEAP implementation, PRDEAP funds were remitted to the ANGAP head office in Antananarivo. Subsequently, representatives o f members o f grassroots communities and village-based associations (COGES118) proposed micro-projects and submitted them to the ANGAP regional office representative. Once approved by the ANGAP regional office representative, micro- projects were presented to ANGAP head office for clearance. Once approved, mini-projects were co-managed by COGES and the site representative. The complexity o f this institutional arrangement partly explains the low transfer rate (37 percent) o f PRDEAP funds that have actually been transferred to neighboring communities during the first two stages of NEAP implementation. 203, Reforms of communityfunding mechanism under EP3 Duringthe preparation o f the third stage o f NEAP implementation (2004-2009), this institutional arrangement will be reformed. PDREAP funds will be placedin a special social fund. Miniprojects will be elaborated by local communities with the technical assistance of the Regional Steering Committee (CRO) that will involve selection o f representatives from all stakeholders present in the area, including representatives of members o f grassroots communities, village-based associations, regional development committees, civil society/development operators, and private and public institutions. In addition, PDREAP funds will be managed under a contractual basis by beneficiary communities, CRO, communes and site representative. These decentralized management principles should bring greater benefits to communities neighboring protected areas. Protected areas do offer benefits to local communities, particularly the creation o f employment. PDREAP funds are augmenting these benefits and helping to offset the opportunity costs borne by local communities. Until today, however, the overall implementation o f this community funding mechanism appears to be slow. 8) RECOMMENDATIONS 204. After ten years o f investment by the Government, Madagascar has established the basis o f a modem environmental management system. A legislative framework is in place. A functioning EL4 process has been developed. An effective system o f protected areas, covering most biodiversity `hotspots,' has been built. The analysis, however, also draws attention to several flaws in the area o f public finance, institutional effectiveness and implementation o f government sector priorities. The present resource envelope does not seem adequate for the investment and operating needs o f the protected area system over the next five years, which will necessitate a focus on revenue generation, the functioning o f environmental institutions and the implementation o f government strategies to ensure an efficient use o f public resources. Thus, Government's action plan should focus upon: 0 Increase tourism revenues. Current tourism rent capture i s extremely l o w in Madagascar. The government should (i) entrance fees for foreign tourists inprotected areas raise (this can potentially double total revenues), and (ii) hotel accommodationtaxes inline with raise other tourist destinations. o Ensure sustainablefunding of theprotected area system. Government should close the projected funding gap for ANGAP over the next five years and secure sustainable funding for the protected area system by tapping into tourism revenues and HIPC resources, and by seeking additional external resources for the protected areas trust fund. ~ Comite de Gestion(COGES) -102- 0 Improve financial management. Government should address existing problems in budget planning, coverage and execution. This will require training and capacity building in the area o f public financial management, as well as the development of a consistent financial management framework for the sector. The existing financial management fkamework should be revised to capture the distribution o f expenditures across different elements o f environmental management. o Ensure adequatefunding of Environmental Impact Assessments. Payment of permit fees for public investment projects should be enforced or an alternative funding mechanism for the management o f ONE shouldbe developed. Increase compensation of communities near protected areas. A social fund using PRDEAP park revenues should be created to improve the compensation o f communities near protected areas. Participation o f these communities in the management o f protected areas should be enhanced by systematically implicating local committees and NGOs. o Streamline the devolution of local resource management rights to communities. As the devolution o f resource management can potentially reduce deforestation, the existing processes for community-based forest management (GELOSE and GCF) should be harmonized and streamlined. -103- Asian Development Bank: "Road Fundsand RoadMaintenance - an Asian perspective", 2003 Bloch-Laid, B de la Biche JP Landau : "Audit du Ministkre des Finances et du Ministhe du Budget ,de l a Decentralisation et duDeveloppement des Provinces Autonomes", Octobre 1999 Brinkerhoff, D. and S. Keener, "District level service delivery in rural Madagascar: Accountability inHealth and Education", World Bank Washington 2003 Cabinet R conseil : "Etude sur 1'Etat des Lieux du RCgime Actuel de RCmunCration du Personnel de l'Etab), V o l I& 11,Antananarivo 1998 Case, A. and A. Deaton, " School inputs and educational outcomes in South Africa." The Quarterly Journal o f Economics, August 1999 Carret J.C., Loyer D. "Comment financer durablement les aires protCgCes de Madagascar H, Notes et Etudes No 3, Agence Franvaise de DCveloppement, Novembre 2003. D e Richecour, Anne Balcerac and Ian G. Heggie: "African Road Funds, What Works and Why" World Bank Washington 1995 Dessureraux, Charles : Final Mission report on the Management Information System for MINESEB, Antananarivo 2004 Francken, Nathalie: Service delivery in Public Primary Schools in Madagascar: Results o f a Budget Tracking Survey, Draft report presented at the World Bank Country Office in Antananarivo, Sept 2,2003 Group 2AC : "Audit des Fonds de Contrepartie de 1'Appui ri 1'Ajustement Structurel dans la Republique de Madagascar", Antananarivo Mai-Juin 2000 Gwilliam, Kenneth and Ajay Kumar: "Road Funds Revisited - A Preliminary appraisal o f the effectiveness o f second generation road funds", World Bank Washington 2002 Heggie Isn G.: "Commerically managed road funds: managing roads like a business, not like a bureaucracy", World Bank Washington 1999 National Statistics Institute (INSTAT) / Madagascar Dial Instat-Orstom (MADIO) : "Rapport des RCsultats: Les Indicateurs de RCalisation du Programme d'Ajustement Structurel," Antananarivo 2003 International Monetary Fund Madagascar Heavily Indebted Poor Countries Assessment and " Action Plan", Washington 2004 International Monetary Fund: "Madagascar: Enhanced structural adjustment facility policy framework paper (1999-2001)," Washington 2003 -104- International Monetary Fund: "Madagascar: Selected issues and statistical appendix," Country report no. 0317, Washington 2003 King, E., L. Rawlings, and B. Ozler: "Nicaragua's School Autonomy Reform: A First Look". World Bank Working Paper Series on Impact Evaluation o f Education Reforms. The World Bank Washington 1996 L e Baron, Alain, Raude, Jean-Yves : ((Audit Institutionnel du TrCsorD,Madagascar, rapport provisoire, January 2004 L e Houerou, PhilippelTaliercio, Robert: " Medium Term Expenditure Frameworks: From Concept to Practice. Preliminary Lessons from Africa", Africa Region Working Paper Series No. 28, February 2002 The World Bank: "Madagascar -Diagnostic Trade Integration Study", volume 1August 2003 Ministry o f Basic and Secondary Education of the Republic o f Madagascar (MINESEB): "Tableaux rkcapitulatifs sure 1'Ctat de lieu informatique des CISCOs", Antananarivo 2003 Ministry o f Basic and Secondary Education o f the Republic o f Madagascar (MINESEB): "Annuaires Statistiques 1999-2003", Antananarivo 2003 Ministry o f Basic and Secondary Education o f the Republic o f Madagascar (MINESEB): "Annuaires Statistiques 2004", Antananarivo 2004 Ministryo f Basic and Secondary Education o fthe Republic o fMadagascar (MINESEB): "Plan National EducationPour Tous", Antananarivo 2003 Ministry of Basic and Secondary Education of the Republic of Madagascar (MINESEB): "Rapport Fiche d'Enqu&e Rapide (FER)", Antananarivo 2003 Mistiaen, J. Berk Ozler, T. Razafimanantena, and J. Razafindravonona: "Putting Welfare on the Map in Madagascar" Africa Region Working Paper Series No. 34. World Bank. Washington 2002. Mita, Miara : "Etude sur les besoins financiers des Communes a Madagascar", Antananarivo 2002 National Statistics Institute (INSTAT), "Enquete Auprks des MCnages 2002 (EPM 2002); RCsultats Preliminaires", Policy Brief, Juillet 2003 OSIPD : ((Etude sur 1'Etat des Lieux du RCgime Disciplinaire et de la DContologie du Personnel de l'Etab), Antananarivo 1998 Overseas Development Institute: "Implementing a Medium-term Perspective inBudgeting in the context o f national Poverty Reduction Strategies", undated -105- Oxford Policy Management and Overseas Development Institute: "General Budget Support Evaluability Study, Phase 1 - Final Synthesis Report", Report to the UK Department for Intemational Development (DFID), December 2002 Rarivoarivelomanana, Josiane and Florence Miauton: "Les Relations entre le FRAM, L e FAF, et 1'Ccole". Etude rCalisCepour la Mission RCsidente de la Banque mondiale a Madagascar 2004 Republic o f Madagascar: Poverty Reduction Strategy Paper and Action Plan, update July 2003 Romani, Mattia : "Growth and poverty in Madagascar, 1999-2001: Impact analysis in a context o f imperfect data comparability", 2004. The World Bank and the International Monetary Fund: "Madagascar: Poverty Reduction Strategy Paper: Joint Staff Assessment (JSA)", October 2003 The World Bank and the Intemational Monetary Fund: Decision Point Document for Enhanced Highly IndebtedPoor Country Initiative -Madagascar, Washington, December 2000 The World Bank Report, "District-level Service Delivery inRural Madagascar", July 2003 The World Bank: ((Madagascar - A Strategy for High Growth and Poverty Alleviation)), Economic Strategy Note 1994 The World Bank : Education et Formation Madagascar : Vers une politique nouvelle pour la (( a croissance Cconomique et la reduction de la pauvretCD, Washington 2002 The World Bank : Rkpublique de Madagascar - Etude de la Gestion des Finances Publiques (( )) Country FinancialAccountablity Assessment), Washington 2003 The World Bank, Les rapports Cconomiques de la Banque Mondiale : "Education et Formation a Madagascar, vers une politique nouvelle pour la croissance Cconomique et la rkduction de la pauvretC", M a i 2002 The World BankKorea Development Institute: "Reforming the Public Expenditure Management System", March2004 The World Bank: " Guinea - Strengthening Public Expenditure Management for Poverty Reduction and Growth", Public Expenditure Review June 2004 The World Bank: " Service delivery in public schools in Madagascar: Results of a budget tracking survey", draft report 2, August 29,2003 The World Bank: "Country Public Procurement Assessment Report" Report No. 26410, Vol. I& 11,Washington, December 2002 The World Bank: "Madagascar - Trackmg Poverty-Related Spending Assessment and Action Plan", Washington 2003 -106- The World Bank: "Memorandum and Recommendation o f the President o f the Intemational Development Association and to the Executive Directors on Assistance to the Republic of Madagascar under the EnhancedH P C Debt Initiative", Washington 2004 The World Bank: "Report and Recommendation o f the Managing Director to the President o f the International Development Association on a Proposed Interim Trust Fund Credit in the amount Of SDR 48.6 million to the Republic o f Madagascar for a structural adjustment credit," Report No, P-7013 MAG, Washington 1997 The World Bank: Madagascar Decentralization. Report No. 25793-MAG, Washington 2003 The World Bank: Poverty and Socio Economic Developments 1993-1999; Report No. 23366- MAG,Washington, 2001 The World Bank: Performance Audit - Madagascar - Ports Rehabilitation project (1752-MAG 1987) The World Bank: Seventh Highway project (1905-MAG 1989 The World Bank: ThirdRailway project (1694-MAG 1986) The World Bank: World Development Report 2004: "Making Services Work for Poor People", World Bank Washington 2003 Patemostro, Stefano, Razafindravonona, Jean, Stifel, David: "Changes inPoverty inMadagascar: 1993-1999", Africa Region Working Paper Series Number 19, July 2001 s 8 3 8 s8 3 8 38 x + k 2a e, 8 .3 a z u E I 5e, p! . . 8 .$3 .3 Y 0 2 I 3 I ru O s ? A A n 3 3 a 0Ei M G 0 8 2B - 0 0 m I W 2 3 I L u W U W 5 W rl 8 3 AE A \o 00 N 8s 8 ;Ei ?E 3E -117- Annex 2: Madagascar's Poverty Reduction Strategy Madagascar submitted a full PRSP to the IMF and World Bank in July 2003, which mapped out an ambitious program to reduce poverty by half inten years through rapid and sustained growth. Achieving this objective would entail three strategic axes: (i) improving governance; (ii) promoting broad based growth; and (iii) providing human and material security. The Joint Staff Assessment found the overall strategic focus to be clear, with coherent sector development plans built on work program agreements with the ministries involved. Most importantly, there was a high degree o f ownership and commitment. At the same time, the JSA noted that only limited costing had taken place, that in many cases the analytical underpinnings were weak and that the government would need to prioritize strengthening capacity in order to implement the program successfully. Interms structural policies to promote growth, the JSA proposed major investments in public transport infi-astructure; commitment to public-private partnerships to improve the management of public enterprises; and reforms intax and tariff policies, land laws and better governance. The macroeconomic outlook inthe PRSP anticipated sustained growth o f 8-10 percent. However, the JSA viewed this as overly optimistic in light o f limited government capacity, possible financing shortfalls, institutional constraints and a weak financial sector unable to support the demand for credit. Thus, while the proposed macroeconomic framework was basically sound, the JSA proposed growthtargets be revised down119. The preliminary investment costs of PRSP, as estimated by Government, came to about MGF 16 trillion, or approximately US$2.5 billion for the 2003-2006 period."' By comparison, the 2001 GDP came to MGF 30 trillion, or approximately US$4.7 billion, in 2001 exchange rates.'*l Planned investment costs for the 2003-2006 period amounted to about half of the 2001 GDP. Government anticipated that the PRSP would be implemented in a high growth environment. GDP growth was estimated to reach 6 percent in2003 (as the country exited the crisis) and 8 percent by 2006. Growth would come from all sectors o f the economy, but would rely primarily on agriculture, the export processing zone (textiles), public works, transportation services, fishery, mining and tourism. Table 2.1 provides some details. The Development Policy Review provides an elaborate analysis. To attain the anticipated growth rate, the PRSP anticipated that the investment rate for the economy would increase from 17.5 to 23.2 percent o f GDP over the period 2003-2006; this comes to an average annual investment rate o f 21.2 percent o f GDP for the four year period. In contrast, gross national savings expected to average 16.9 percent o f GDP for the 2003-2006 period; this implies that Madagascar would require average annual net imports, and external current account deficit, o f about 4.3 percent o f GDP for the 2003-2006 period, or net imports o f MGF 6 trillion in 2002 prices. The external current account deficit would be financed, about half by Government borrowing and about half by capital grants to Government. The external accounts include inflows o f about 1.3 percent o f GDP a year in foreign direct investment, which i s equivalent to the average annual external reserve accumulation. The PRSP ` I 9IDA and IMF 2003, "Madagascar: PovertyReductionStrategy Paper: Joint Staff Assessment (JSA)". I 2 OThe PRSP presents investment costs for the 2003-2006 period in nominal prices. In 2002 prices (time of drafting of the PRSP) ,this comes to about MGF 14trillion. 12' The reader is advisedthat most comparisons for the PRSP expenditures (2003-2006) are not made against 2002 since that was the crisis year and the economy and the public sector came to a halt, but rather against 2001 which i s considered a normal economic and budget year. -118- does not explicitly discuss monetary policy, but assumes an average annual inflation rate o f about 5 percent o f the 2003-2006 period. The macroeconomic framework reserves a very significant role for the Government sector over 2003- 2006 period. Government i s expected to spend on average about 18 percent o f GDP a year in non- interest expenditures for the 2003-2006 period, compared to 15.4 percent o f GDP in 2001. At the same time, Government's revenues are expected to increase to reach to 12.5 percent o f GDP in2006, up from 10.1 percent o f GDP in 2001, while interest payments are expected to half to about 1percent o f GDP by 2006, inlarge part due to the HIPC debt relief program. The overall fiscal balance i s expected to average -4.8 percent o f GDP for the 2003-2006 period, up from -4.4 percent o f GDP in 2001. The PRSP does not explicitly discuss the financing o f the deficit. It is important to place the admirable ambitions o f the PRSP in the context o f the previous accomplishments of the Malagasy economy. Inthe 1997-2001 period, GDP per capita growth amounted to about 6.1 percent on a cumulative basis; inthat same period, the national headcount index fell from 73 to 68 percent. The implied poverty elasticity for the 1997-2001 period was close to close to 1.1; this means that for a one percent increase inper capita income, the head count index fell by 1.1percent. For the 10 year growth period (2003-2013) the PRSP i s seeking to initiate, the 9.3 percent GDP growth would amount to a cumulative increase inper capital income o f about 84 percent. A concurrent reduction inthe headcount index from 71.6 percent to 34.2 percent would imply a poverty elasticity o f 62 percent.'22 This seems less ambitious than the 1997-20001 experience. However, one must look at the composition o f the growth and the composition o f poverty reduction between urban and rural areas. Over the 1997-1999 period rural poverty actually increased while urban poverty fell; in the 1999-2001 period rural poverty fell, but by much less than urbanpoverty. Macroeconomic outcomesfor 2002-2003. After the 6 month economic crisis inthe first half o f 2002, the Malagasy economy contracted by 12.7 percent. Inflation reached almost 16 percent and investment fell from 18.5 to 14.3 percent o f GDP. The external current account balance worsened from -1.3 to -5.7 percent o f GDP. On the fiscal side, revenues collapsed from 14to 10percent o f GDP, as did expenditures, from 17.6 percent the previous year, to 15.6 percent in2002. The year 2003 saw an unexpectedrecovery. GDP growth reached 9.8 percent compared the planned 6.0 percent growth that had been planned. Although the high grow rate was due to a catch-up effect (relative to the 2002 crisis), it raised significant expectations about the mediumterm. Public expenditure increased to 18.2 percent o f GDP while revenues (and grants) reached 14.4per percent o f GDP, matching the pre- crisis 2001rates. Although the political crisis o f the year 2002 was a setback implementation o f the PRSP i s proceeding reasonably well. A first progress report on the PRSP was submitted in 2004; it indicates reasonable progress on all strategic axes o f the PRSP. 122 This comparative analysis can best be done by looking at extreme poverty instead o f the headcount indeed, but measures of extremepovertyare not availableinthe PRSP. -119- Table 2.1: PRSPMacroeconomic and Fiscal Framework, Prepared July 2003 National Accounts and Prices RealGDPmpgrowth 6.0 6.0 4.0 4.5 5.0 Inflation, CPI (p.a.) 7.4 8.0 7.0 6.5 6.0 Primary Sector 4 -1.8 2.3 2.5 2.8 Secondary Sector 7.6 17 4.9 5.6 6.3 Zone (EPZ) 40 50 10 11 12 Tertiary Sector 6.1 9.4 4.6 5.2 5.7 Public Works 13.5 20 12.5 13 14 Goods Transportation 5.7 11.5 5 5.5 6.5 GDPpm (MGF bill, cur.Pri) 29,843 33,978 37,811 42,081 46,836 Per CapitaGDP (cur USD) 302.0 307.0 316.9 326.6 338.7 National Accounts (YOGDP) Gross DomesticInvestment 18.5 17.5 15.4 16.1 20.2 Public Sector 7.3 7.3 7.4 7.9 8.4 PrivateSector 11.2 10.2 8.0 8.2 11.8 GrossNational Savings 17.2 12.9 11.2 12.0 16.6 Fiscal WOGDP) Total Revenue& Grants 14.0 15.0 13.4 13.7 14.4 Total Revenue 10.1 10.6 10.3 10.6 11.3 oiw Taxes 9.7 10.3 10.0 10.2 10.9 Grants 3.9 4.4 3.1 3.1 3.1 Total Expenditure 17.6 18.2 17.5 17.8 18.0 oiw InterestDue 2.0 1.8 1.3 1.2 1.1 o/w Current (excl. interest) 8.1 8.4 8.6 8.5 8.3 oiw Capital 7.3 7.3 7.4 7.9 8.4 StructuralReforms (net) -0.7 -0.1 -0.1 0.0 0.0 Overall Balance (comm.) -4.3 -3.3 -4.2 -4.1 -3.6 ExtemalCAB (% GDPj CAB excl. Official Transfers -2.0 -7.3 -5.2 -5.2 -4.6 CAB incl.Cur. Off. Transfers -1.3 -4.6 -4.2 -4.2 -3.6 ACB incl. Cur. & Cap. Transf. 3.7 -1.1 0.3 0.4 1.o Gross Off. Res. (mill of ..) 317.5 324.0 466.5 570.2 697.5 Inweeks of G&S 14.4 15.2 20.8 22.9 25.6 ExchangeRate (p.a.) MalagasyFrancsiSDR 8,391 9,308 9,750 10,241 10,696 Source: Madagascar PRSP -120- Annex 3: Assessment of Madagascar'sbudgetmanagementcapacity to track poverty-reducing spending'23 The four benchmarks that were met indicate that: (a) the state's activities are not financed significantly by improperly reported extrabudgetary sources; (b) budget expenditure is classified on a detailed administrative, economic, and functional basis in conformity with the Classifications o f the Functions o f Government (COFOG); (c) multi-annual forecasts are integrated into the budget formulation process; and (d) tracking surveys are being carried out. Significant weaknesses, however, do exist, particularly inthe areas o f in-year budget tracking and execution. The main weaknesses are: (a) the substantial gap between budget outturn and the initial budget; (b) no clear identification o f poverty-reducing spending in the budget law or in execution tracking (only HIPC Initiative expenditure is clearly identified); (c) domestic arrears remain substantial and are not fully reflected; (d) internal control procedures are still inefficient; (e) accounting and banking data are not reconciled; (f) fiscal reports are received with excessive lags; (g) the complementary period i s still overly long; (h) the draft budget execution laws (lois de reglement) are not submitted to Parliament within 12 months after the end o f the fiscal year. A. BUDGET FORMULATION Indicator 1.Coverageof the budgetor fiscal reporting entity Benchmark 1:Fiscal reporting covers the Government Finance Statistics definition o f the general government sector, Le., including central, regional, and local governments, and all government operations, whether funded through the budget or not Assessment: The benchmark is not being met. General budget coverage i s limited to central government as defined in the 2001 edition o f the GFSManual. Local government budgets and their execution are not consolidated with those o fthe central government Indicator 2. Degree of spending being funded by inadequately reported extra budgetary sources Benchmark 2. Government activities are not funded through inadequately reported extra budgetary sources to a significant degree (less than 3 percent o f total expenditure) Assessment: The benchmark is being met. Extra budgetary funds exist and information on execution i s not consistently accurate or transparent. The overall amount, however, i s considered insignificant inrelation to total expenditure (less than 3 percent) 123The assessment is based on a common framework comprising 16 indicators, that is applied to all beneficiary countries o f IMF and World Bank Heavily Indebted Poor Country (HIPC) Initiative. An assessment and action plan was completed in 2001 and an update was performed in 2003. A comprehensive re-assessment was conducted in 2004. The 2004 assessments use a new 16" indicator on public procurement. Accordingly, the rating for 2004 i s higher for five benchmarks than in2003. On the whole, a modest progress inthe results o f the 2004 assessment was observed as compared with the progress registered in2003. -121- Indicator3. Reliabilityof budgetas a guideto future Benchmark3. Budgetoutturn data are quite close to the originalbudget. Assessment. The benchmark is not being met. The overall budget utilization rate is low, particularly where capital expenditure i s concerned Indicator4. Inclusionof donor funds. Benchmark 4. Budgets and/or fiscal reports at all levels o f government include, without exception, grants projected to be provided by donors, and the capital and current expenditure o f all multilateral and bilaterals on government activities. Assessment. The benchmark is not being met. Meeting this benchmark requires that all activities financed by donors and lenders be fully reported ex ante, and in a timely manner, on an ex post basis: which i s not the case. Indicator5. Classification. Benchmark5. Budget expenditures are classified on an administrative, economic, and detailed functional and programmatic basis. Assessment. The benchmark is being met. The budget nomenclature meets international standards. It includes an administrative, economic, functional, and program classification. Indicator6. Identificationof poverty-reducingspending. Benchmark6. Poverty-reducingexpenditures are clearly identified. Assessment. The benchmarkis not being met. With the exception o f HIPC Initiative spending, other poverty-reducing expenditures are not clearly identified in the existing budget classification, and the PRSP priorities are not generally reflectedinthe budget law. Indicator7. Integrationof medium-termforecasts. Benchmark7.Multiyear expenditureprojectionsare integratedinto the budget formulation process. Assessment. The benchmarkis beingmet. Projected multiyear expenditure is integrated into the budget formulation process. -122- B.BUDGET EXECUTION Indicator 8. Evidence o f budget execution problems-Arrears. Benchmark 8. Small stock o f expenditure arrears, with little accumulation o f arrears over the previous year. Assessment. T h e benchmark is n o t being met. Indicator 9. Effectiveness o f the internal control system. Benchmark 9. Internal control i s effective. Assessment: the benchmark is n o t being met. Internal control on the conformity o f expenditure to laws and regulations i s relatively satisfactory, but remains insufficient to improve the quality o f budget and financial reporting, as well as the efficiency and effectiveness o f expenditure. Indicator 10. Tracking surveys are in use. Benchmark 10. Tracking surveys are used, where necessary, to supplement internal control, but may not yet be a regular feature o fthe PEMsystem. Assessment: T h i s benchmark has been met. Tracking surveys are conducted to supplement internal controls. Indicator 11.Quality o f fiscal information. Benchmark 11.Satisfactory reconciliation o f fiscal and banking records i s undertaken routinely. Assessment. This criterion i s n o t being met. Reconciliationstatements are not being prepared. C. FISCALREPORTING Indicator 12. Regularity o f timely internal fiscal reporting. Benchmark 12. Internal budget reports are received within four weeks o f the end o f the relevant period. Assessment. This benchmark i s n o t being met. Internal fiscal reports are filed with lags exceeding four weeks. Indicator 13. Periodic fiscal reports t r a c k poverty-reducing spending. Benchmark 13. Good-quality classification o f poverty-reducing spending i s reflected in the in-year budget reports. Assessment. This benchmark i s n o t being met. Poverty-reducing spending cannot be tracked as it i s not identified. -123- Indicator 14. Transactions are recorded inthe accounts in a timely fashion. Benchmark 14. Routine transactions are entered into the main accounting system(s) within two months after the end o f the fiscal year. Assessment. T h i s criterion i s n o t being met. The supplementary fiscal period exceeds the two- month lag provided by law, although it i s in fact less than six months. Indicator 15. Timeliness o f audited financial information. Benchmark 15. An audited record o f the financial outtum should be presented to the legislature within twelve months from the end o fthe fiscal year. Assessment. T h i s benchmark i s n o t being met. The latest budget review law passed by Parliament dates back to 1996. D.PUBLICPROCUREMENT Indicator 16. Efficiency and effectiveness o f the public procurement process. Benchmark 16. The procurement system promotes efficiency and effectiveness in the expenditure o f public funds through clear and enforceable rules that promote competition, transparency, and value for money. Assessment. This benchmark is n o t being met. -124- Annex 4: Reformof the PublicFinanceManagementSystem GovernmentActionPlan PLAND'ACTIONS PFUORITAIRES2004 15 octobre 2004 Domaines Indicateursde Dates debut Lesponsablesde d'activites Actions A mettreen aeuvre rksultatsfin 2004 et fin la mise Etat d'avancement fin en aeuvre septembre2004 ~~ luivi et Crier une cellule de .a cellule crkee, ses Fevrier G Cellule crCCe depuis Juin 2004 2oordination des suivi et de coordination des nembres recrutes -Coordonnateurdes !Cformes rkformes au sein du MEFB RCformes nommi et dCja et definir son cahier de ?Cvrier opkrationnel : consultant charges ruin recrute au titre du PGDI ProcCder au Uncadre encours de recrutement des membres recrutement. de la cellule 2oordination avec vlettre en place un mecanisme Iccord entre le FCvrier G Accord entre bailleurs de es bailleurs de fond )ennettant d'assurer le dialogue dEFB et les fonds et MEFB Ctabli pour echnique entre le MEFB et les )ailleurs de fonds une reunion trimestrielle. L a )ailleurs de fonds sur les :urle mecanisme premitre reunion a eu lieu l e -iformes 21 septembre 2004 sur les problkmes de la Douane (voir compte rendu de reunion a ce sujet) tenforcement des - Definir les termes de rermes de rCfCrence Fevrier ;G Encours :apacitCs techniques refkrences determinant ipprouvks iuMEFB la strategie de - formation du MEFB Zonsultant engage Avril Recruter un consultant 3ystime integrk de - Faire receptionner l e >es fonctionnalites Fevrier ;G Dtveloppement du logiciel de Sestion des volet A par le Comitk ncluses dans l e 'resident du programmation effectuC dans Finances Publiques de pilotage y inclus l a iolet B son1 :omit6 de l e cadre de la preparation du :SYGFP) validation technique el itilisees dans le site FCvrier i 'ilotage budget programme 2005 I'evaluation de )h SYGFF AoQt XI l'environnement. - mplante Mettre en Ceuvre Avril A Juin immediatement l e Fevrier - volet B du SYGFP i Finaliser lec Juin procedures en prenani en compte les progrkc - de l'informatisation Septembre i DCcider la plate forme DCcembre technologique er prenant en compte le: choix ou option: technologiques dL Gouvemement - Etendre le SYGFP dans quatre (4) Chefs lieux de Province -125- Domaines Indicateursde Dates debut Responsablesde d'activitks Actions mettreen ceuvre rksultatsfin 2004 et fin la mise Etat d'avancementfin en oeuvre septembre2004 Preparation de la - Consolider le processus de La loi de Finances 4vril DGDP ,oi de Finances 2005 en cours Loi de Finances preparation budgktaire pour la 2005 est prCparCe le preparation presentee sous Loi de Finances 2005 (y compris selon cette brme de Budget Programmes -leEtablir calendrier de preparation) reglementation Mars a DGDP les phases et activitks Le processus de 3ctobre de mises en Oeuvre de budgets - mise en Oeuvre des programmes budgets programmes dibute Chaine de la - Elaborer les textes pour i)la Les conditions pour Juin DGDP DCcret n"2004-571 du 4 Juin depense I fusion des fonctions de sous la fusion sont 2004 dkfinissant les Fusion des ordonnateur et de gestionnaire remplies ittributions et la responsabilite fonctions de sous de credit et ii)le statut de cette fie I'Ordonnateur dans les ordonnateur et de cattgorie de personnel Septembre phases d'exkcution de l a gestionnaire de - DCfinir le plan e formation Plan de formation fiepense credits dans l e cadre de la strategie Ctabli Septembre Statut des corps des -globale du MEFB. Administrateurs Financiers Identifier les Ministkres Pilotes Ministtres Pilotes dCsin6s Chaine de la - Elaborer un arret6 fixant la Cette nomenclature FCvrier B DGDP Dkcret nO2004-282 fixant 15 depense I1 nomenclature des pikes entre en vigueur Juin nomenclature des pitce: Nomenclature des justificatives de la dipense dans les services justificatives des dossiers 5 pitcesjustificatives - Diffuser cette nomenclature au opkrationnels Juillet A soumettre au visa du CDE sein des services budgktaires concemks Novembre En cours de finalisation proje concemks (Ordonnateurs, de dCcret concemant toutes le! Comptables et ContrBleurs) chaines de la dkpense Nomenclature d6jB publie( dans le JORM Chaine de Depense - Diffuser l e manuel 1er semestre DGIGE Manuel deja diffuse depuis le 111 d'inspection aux inspections el 05/07/2004 a I'heure actuelle Manuels de ContrBles d'Etat le` semestre DGT mise jour compte tenu des procedures - Diffuser l e manuel de nouveaux textes ComptabilitC GCnCrale du Tresor Les services on1 1er semestre - Diffusion et formation de mai Diffuser l e manuel directement accks 51 juillet 2004 d'inspection destine a la Brigade aux textes financiers zesemestre SG -duAmenager au sein du MEFB TrCsor fondamentaux Diffusion et formation de mai Les services 1 semestre DGDP 51juillet 2004 un centre de documentation utilisent les manuels -accessible le a tous les Services Diffuser manuel d'exkcution du budget au niveau de contrBle Circulaire budgetaire financier no519lMEFB du30/12/2003 -126- Domaines Indicateurs de Datesdebut Responsablesde d'activitbs Actions imettreen oeuvre rbsultats fin 2004 et fin la mise Etat d'avancement fin en oeuvre septembre2004 lIarchCsPublics - Finaliser et dCposer a -e cadre ICgal et Mai Juin DGDP ,oi no2004-009 du 26 juillet 1'Assemblke Nationale le projet nstitutionnel des ,004 portant code des de Code des Marches publics - narchCs publics est darch6s publics PrCparer et diffuser les .enf o r d MUS a PrCsident ComitC documents types affirents B la iuillet de Pilotage 'dR des Consultants aregu un nouvelle rkglementation .vis favorable de la Banque- - Signer le programme de Avril ippel B Manifestation renforcement du cadre I'interh paru dans les -Institutionnel la formation du Juin a luotidiens de la Capitale- DCmarrer dCcembre )ate de clbture fix6e au 30 personnel par la mise en ceuvre eptembre2004 du nouveau Code des march& avec priorit6 aux 6 Ministttres :ormation dCmarrCe dtts Pilotes inalisation des documents YPes reglementaires l'application du Code de$ vIarchCs Publics jestion de la Solde - DCfinir les TdRs pour Le MEFB Avril DGDP It des Pensions I'Cvaluation de la gestion de la :ntreprend une Solde et des Pensions dans rCflexion sur l e I'objectif de sa dkconcentration jtratCgie de cette Juin a - Recruter un Consultant et iCconcentration octobre rCaliser cette Cvaluation 3estion de la dette Former l e personnel concern6 Les donn6es de li Mars a DGT Formation de : )ublique extCrieure sur le logiciel Sygade dette publiquc octobre agents de la DDP B Genttvl extCrieure son (finmars) mises fi jour Mise 9 jour el continu, au fur et a mesur des tirages f remboursements Attente r6conciliation ave l'ancien systkme -127- Domaines Indicateursde Datesdebut tesponsablesde d'activitks Actions mettreen oeuvre ii resultatsfin 2004 et fin la mise Etat d'avancementfin en oeuvre septembre 2004 Zomptes de 1'Etat . Finaliser la balance consolidee Clette balance est ruin )GT Balance provisoire 2003 deja 2003 des Comptes du Tresor :ransmise a la Cour ClaborCe les Comptes La balance definitive sera transmise ala Chambre des . Finaliser les comptes de Cles comptes de Septembre Comptes dans la loi de zestion 2002 ;estion sont deposks Reglement 2003 ila Chambre des Xcembre .Finaliser les lois de rkglement Clomptes 87 % Budget General transmis 2002 et 2003 a la Chambre des Comptes Comptes de gestion transmettre a fin septembre ruin la Chambre des Comptes centralises a 1'ACCT et en - Deposer les lois de rkglement cours d'examen 1997 et 1998 a I'Assemblee Loi de rkglement 2002 deja Nationale transmise a la Chambre des Comptes finjuin L o i de Reglement 2003 A transmettre 21 la Chambre fin octobre Loi de Reglement 1997 dejf adoptee par I'AssemblCc Nationale Loi de Reglement 1998 dCposCe au Gouvemement (i verifier) _____ Documentation - Definir une strategie visant a Les textes financiers Mai DGDP financiere accroitre la disponibilite de la fondamentaux et l e documentation financibre au manuels de sein des services du MEFB procedures son1 DCcembre - Implanter un centre de rendus l a documentation disponible aux disposition des agents du MEFB agents du MEFB -128- Domaines Indicateursde Dates debut iesponsables de d'activitks Actions h mettre en oeuvre resultatsfin 2004 et fin la mise Etat d'avancement fin en oeuvre septembre2004 Irgane de ContrGle - Eriger la Brigade du TrCsor en )es contrdles sont ruin >GT . Brigade du TrCsor deja Direction :xerces sur les :rigCe en Direction suivant le 3rigade du Tresor Iostes comptables FCvrier a iemier organigramme du - Recruter 8 vkrificateurs IuTresor Decembre MEFB dCcret no2004-570 du -suppltmentaires des Avril 9 le`juin2004 Mettre cellules DCcembre 5 VCrificateurs deja recrutbs - d'inspection dans chaque Mars 3 a recruter d'ici dCcembre TrCsorerie Principale DCcembre .Cellules dejamises en place - Former le personnel chargC des jans chaque TrCsorerie verifications Principale - Formation terminCe enjuillet 2004 3rgane de contrdle - PrCparer les decrets instituanl >es conditions Juin 3G Decret n"2004-573 du le` Juir :IInspection et organisant I'IGF ikcessaires au 2004 portant creation 3CnCrale des Tonctionnement organisation e ?inances - Recruter 8 inspecteurs sur les :orrect de I'IGF Juillet $ 3G fonctionnement de I'IGF -15Former les inspecteurs recrutCs postes prevus DCcembre Septembre P Liste des inspecteurs dCji DBcembre pr&te Attente nomination en Consei des Ministres Une premikre formation des Inspecteurs d'Etat sera faite les 11 et 12 Octobre 2004 par les IGF franqaise et Marocaine Formation sous 1'6gide du Organe de contrdle Entreprendre la formation de: Les connaissance: Avril i DGCDE CRROC en 2003. I11: Contrdles Contrdleurs financiers dans 1c techniques de! octobre Actuellement : formation en Financiers Cadre de la Strategie dc Contrdleurs cours (CF) formation du MEFB financiers son amdioris -129- Domaines Indicateursde bates debut Lesponsables de d'activitb Actions A mettreen oeuvre rbsultatsfin 2004 et fin la mise Etat d'avancement fin en oeuvre septembre2004 lecettes I Rediger et diffuser les textes ,es services fiscaux 4ai A IGI Mise la disposition des mp6ts I'application du CGI rccroissement leur ICcembre usagers du Bulletin Accroitre la documentation naitrise de la 4ai a Fiscal officiel n02 et lisponible dans les Services 6gislation fiscale et ICcembre preparation de I'Cdition iscaux et a l'attention des LmCliorent leur `Cvrier a duBFO n"3 :ontribuables endement )Ccembre Distribution aux Etendre I'informatisation des 4ars a inspecteurs du PCG 2005 :entres fiscaux .ovembre et du CGI a tous les Editer et disskminer les personnels. Mise a la nanuels de procedures fiscales disposition des guides pratiques fiscaux a l'attention des contribuables ; L'objectif de 15 centres fiscaux A informatiser cette annte a et6 largement depasse, 12 realisation actuelle est de 20CF informatisks. Debut de la validatior auprts des centre: fiscaux: 15 Octobrc 2004 Zecettes I1 . Installer Sydonia++ dans 3 >a Douane amCliore Vovembre )GD L'installation dc Iouanes mreaux de Douanes a gestion des Sydonia++ dans le! . Elever le niveau de formation .ecettes douanitres Bureaux de Douane: iu personnel douanier dans le :t Cltve le niveau de d6butera en janvier 2005, :adre de la stratkgie globale du Zonnaissance du fait du retard enregistr6 MEFB .echnique de ses 9vril 5 dans I'exCcution du Projet . Etendre la mise en place de ?,gents Xcembre (mai 2004 au lieu de pichet unique dans 2 bureaux janvier 2004). Au titre de fie Douane l'annee 2004 les activitks porteront sur la construction et le test du prototype du logiciel a la DGD et au niveau du site pilote (Tamatave Port) L a strategic de formation de personnel est en cours d'klaboration. Toutefois, 1'Ecole Nationale des Douanes B Tamatave reprendra ses activites avant la fin de I'AnnCe 2004, avec l a formation de 45 Cltves inspecteurs des Douanes. L e guichet unique est install6 dans le Bureau des Douanes de Diego et de Tulear depuis le 2e trimestre 2004. -130- Annex 5: StatisticalTables (Public Finance) Table 5.1.: Madagascar: Real annual growth ratesin sectoralbudget allocations1997-2004 (percent change in share of budgetary allocations) Aver. 1998 1999 2000 2001 2002 2003 2004 97-04 97-02 Total allocation 18.2 24.3 21.9 12.3 9.3 (9.5) 17.7 13.5 17.2 NonInterest Expenditure 16.0 29.6 20.0 18.1 6.3 (10.4) 15.5 13.6 18.0 Gouvernance 26.8 51.2 10.2 11.5 (11.8) 9.7 9.0 15.2 17.6 General Administration 28.0 58.1 11.7 11.3 (15.8) 5.6 6.8 15.1 18.7 olw budget, finance & eco.adm 4.4 128.6 1.5 11.8 (19.3) 22.3 8.7 22.6 25.4 Defense 56.2 (12.6) (1.3) 11.7 15.4 34.6 19.9 17.7 13.9 Public security (7.0) 66.6 2.8 14.6 12.8 24.8 15.4 18.6 17.9 oiw Justice 97.5 30.2 (19.0) 43.0 4.7 7.9 15.4 25.7 31.3 Growth 0.3 7.3 26.5 34.6 30.8 (34.7) 20.6 12.2 19.9 Energy (59.4) 149.4 (12.7) 30.2 (15.7) (3.3) (0.7) 12.5 18.4 Agriculture (3.7) 5.1 10.4 6.5 12.1 (28.8) 22.2 3.4 6.1 Environnement (2.3) 2.4 7.4 62.5 39.4 (33.2) 6.8 11.8 21.9 Transport & Communication 10.0 (10.3) 46.5 65.0 7.1 (25.0) 33.4 18.1 23.7 olw Transport 15.5 (10.7) 48.0 66.5 8.2 (26.6) 34.0 19.2 25.5 Others 28.5 21.9 43.9 6.9 128.3 (54.9) 4.6 25.6 45.9 Social 19.0 19.1 33.2 14.6 9.9 (6.5) 21.0 15.7 19.1 Education 39.8 20.2 32.3 14.0 22.6 (5.3) 40.1 23.4 25.8 o/w PrimaryEducation 91.5 18.7 10.0 (7.2) 134.1 12.1 11.2 38.6 49.4 Health 4.1 28.3 37.5 18.6 0.6 (11.7) 2.3 11.4 17.8 oiw Primary Healthcare (3.8) 19.2 77.6 (32.2) 38.0 (29.7) 33.5 14.7 19.8 Others 8.6 7.8 29.8 10.4 (1.8) (1.8) 0.3 7.6 10.9 Source: MFEB, World Bank -131- Table 5.2.: Madagascar: Real annual growthratesof current and capitalbudget allocations(1997- 2004) (percent change in share of budgetary allocations) 1997 1998 1999 2000 2001 2002 2003 Total 84,2 96,2 79,7 79,4 75,O 43,3 91,9 Administration 74,3 83,6 56,6 793 77,7 51,2 86,8 National Defense 98,8 68,l 100,9 103,7 97,5 93,O 95,2 Public security 97,9 131,7 93,O 97,O 93,3 78,l 99,7 Education 103,9 101,3 112,3 83,9 78,l 63,5 97,O Health 85,4 101,3 88,8 65,6 68,l 37,7 97,l Social protection 56,6 91,7 84,O 81,7 72,6 68,9 82,O Urbanism 62,O 66,3 84,2 50,O 46,8 24,l 64,6 Culture 64,2 76,8 62,8 101,9 79,3 68,O 86,6 Energy 33,7 174,O 73,3 112,4 69,9 8 2 62,9 Agriculture 78,6 121,8 79,9 84,5 72,7 23,4 84,6 Environment 124,O 167,7 205,2 195,6 143,9 41,7 85,l Industry 100,9 70,2 78,8 92,5 74,l 21,5 83,l Transport 94,8 98,7 92,O 53,O 51,l 21,5 89,3 Other economic affairs 126,4 116,3 102,3 81,O 99,l 17,9 168,O Not classified 101,8 38,7 99,0 733 92,3 19,5 96,7 Yource: MEFB, Banque Mondiale -132- Table 5.3: Functionalrepartitionof Budgetallocation1997-2004 (% ofnoninterest expenditures) 1997 1998 1999 2000 2001 2002 2003 2004 Total 100 100 100 100 100 100 100 100 Administration 32.1 35.5 43.2 40.3 37.9 30.1 35.5 32.8 National Defense 3.4 4.6 3.1 2.6 2.4 2.6 3.9 4.1 Public Security 4.1 3.3 4.2 3.6 3.5 3.7 5.2 5.2 Education 10.5 12.6 11.7 12.9 12.5 14.4 15.2 18.4 Health 8.8 7.9 7.8 8.9 9.0 8.5 8.4 7.4 Social Protection 0.9 0.7 0.6 0.5 0.6 1.o 0.6 1.3 Urbanism 5.9 6.2 4.9 5.4 4.8 4.1 5.2 3.8 Culture 1.5 0.9 1.o 1.1 1.1 0.9 0.8 0.6 Energie 3.2 1.1 2.1 1.6 1.7 1.4 1.5 1.3 Agriculture 10.4 8.6 7.0 6.4 5.8 6.1 4.9 5.2 Environment 3.1 2.6 2.0 1.8 2.5 3.3 2.4 2.3 Industry 1.4 1.2 1.1 1.o 1.o 5.7 1.5 1.3 Transport 11.9 11.3 7.8 9.5 13.3 13.4 11.2 12.9 Other economic affairs 3.0 3.6 3.4 4.4 4.0 4.9 3.9 3.5 Not classified 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 'ource: MEFB, WorldBank -133- Table 5.4.: Madagascar's budget executionrates 1997-2003 (actual expenditures as percent ofbudgetary allocations) 1997 1998 1999 2000 2001 2002 2003 Total 84.2 96.2 79.7 79.4 75.0 43.3 91.9 Administration 75.2 84.9 56.7 78.9 76.8 50.6 87.1 National Defense 98.9 68.1 100.9 103.7 97.6 93.0 95.3 Public Security 97.9 131.7 93.0 97.0 93.3 78.1 99.7 Education 104.0 101.2 112.2 84.5 79.9 65.2 97.3 Health 78.2 96.8 88.6 59.9 67.8 37.9 92.2 Social Protection 69.5 91.7 92.0 76.2 46.8 25.1 73.1 Urbanism 65.5 64.8 84.9 38.0 42.8 28.3 65.4 Culture 95.7 97.0 97.5 109.3 78.9 69.3 88.3 Energie 23.O 137.0 72.9 106.5 66.4 11.1 62.1 Agriculture 75.7 118.7 81.6 86.2 73.1 25.2 82.1 Environment 123.1 171.7 205.5 203.9 146.5 41.7 84.9 Industry 62.7 57.7 66.3 45.3 34.4 15.9 74.5 Transport 104.6 108.6 94.2 59.0 60.4 25.0 100.2 Other economic affairs 98.8 99.3 92.7 85.8 96.7 21.1 166.4 Not classified 101.8 38.7 99.0 73.5 92.3 19.5 96.7 Source: MEFB, WorldBank Table 5.5: Publicexpenditures of selected Sub Saharan African Countries 2004 (as a % of GDP) Countries Total gov. expend. Sub SaharanAfrica 25.9 Gabon 30 Senegal 21.8 Tchad 21.5 Madagascar 19 CGte d'Ivoire 16.5 Cameroon 15.6 Source: WorldDevelopment Indicator 2004 Table 5.6: Share of some sectoral allocations (%of GDP) Health Education Wage bill SSA Cameroon Mali Senegal Madagascar 192 Source: WorldDevelopment Indicator 2004 -134- Table 5.7: Madagascar: Real annual growth rates of actual expenditures 1997-2003 (percent change in share of actual expenditures) Aver. 1998 1999 2000 2001 2002 2003 97-03 97-01 ExecutedNonInterestExpenditure 37.6 10.7 19.7 6.3 (36.8) 90.7 21.4 18.6 Gouvernance 37.7 10.2 45.1 8.6 (37.2) 71.3 22.6 25.4 General Administration 44.1 7.1 56.7 8.7 (44.4) 78.9 25.2 29.2 olw budget, finance & eco.adm 31.0 28.3 71.7 14.3 (48.6) 116.9 35.6 36.3 Defense 7.7 29.4 1.4 5.1 10.0 37.9 15.2 10.9 Public security 25.1 17.6 7.3 10.2 (5.5) 59.2 19.0 15.0 olw Justice 96.1 40.2 (13.6) 26.6 (36.0) 93.6 34.5 37.3 Growth 29.7 (9.9) 6.5 18.9 (59.0) 181.8 28.0 11.3 Energy 109.3 5.1 33.8 (19.1) (90.1) 642.4 113.6 32.3 Agriculture 49.1 (31.O) 16.8 (8.4) (63.9) 157.0 19.9 6.6 Environnement 32.2 25.3 2.4 19.6 (59.6) 36.2 9.3 19.9 Transport & Communication 14.5 (16.4) (15.6) 59.0 (55.0) 211.7 33.1 10.4 olw Transport 20.3 (17.4) (16.2) 62.6 (55.3) 211.8 34.3 12.3 Others 13.9 12.3 23.9 21.1 (5 1.8) 228.3 41.3 17.8 Social 29.2 25.2 (0.8) 9.7 (19.9) 67.8 18.5 15.8 Education 36.3 33.2 (1.2) 6.1 (0.4) 44.5 19.8 18.6 olw Primary Education 58.9 107.3 (24.4) (14.0) 66.6 45.6 40.0 32.0 Health 23.5 12.4 1.6 23.2 (44.3) 127.5 24.0 15.2 oiw Primary Healthcare 42.0 12.9 (12.0) 12.2 (44.5) 73.2 14.0 13.8 Others 22.3 25.1 (3.3) 0.2 (31.4) 76.8 15.0 11.1 Source: MFEB, WorldBank -135- Annex 6: The PhasesandProceduresof the BudgetCycle I.Phaseadministrative(ordonnateur) a) Engagement: c'est l'acte qui engage juridiquement et financikrement I'autoritC publique. Les cinq Ctapes de l'engagement de la dCpense 1. Expression dubesoin par le gestionnaire de crkdit. Ce dernier organise une mise en concurrence ; selon le cas (i.e. en fonction du montant de la commande et conformkment aux seuils fixks par la rkglementation) : simple consultation de trois fournisseurs ou Appel d'offres. 2. Propositiondes foumisseurs. Selon le cas (cf. ci-dessus) propositionpar facture pro forma ou devis ou rkponse a Appel d'offres. 3. L e gestionnaire de crkdit effectue le choix du fournisseur. Selon le cas (cf. ci-dessus) : visa sur la facture pro forma ou mise en place d'un march6. A cette ktape, le gestionnaire de crkdit elabore un dossier d'engagement qui comporte : 0 une demande d'Engagement Financier, viske par le Gestionnaire, 0 untitre d'engagement financier, nonencore vis6 par leGestionnaire, 0 la pro forma du foumisseur choisi, viske par le gestionnaire, ou le marchk vis6 par l'autoritk compktente, 0 le Bon de commande, non encore date. I1transmet ce dossier de projet d'engagement au Contrale des Dkpenses Engagkes (CDE). 4. L e CDE exerce ses contrdles de conformitk sur la forme et sur le fond. Notamment, a ce titre, il vkrifie : la correcte imputationbudgktaire, la disponibilitk des crkdits, la cohkrence administrative des supports, le respect de la mise en concurrence, la moralitk des prix. L e CDE vise : le TEF, la facture pro forma et le Bon de commande ; ilconserve la DEF. L e visa sur le TEF peut comporter une rkserve : celle-ci devra &re levke lors de la certification de la facture. L e CDE retourne le dossier au Gestionnaire de crkdits. 5. L e Gestionnaire des crkdits vise le TEF et le Bon de commande (cachet, timbre et date) et procede a l'engagement de la dkpense en adressant au foumisseur un exemplaire du TEF et le B o n de commande ou 1'Ordre de service s'il s'agit d'un marchk. b) Liquidation :ils'agit de la vbrificationdu service fait et de la fixation du montant de la dCpense. Les deux Ctapes de laliquidation de la dCpense 1.Le foumisseur exkcute sa prestation et produit sa facture definitive. 2. L a phase de liquidation fait intervenir le depositaire comptable et le gestionnaire des crkdits et, le cas kchkant, le CDE : L e dkpositaire comptable procede a la rkception des fournitures ; en fonction de la nature des fournitures illes prend (ou non) en charge en comptabilitk matibre. L e gestionnaire des crkdits procede aux actes de liquidation : verification arithmktique et cohkrence des pieces ;certification de la facture. L e CDE, dans les cas ou son visa comportait une rkserve, peut participer a larkception. -136- c) Ordonnancement :c'est l'actepar lequelle sous-ordonnateurdonnel'ordre au comptablepublic de procCder au reglement de la dkpense.Les deux &apesdel'ordonnancement de la dCpense 1. Le gestionnaire de credits prepare le dossier d'ordonnancement ; ce dossier comprend : le bordereau des pikces, les pikces justificatives, le projet de mandat de paiement, le bordereau &emission et le (ou les) bon(s) de caisse ou le (ou les) avis de credits. Si le visa du CDE sur le TEF Ctait avec reserve le gestionnaire de credit transmet le dossier au CDE pour validation definitive : le CDE appose la mention "Confirmation de validation" sur le bordereau des pikces. L e gestionnaire de credits transmet ce dossier au sous-ordonnateur. 2. L e sous-ordonnateur prockde au mandatement aprks verification du dossier d'ordonnancement : conformite des pikces avec le bordereau des pieces qualite (i.e. originaux) des p i k e s coherence du dossier (identite dubenkficiaire sur toutes les pikces) contrble arithmktique des pikces (en chiffies et en lettres) conformite des signataires avec specimens. I1 vise les mandats, les bons de caisse et/ou avis de credit, appose la mention "Vu" + cachet et sa signature sur le bordereau des pikces. 11.Phasecomptable(comptablepublic) Paiement :c'est l'actepar lequel1'Etat (ou la collectivitk locale)se libkrede sa dette. Les deux Ctapes de la phase comptable de la depense 1. L e comptable assignataire vkrifie la correcte mise en ceuvre des lois et rkglements relatifs 51. la dkpense publique ;ilcontrble notamment : l'existence des credits la regularit6 de l'engagement, de la liquidation et de l'ordonnancement la concordance entre les pikcesjustificatives et les titres de paiement, et entre ces titres et le bordereau d'emission la coherence arithmktique du dossier le caract6re liberatoire dupaiement L e comptable assignataire appose le cachet "Vubon 51. payer" sur le titre de paiement. 2. Le comptable met la dkpense en paiementpar caisse oupar virement. i d a s I I d m 3 I I I I C 1; iE C c ! I e, CO ~ e, .-Ea '-i 3 z 0 0 e, e, -0 a 0 ._ C 0 c) * e, .-'cr g 8a ,: h L 0 rh E .r .ev 2 c I < 1 CI 0 b, L t -0 a, 142 Annex 9: The MalagasyEducationSystem- Structure andIndicators I.Organisationalandfunctionalset-up Structure of the Educationsystem. The government restructured its education sector administration in 2003/04. The three separate education sector ministries (basic, vocational and higher education) were combined into one: the Ministry of National Education and Scientific Research (MENRS). Adult and Literacy education, which comprises over 80 percent of the population above 15 years o f age, fall under the responsibility of the Ministry o f Population (MINF'OP). In 2003 the government also updated its national Education For All (EFA) strategy, which now comprises the objectives o f (a) universal primary education (of 5 years duration, defined as the first basic education cycle), and (b) a second basic education cycle ounior secondary education cycle with a duration of 4 years), with an adequate enrolment'24 and satisfactory quality. The tables below present the distinct levels and responsibilities in Madagascar's education system: Level Organisation Center Ministryo f Secondary and Basic Education(MENRS) Province Interregional Directorate o f Secondary and Basic Education (DIRESEB) (Faritany) District District EducationOffice-Circonscription Scolaire (CISCO) :Fivondronana) ~ Commune Zone administrative et pkdagogique (ZAP) (Firaisana) Public primary school (EPP) Lower secondary school-- Collbge d'enseignement genkral (CEG) Higher secondary school (Lycke) Vi11age Public primary school (EPP) (Fokotany) Parents-school partnership association (FAF) Association o f parents o f students (FRAM) lZ4EFA implementation depends on donor financing. The MENRS strategy includes a satisfactory enrolment for secondary levels, based o n the needs for economic and social development as defined in its PRSP. Estimates are that this would require about 40-50 o f the relevant age group to be enrolled and complete this junior secondary cycle. This would also secure a sufficient teacher supply for the primarylevel. -143- Level Function MENRS Sets national policy and does country-wide planning for the sector. Prepares annual investment and operating budgets. Allocates resources to the provinces and school districts. Carries out financial and technical oversight, and quality control o f public education institutions, and monitors private institutions. Manages personnel, including hiring, promotions, and firing. Manages the insurance fund for school-related accidents (PASCOMA Protection Accidents Scolaire de Madagascar). Sets curriculum content and pedagogical standards. Assures teacher training and slulls upgrading. Collects, analyzes, and reports on education data and statistics. DIRESEB Conducts province-level planning, and preparation o f the annual plan. Tracks and reports on provincial education indicators. Does financial and technical oversight o f districts and their schools. Allocates staff within the province. Provides training and technical support for districts and schools. Serves as intermediary for CISCO reporting to MINESEB. CISCO Supervises pedagogical activities, administration, and expenditures o f thl primary and secondary schools inthe district (EPPs, CEGs, and LycCes). Maintains student records and manages the annual examination for promotion and award o f diplomas. Allocates civil service staff withinthe district. Manages teachers hiredo n a contract basis by communes and FRAMs. Prepares an annual district workplan. Manages the non-salary portion o f the recurrent expenditure budget. Manages the collection and accounting procedures for school fee: PASCOMA premiums, and other Caisse Ecole activities. Handles distribution o f supplies and equipment to schools through FAFs. Compiles and reports o n educational statistics for the district. ZAP e Serves as the administrative and technical interface for the CISCO with ab01 10-20 schools Provides oversight o f and support to school directors. Facilitates informationtransmission to and from schools. Assures monitoring at the school level. Assists with the distribution o f school supplies and equipment. Interacts with members o f the FAF and FRAM. -144- EPP, CEG, Manages the instructionprovidedto students. Lycte Director Supervises teachers, bothregular civil service employeesand contract teachers (enseignants suppltants). Maintains schoolrecords. Transmitseducationalstatistics for their schools to the CISCO via the ZAP. Oversees andaccounts for supplies and equipmentprovidedby the CISCO. Serves as a designatedmember of the FAF management committee. Interactswith the FRAM and individual parents. Community organization (FRAMs and FAFs). At the local level, there are two community organizations in the education sector. The FRAMi s the association o f parents o f students. It i s supported by voluntary contributions from its members; in communities whose schools do not have enough teachers, FRAMshave hired teachers on a contract basis, paying them with a combination o f money, bags o f rice, and donated agricultural labor. FRAMmembers also provide in-kmdsupport to school operations and rehabilitation, volunteering to carry materials and supplies from CISCO drop-off points to remote schools where vehicles cannot reach, and contributing labor to school projects as needed. FRAMleaders are elected by the community. The FAF, a government-community partnership organization for school development (know by its Malagasy acronym), was created by MINESEB decree in September, 2002, largely in response to the need for a formal organization to receive HIPC funds.'25Its partnership structure combines civil servants (school directors) with elected community members to manage resources devoted to support educational establishments through a Caisse Ecole, whose transactions are publicly postedto assure transparency. Private sector role. Private providers inprimary education play an important role inMadagascar, but the MENRS has not yet developed the capacity nor the legal and administrative framework to provide incentives and a basis for effective cooperation with the private schools. Many urban and rural-isolated communities contribute to the establishments o f primary schools (private or public) and pay for a teachers (often uncertified). The so-called FRAM teachers are estimated to number about 16,000 in 20003-4, compared to a total of about 62,000 staff employed by the MENRS (of which about 38,000 are actually teaching). 11.Madagascar'seducationsystem Education systems in Africa usually vary in structure and management. In the case o f Madagascar the system i s structured as follows (see chart below): '25 I t s predecessor, the FRAM or parents' association, does not have a formal legal status and thus cannot receive funds. -145- MadagascarBasicEducationStructure(January 2004) Basic Education: First Cycle Preparatory (EF 1) Course r------- (2 years) 1 5 years In PrimaryPublic Schools (EPP) Elementary Course I Medium Course Basic Education (5+4 Years) Basic Education: Second ~~ (EF II) Courses 4 years in "Colleges d'EnseignementGeneral" Orientation ---- (CEG) -- (2 vears) ,Courses . Certification Exam (BEPC) - - ~ - ---- --_I-Entrance Examinations Istyear: common classes Senior Secondary 2n year. --_^_I_ b T - - Education (3 years) Branches : One cycle Literature- Economics - Science - In "Lycees d'EnseignementGeneral" Secondary General High Schools --~. Social Branches : Literature- Economics - Science - Social ~ - ~ - - Certification Exame (BAC ;) Source: MENRS website: www.mitieseb.pov.ntphvebsitdstedi~eg.is0 -146- GeneralEducation - Pre-school or pre-primary school: This level i s offered to children 5 years old and younger. Enrolment at this level i s very low. The majority o f the pre-schools inthe country are private. - Primary school orJirst cycle of basic education: This level i s offered to children older than five. It is composed o fthree sub-cycles. The first, or preparatory cycle, includes the first two years o f primary (grades ll* and lo*). The second or elementary cycle i s only one year (grade 9") and the final or medium cycle includes the last two years o f primary (grades 8* and 7*). - Basic or elementary education has beenredefinedby the government as consisting o f two basic'26 cycles: the first basic education cycle (5 years o f primary education), and the second basic education cycle (4 years o fjunior secondary or the "College"). About 85 percent o f the primary age group was enrolled in 2003-04, but only 40% o f the cohort i s currently completing primary school (one of the lowest rates inthe region). Transition to the junior secondary cycle requires a successful exam at the endo f primary. - Lower secondary or second cycle of basic education: To be able to continue to this level o f education, students most pass a primary certification exam known as CEPE. This cycle, previously known as collkge, includes four years o f education. - Upper secondary school or Lyce`e: Students enter at this level o f education if they pass an entrance exam and if they have also passed a lower secondary certification exam known as BEPC. - Higher education: To enter in higher education institutions students most pass a high school certification exam known as BACC. This level i s open to general high school graduates and some students from vocational and technical high schools. This level o f education i s offered by universities, almost all public, and lately by post-secondary professional and technical institutions. Vocational and TechnicalEducation Besides general education, the Malagasy system includes vocational and technical branches, mostly at secondary school level. The education and training offered by these branches vary in duration o f the courses, the pre-requisites o f entry, and the certification they offer. The two levels are: (a) Centers of Vocational Training (Centres de Formation Professionnelle); these centers offer training to primary school graduates (students do not need to pass the CEPE to enrol in some o f these centers); and (b) Vocational and Technical High Schools (Lyce`es Techniques et Professionnelles); these centers offer training to junior education (collkge) graduates. 126Ths new structure comprises two "cycles fondamentals" or basic education cycles, defining basic education under the government's EFA strategy as 9 years (5 years primary and 4 years junior secondary). The new Education Law was approved inlate 2004. -147- 111. Administrative proceduresto pay a supplier At the beginning of the expenditure chain the credit administrator, inthis case the Chef CISCO issues a spending request and sends it for approval to the Control des Dkpenses Engagks (CDE). The CDE carries out a form o f ex ante control, i.e. verifying that funds are available and that procurements will be made fiom eligible budget lines. Once the service i s rendered, the normal procedure o f payment entails three administrative steps that need to be followed. First, the credit administrator, in this case the Chef CISCO, has to verify that the service was rendered. This step might involve long delays as CISCOs are usually located at long distances fiom some o f the schools under their supervision. These long distances make it easier for the CISCOs to procure goods and services to those schools closer to the district center, leaving the ones further away under-provided. Second, the Chef CISCO has to send a file o f the expenditure to the sous-ordonnateur for approval. Third, in case o f approval, the sous- ordonnateur has to sendthe expenditure file to the local treasury which will then pay the supplier. Thispayment procedure has three major drawbacks: (i) delays to pay a provider; (ii) obligation long the for district staff to frequently visit control organs that can be at some distance from the CISCO (these are only 18 CirconscriptionsJinanciers inthe country, hence only few CISCOs have control organs close by); and (iii) inability o f the CISCO to track the status o f a payment once a file has been sent to the sous- the ordonnatuer, which can create difficulties with localproviders when payments are delayed. There i s a special and faster procedure to pay suppliers calledDdldgation de Cridit where the approval of the sous-ordonnateur i s not needed for the payment to be rendered. This procedure allows CISCOs to open accounts at the local treasury office (usually at the district level) rather than at the circonscription financier. This allows the CISCO to keep better track o fpayments and reduces the overall time needed to release the funds. For instance, in 2001 the MADIO/INSTAT survey found that on average it took 27 days between the assessment o f the expenditure and the payment o f the supplier, 13 days between the assessment, and the approval and 14 between the approval and actual payment. This delay was longer when the normal procedure was followed, 28 days vs. 10 days, and in case the payments was done through a deposit ina bank account rather than when paid in cash. Nevertheless, neither procedure to pay suppliers (the normal or the Dildgation de Crddit) entail an efficient control o f the "real" sources o f leakage identified during the traclng survey studies. For example, CDEs and sous-ordonnateurs are not required to carry out any regular control o f procurement procedures and prices (and in the cases where they do they usually do not have updated price lists at hand).'" Additionally, neither the CISCOs nor the control organs are required to keep inventories o f delivered goods that would permit to track the use and distribution o fpurchased equipment over time. The regular control organ for this issue (the Cellule Centrale des Marchies) i s only consulted in the case o f procurements above a certain threshold. These however, are rarely met by CISCOs 148 Annex 10: Statistical Tables (Education Sector) Table 10.1: Distributionof staff in2004 Source : MENRS Table 10.2: CISCO criteriato distribute suppliesto schools Source: Budget TrackingSurvey, WorldBank,April/Mai 2003;limited sample of 24 CISCOs. P I = Priority 1 P2 =Priority 2 P3 =Priority 3 149 Table 10.3: Per primary student expenditureacrossFivondronanasrankedby average per capita income 2001 Faritany Quintile Antananarivo Fianarantsoa Toamasina Mahajanga Toliara Antsiranana I 131,762 205,799 145,077 186,715 I1 112,150 189,827 125,154 104,661 276,350 107,833 I11 144,313 161,085 127,043 162,991 348,680 IV 152,592 123,736 190,735 193,474 96,807 V 224,493 355,881 194,811 322,246 436,175 167,807 Total 163,032 191,302 149,568 226,363 297,886 145,365 Source: This table was constructed using actual exuenditure data from 2001 at CISCO and EPP level and data from thepoverty map exercise (Mistiaen, et. al. 2002) I Table 10.4: Studentsinrural areas across type of school ownership in 2001 (proportions) Quintiles Schooltype I I1 I11 IV V Total PublicPrivatelyrun 0.03 0.04 0.04 0.03 0.00 0.03 Public 0.84 0.89 0.79 0.79 0.49 0.79 Private 0.01 0.01 0.06 0.07 0.24 0.06 Private nonlucrative 0.01 0.01 0.03 0.02 0.10 0.02 Confessionary 0.09 0.04 0.09 0.09 0.17 0.09 Community School 0.03 0.01 0.00 0.00 0.00 0.01 Total 1.oo 1.oo 1.oo 1.oo 1.oo 1.oo Source: EPM 2001 Table 10.5: Teacherspaidby FRAMs across years 2001102 2002103 2003104 Number of FRAMteachers 5,849 7,107 12,323 Source: MERS 2004. Report FER 150 Annex 11: The Environment Sector inMadagascar 11.1. Biodiversity, deforestation and the protected area (PA) system Madagascar i s an island o f 590,000 km2that separated from Mainland Africa during the Jurassic. Madagascar's long history o f isolation i s reflectedinthe highlevel o fbiological endemism. O f the 12,000 plant species present, 9,700 are endemic, and more than 80 percent o f those are forest or woodland plants. The island includes over 250 species o fbirds, over 40 percent o f which are endemic, and over 115 species o f mammals, 90 percent o f which are endemic. Of the 189 amphibian species 187 are endemic. Inmost cases the majority of those animal species live exclusively in either the humid forests o f the eastern mountains and coastal zone, as well as the deciduous forests and woodlands o f the southern and western coasts. The loss o f forest cover over the last few decades has been a cause o f concern. The origin o f forest conversioni s mainly slash-and-bum agriculture (tavy) to produce rainfed rice inthe Northern and Eastern regions, and unsustainable fuel wood collection practices for domestic energy in the Western and Southern regions of the country. These two destructive practices are, in addition, accompanied by extensive exotic flora and fauna (and marginally pharmaceutical plants) collection on the periphery o f cleared areas. Farmers income generated by forest conversion i s relatively low because it occurs on marginal agricultural lands (most o f the remaining forests are situated in the highlands) and i s not sustainable. Several estimates o f national forest cover have been made usingremote sensing data collected duringthe past 50 years. An assessment based on aerial photography fi-om 1949 produced an estimate o f 17 million hectares o f total forest. In 1988, based on digitizing over LANDSAT image prints fi-om the middle 1970s, the cover o f all forest were estimated at 10 million hectares. Another recent estimate o f forest cover based on classification o f 1 km-resolution data from the SPOT satellite also near 10 millions hectares. These estimates suggest that over one-third o f the forest area has been destroyed since 1950. A first satellite-based estimates o f forest andwoodland clearance for the entire island was made with data for circa 1990 and circa 2000 o f LANDSAT images processed by NASA. Results show that deforestation continued but at slower pace than before NEAP implementation: "only" 1 million hectares o f remaining natural forests has been lost during the last ten years and deforestation rate i s near zero in the protected area system. There i s no comprehensive inventory o f the stock o f biodiversity o f 46 protected areas. An inventory o f the endangered vertebrate animal species encountered i s currently under way. Although middle altitude dense humidforests are currently over represented against dry and thomy forests, the representation o f the island forest habitats i s assumed to be satisfactory. Moreover, it i s believed that the system shelter thousands o fplant species that are unique on earth and 38 o f the 45 inventoried lemur species. -151- 11.2. Donor-SupportedEnvironmentProgramsinMadagascar Madagascar's National Environmental Action Plan (NEAP)has provided the basis for a 20-year program o f investment aimed at assisting the Malagasy population to protect and enhance the environment for more sustainable development. The sequence o f three Environment Programs (EP) has been explicitly designed to tap both local and global conservation finance and, inthe case o f EP2 and EP3, to learn from previous experience inprogramimplementation. The key features o f eachprogram are described below. First Environment Program 1990-1997 EP1 assisted Madagascar inputtinginplace the legal and institutional framework and the skills and tools needed to manage its environmental heritage. Major goals included the conservation o f sites o f world renowned biodiversity, reductions in soil loss and the implementation o f environmentally sound programs o f rural development. The project comprised 7 components : 1) protecting and managing unique ecological systems together with development o f peripheral zones; 2) promoting soil conservation, agro-forestry, reforestation, and other rural development activities inpriority zones, including several large watershed areas; 3) developing maps and geographic information; 4) improving land security through titling; 5) training environmental specialists and promoting environmental awareness and education at all levels o f society; 6) launching environmental research programs on land, coastal and marine ecosystems; and 7) developing support activities, composed o f institution building, adoption o f environmental assessment procedures, strengthening the environmental database, monitoring and evaluation, and studies. The $85.5 million dollar programwas financed as follows: EP1 Sources o f Financin Other Donors Government 17.2 Total 85.5 Second Environment Program 1997-2003 The objectives EP2 were to reverse current environmental degradation trends and to promote sustainable use o f natural resources, including soil, water, forest cover and biodiversity. Another key objective was to create the conditions for environmental considerations to become an integral part o f macroeconomic and sectoral management o f the country. It aimed to continue and strengthen activities already launched under the first phase and initiate work innew areas where environmentproblems are important. The objective o f the Global Environment Facility (GEF) support to the program was to curb the loss o f globally significant biodiversity. The program consisted o f components dealing with field operations, strategic activities and support activities, as follows: 1) sustainable soil and water management; 2) multiple use forest ecosystem management; 3) development o f national parks and ecotourism; 4) marine and coastal environment management; 5) urban environmental management; 6) support to local natural resource management; 7) support to regional development programs and spatial analysis; 8) creation o f a regional fund for environmental management; 9) upgrading o f the legal framework and formulation o f environmental -152- policies; and 10) assisting sector ministries in implementing environmental policies and making environmental impact assessment (EIA) operational. Major sources o f funding included the following: EP2 Sources ofFinancing $m IDA 30.0 GEF 20.8 IFAD 8.1 Other donors 49.6 On-goingprojects 15.5 Government 31.0 Total 155.0 ThirdEnvironment Program 2004-2009 The Third Environment Program Support Project (approved by the World Bank in M a y 2004) aims at improving the protection and sustainable management o f critical biodiversity resources, mainstreaming conservation into macroeconomic management and sector programs, and establishing sustainable financing mechanisms. It seeks to assist the Government o f Madagascar in implementing selective elements o f the environment program, for which two subsidiary development objectives have been specified: conservation and sustainable management o f eco-regions with active multi-stakeholder participation; and strengthening o f environmentalmanagement by incorporating environmental objectives inpolicy making andpublic investments. The project comprises three main components: 1) Forest ecosystem management will support the formulation and implementation o f forest zoning, forest control, and, the establishment o f an information system. Activities will improve forest governance, and strengthen the allocation o f concession rights, and the institutional framework for regulatory enforcement. Conservation sites will be created for the preservation o f biodiversity, and watersheds; transfer o f forest management rights to local communities will be fostered; reforestation practices shall create land reserves; and, household energy (charcoal and biomass) production will be improved. 2) Management o f protected area systems, partially supported by the Global Environment Facility (GEF), will provide capacity building, increase public awareness, and enhance community participation, aligning resources to expand and manage ecosystems under the national protected area system. Conservation management programs will monitor activities, and critical infrastructure and services to stimulate ecotourism will be developed. Finally, financial resources will support biodiversity conservation over the long-term. 3) Establishment of an environmental information system, development o f training materials targeting communities, and the mass media, and support to the Direction Generale de 1'Environnement (DGE) inenvironmental training and dissemination activities. A key innovation o fEP3 is the creation o fthe ConservationTrust Fund, with IDA anddonor resources, in order to increase the sustainability o f funding for conservation in Madagascar. The principal sources o f financing for EP3 are as follows: 1 ::::1 EP3 Sources ofFinancing $m IDA 40.0 GEF NGO Other donors 65.7 Government 18.5 Tntal 150.0 -153- 11.3. New institutionalframework ofthe environmentsector Level of the Ministry of Environment and Forestry General Directorate for General Directorate for General Coordination of Environment (DGE) Water and Forests (DGEF) Projects (CGP) Policy-making and Policy-making and Coordination of donor coordination coordination projects Level of the autonomous agency ONE (Quasi-governmental organization) EIA andpollution management, ONE would be largely self-financing out o f environmental permits that it would issue, but could be funded by MinEnvEFto carry out specific functions, including (i)establishment o f an environmental information center, (ii)preparation o f environmental accounts, and (iii)establishment o f sectoral `observatories,' such as OSF, monitoring governance issues. ONE would continue to be the principal agency executing the MECIE (EIA) rlprrw Level of the executing agencies ANGAP ANGEF (Quasi-governmental organization) (executing agency) Establishment, operation and maintenance of Forest resources management theprotected areas system ANGEF would enforce silvicultural obligations for commercial operations and engage in reforestation activities Community level -154- Annex 12: StatisticalTables (Environment) Table 12.1.: ANGAP: Average annualcosts andfinancingof protectedareasduringEP3 2004-2009 (in MGF billion) costs Financing Capital investment 21.0 Government 2.4 - expansion o f PAS 2.1 Entrance fees 1.7 - biodiversity management 11.6 ODA 30.0 - ecotourism infrastructure 6.5 7.8 - public PA trust fund awareness 0.8 Current expenditure 32.2 Total costs 53.2 Total financing 41.9 Source: FFHMFinance -Biodev (2004), EP3 Task Force (2003) Table 12.2.: A comparisonof revenuesin2001 and potentialecotourismrevenues (in MGF billion) 2001 Potential Total internal revenues 16.9 18.3 Entrance fees 1.4 2.8 PA trust fund 15.5 15.5 External revenues 21.2 43.5 Visa fees 18.7 18.7 Hotel tax 2.5 24.8 Total revenues 38.1 61.8 Source: WorldBank -1.55- Table 12.3.: Qualitativeassessment ofinstitutionalculture in the environmentsector I .... .... DGEF I DGE IANGAP IONE ..... ..... Mission/ strategic Dosition .... I Claritv with resnect to environment Dolicv Ie. I e. I I I ...... ..... ~~~~Clarity with respect to EP3 e.. ..eo Mission reflected inorganization structure e. e. ..eo Mission within EP3 accepted and shared e. ..eo Objectives .... .... Clarity o f objectives ..eo e.e.0 ...... ...... Obiectives reflected inaction plans ..eo ..... ...... ...... Acceptance o f objectives within organization I ? I ? ..... ...... ...... ..... ...... ...... Results orientation ..eo ..... Client orientation I ? I ? ............ .... Accountabilitv ..eo Presence `on the ground' na Adeauate staffing and level of skills e. e. ..eo Source: World Bank Notes: Scoring on a scale from 1to 6; o indicates ?4score. Table 12.4.: Qualitativeassessmentof organizationalfunctions in the environmentsector Source: World Bank Notes: Scoring on a scale from 1to 3; o indicates % score. -156- Table 12.5: Time to processEnvironmental ImpactAssessments (average no. of days) Year Eligibility Evaluation Permitissuance assessment 1997 154 196 8 1998 42 121 47 1999 26 185 20 2000 16 167 38 2001 22 230 133 2002 33 266 103 2003 47 192 76 2004 23 90 39 Source: ONE Table 12.6.: Potentialcosts andpermit fees for EnvironmentImpactAssessments ofpublicinvestments (MGF billion) 2004 2005 2006 Total Productive sector investments Agriculture, livestock, fisheries 65.9 39.3 38.1 143.3 Industry 0.0 4.1 0.0 4.1 Mines 0.0 0.0 0.0 0.0 Tourism 6.8 0.0 0.0 6.8 Environment 0.0 0.0 0.0 0.0 Infrastructureinvestments Public works 598.7 2905.8 2315.9 5820.4 Air transport 0.6 22.9 0.0 23.5 Urban and rural planning 26.1 145.9 205.6 377.6 Telecommunications 0.0 125.0 120.0 245.0 Energy 41.4 1 .o 1.o 43.4 Water 0.0 0.0 0.0 0.0 Administration investments Public security 0.4 0.5 0.0 0.9 Public investmentspotentially subject to EIA 739.8 3244.5 2680.6 6664.9 less operational expenditures 148.0 648.9 536.1 1333.0 less foreign financed 0.0 1090.2 1136.6 2226.7 Totalpublic investmentssubject to EIA 591.8 1505.5 1007.9 3105.2 Total cost of EIAs ( 0.75% average rate) 4.4 11.3 7.6 23.3 Total permitfees ( 0.35% average rate) 2.1 5.3 3.5 10.9 Source: WorldBank Note: It i s assumed that the foreign-financedinvestmentshave their own EIA process, which are not included in the data. For each sub-sectorin the table, only those investmentswith a likelyrequirement for an EIA are included.