Report No. 29154-BUR Burkina Faso The Budget as Centerpiece of PRSP Implementation Public Expenditure Review June 25, 2005 PREM 4 Africa Region Document of the World Bank MFB MinistryofFinanceandBudget MESSRS Ministry of Secondary and Higher Educationand Scientific Research MOH Ministry of Health MTEF Medium-Term Expenditure Framework NGO Non-government Organization PAP PriorityAction Plan PER Public Expenditure Review PDDEB Ten-year Basic Education Development Plan PNDS National Health Care Development Plan PRGB Budget Management ReformPlan PRGF Poverty Reductionand Growth Facility PRSC PovertyReduction Support Credit PRSP Poverty Reduction Strategy Paper PRSP-PR Poverty Reduction Strategy PaperProgressReport ROSC Report on the Observanceof Standardsand Codes SP-PPF Permanent Secretariat for the Supervision ofFinancialPoliciesandPrograms STC-PDES Technical Secretariat for the Coordination of Social and Economic Development Programs TOFE Government Financial Operation Table UNDP UnitedNations Development Program UNICEF UnitedNations InternationalChildren Emergency Fund WAEMU West African Economic andMonetary Union WHO World Health Organization Vice President: Callisto Madavo Country Director: A. DavidCraig Sector Director: Paula Donovan Sector Manager: Robert R.Blake Task Team Leaders: Siaka Coulibalvljan Walliser ... 111 TABLE OF CONTENTS I INTRODUCTION . .............................................................................................................................. 1 A. Country Background........................................................................................................................ 1 B. . ProgressinBudget Management...................................................................................................... 3 I1 . PRSPPRIORITIESAND THE BUDGET1998-2002 .................................................................... 4 A. Productionof BudgetDataandResults Orientationofthe Budget.................................................. 6 B. Participatory NatureoftheBudgetingProcess.............................................................................. 12 C. Budget Allocations and Executionunderthe PRSP....................................................................... 15 D. Budgetary Spending and Donor Financing.................................................................................... 27 E. Conclusions and Recommendations............................................................................................... 30 111 INTEGMTINGBUDGETPROCESSESAND THE PRSPREVIEW . ...................................... 33 A. Program BudgetReform: Measures, Results andProspects.......................................................... 34 B. LinksbetweenProgramBudgetsandthe PRSP............................................................................. 40 C. lntegrating PRSP ProgressReports and ProgramBudgets: A Proposal ........................................ 42 D. Layingthe Foundation for a SuccessfulReform............................................................................ 44 IV HEALTHSECTOR POLICIESUNDERTHE PRSP:PROGRESSAND CHALLENGES 48 . .... A. Health Sector inBurkina Faso: Outcomes, Resources,and Utilization......................................... 48 B. Household HealthExpenditure andFinancialAccess toHealthCare........................................... 58 C. HealthFinancing andExpenditures ............................................................................................... 61 Figure 28. Delegated credit allocations and poverty rates..................................................................... 67 D. Government BudgetAllocation and ExecutionProcessinthe Ministry ofHealth........................ 70 E. Conclusions and Recommendations............................................................................................... V. THE EDUCATIONSECTOR INBURKINA FASO: PERFORMANCEAND PROGRESS 83 ..80 A. Education Trends and Outcomes.................................................................................................... 83 B. Education Financingand Expenditure Management.................. ......... ...................... 93 C. Program Budget Allocation and Execution...................... .......................................................... 98 D. Conclusions and Recommendations............................................................................................. 108 TABLES: Table 1:National Assembly's Priorities: 4 Percentage Change from Original Allocations ......... 14 Table 2: Education and Health Expenditures as a Percentage o f GDP......................................... 26 Table 3: List of PDDEB Objectives andIndicators for the Period2004-2006 ............................ 36 Table 4: AdditionalResources Available under the 2004-2006 MTEF....................................... 47 Table 5: Health Indicators inBurkinaFaso Comparedto Sub-saharan African Countries..........49 Table 6: Trends inMortality, Malnutrition andFertility .............................................................. 50 Table 7: Contraceptive Use and Knowledge among Married Women (percent) .......................... 50 Table 8: Evolution o f Child Mortality by Income andResidence................................................ 51 Table9: Evolutiono fHealth Infrastructure.................................................................................. 54 Table 10: Evolution o fHealth Sector Personnel........................................................................... 55 Table 11:Percentageof Individuals UsingHealth Services inpast Four Weeks......................... 56 Table 12: Evolution o f Health Indicators...................................................................................... Table 13: Incidence o f IllnesslInjury and HealthCare Utilization (past month).........................56 57 iv Table 14: Average HouseholdOut-of-pocket PaymentsonHealth. 1998. 2003 (CFAF per month. nominal terms) .............................................................................................................................. 59 Tabfe 15: EvolutionofUser Fees at Health Clinics (CFAF) ........................................................ 60 Table 16: Change inAverage Prices for Certain MedicalServices at the CMA Level (CFAF) ..60 Table 17: Total Private Spending on Health, millions of CFAF .................................................. 62 Table 18: Public Expenditures for Health, millionsofCFAF ...................................................... 62 Table 19: Budget Allocations by Level, millions o fCFAF (includes government budget, HIPC, external financing includedinbudget) ......................................................................................... 63 Table 20: FinancingofVaccination, Malaria, Reproductive Health, andNutritionPrograms in 2003 (billions CFAF) .................................................................................................................... 66 Table 21: Allocation of delegatedappropriations within districts (thousands of CFA) ...............68 Table 22: Total DelegatedAppropriations Allocated for Maintenance (millions ofCFAF) ......68 Table 23: Average Income andExpenditureso f CSPS by Category, 2002.................................. 69 Table 24: HIPC HealthAllocations andExpenditures (millionsofCFAF) ................................ 76 Table 25: Gross Enrollment Ratesby InstructionalLevel, 1998-2002 ........................................ Table 26: DistributionofEnrollmentby Type of Establishment (publiclpnvate), 1998-2002...83 86 Table 27: Primaryand Secondary Enrollment Ratesby Location, Sex and Standardo fLiving Quintiles, 2003.............................................................................................................................. 88 Table 28: Repetition andDropout Rates at the Primary Level, 1998-2003.................................. 89 Table 29: RepetitionandDropout Rates at the Secondary Level, 1998-2003.............................. 90 Table 30: Average Cost per StudentCompleting a Cycle ............................................................ 92 Table 31:Status of Student Grants, Aid and Loans for Higher Education, 200112002................93 Table 32: Sources ofFinancing for Education (inbillions o f CFAF), 1998-2002....................... 94 Table 33:EducationSector Expenditure Indicators, 1998-2002.................................................. 95 Table 34: Distributiono f Education Expenditure by Administrative and Operational Level, 1998- 2002............................................................................................................................................... 96 Table 35: Distributionof Education Expenditures by Instructional Level, 1998-2002................97 Table 36: Public Expenditures for Education, by Student andby Instructional Level, 1998-2002 ....................................................................................................................................................... 98 Table 37: DistributionofProgramBudget by Level of Education, 2000-2003 ........................... 99 Table 38: Distributionof Program Budget by Program Objective, 2000-2003.......................... 100 Table 39: Budget Execution for Basic Education, 1998-2002................................................... 101 Table 40: Execution of DelegatedAppropriations ofthe DREBA, Cascades Province, 2003 .. 104 ListofBoxes: Box 1: LessonsLearnedfrom the Use of ProgramBudgets in OECD Countries.................46 V ACKNOWEDGMENTS This report was prepared by a Bank team led by Siaka Coulibaly and Jan Walliser (AFTP4). The core Bank team included Rosa Maria Alonso Terme (WBIPR), Timothy A'. Johnston (AFTH2), Tankien Day0 (Consultant, AFTP4), Mercy Tembon (AFTH2), and Safaa El Tayeb El Kogali (AFTH2). Team support was provided by Judite Femandes (AFTP4), Suzanne RayaTsst Kabort, and BintouSogodogo (AFMBF). The report has benefited from the overall guidanceofEmmanuelAkpa andRobert Blake, Sector Managers (AFTP4) and A. David Craig, Country Director (AFC15), as well as from feedback provided by peer reviewers Vinaya Swaroop (SASPR) and Michael L. 0. Stevens (Consultant, AFTPR), andcolleagues from the InternationalMonetary Fund, The team benefited enormously from close cooperation with staff at the Ministry of Finance andBudget (MFB), the Ministry of Economy and Development (MEDEV), the Ministry o f Health (MOH), the Ministry o f Basic Education and Literacy (MEBA), the Ministry of Secondary, Higher Education, and Scientific Research (MESSRS), and the Ministry of Agriculture, Water and Fishery Resources (MAHRH). The team wishes to thank in particular staff at the General Directorate for Budget (DGB); the General Directorate for Cooperation (DGCOOP); the General Directorate for the Treasury and Public Accounting (DGTCP); the Secretariat for the Monitoring of Financial Programs and Policies (SP-PPF); the Secretariat for the Coordination of Economic and Social Development Programs (STC-PDES); the Directorate General for the Economy and Planning (DGEP); as well as staff at the Planning and Study Directorates (DEP), Finance and Administration Directorates (DAF), and Human Resource Directorates (DW) at the Ministries o f Health, MEBA, and MESSRS, the regional and provincial directorates of the MEBA and the Ministryo f Health in the Cascades region, and the headmaster o f the Lompolo Kone lyceum for their availability, close collaboration and feedback during the preparation of this study. The team also thanks the President of the National Assembly's Finance and Budget Commission and heads of several non-government organizations (NGOs) who sharedtheir views andperceptionsofthe budget process. vi EXECUTIVESUMMARY This study takes stock o f budget procedures and budget implementation after three years of PRSP implementation. It reviews progress in budget management and budget priorities, proposes a closer institutional integration o f programmaticbudgetingwith the PRSP process, and reviews recent developments and expenditure policies inthe health and education sectors. PRSPPrioritiesand the Budget1998-2002 The elaboration of the 2000 PRSP built on a longstanding participatory tradition and prior work undertaken by the government in 1995-98,notably the Letter o f Intent for Sustainable HumanDevelopment Policies. The PRSP focuses on four strategic pillars (accelerating equitable growth, improving social services, creating income for the poor, and improving governance), and three priority sectors: health, education, andrural development. A revised PRSP for 2004-06was adopted inOctober 2004, and it benefited from a fairly broad participatory process, as confirmed bycivil society organizations. The revisedPRSP substantially broadens priorityareas, which will pose a challenge for sufficient resource mobilization for its implementation. Productionof budgetdata and resultsorientationof the budget. The coverage of the central government budget is fairly comprehensive given the small size o f local governments. However, the central government budget does not cover the spending o f autonomous public institutions. The presentation o f budget data tracked inthe electronic budget management system relies exclusively on administrative and broad economic categories inabsence o f a hnctional or programmatic classification. A functional classification was introduced with the 2004 budget but does not identify priority spending categories. Spending financed with resources of the H P C Initiative has been tracked through a special account since 2001, separating HIPC-related expenditure fkom the remainder o f the budget and reducing the overall programmatic vision o f the budgeting process. Of particular concern is the weak reporting on foreign-financed investment spending, which is not tracked inthe computerized system and for which data is often not reported by donors using direct disbursement mechanisms. Budget controls and execution procedures are still overly centralized at the Ministry o f Finance and Budget. Some limited deconcentration o f payment-order issuance to the regions has taken place recently, but all procedures are still executed by directorates and staff o fthe Ministry o f Finance and Budget. Medium-term expenditure frameworks and programmatic budgeting have been introduced since 1998-99, with important progress made in the realism o f tax revenue projections. Program budgets are elaborated in all ministries but their quality varies and they are under-utilized, limiting their usefulness for results-oriented budgeting and giving limited incentives to line ministries to devote time and resources to program budgets. Overall, the PRSP is improving coordination within and between ministries. The PRSP has fostered a culture o f reporting on results of policy implementation but the culture o f evidence-based policymaking is still tenuous. The recent split o f the Ministries of Finance and Economy requires additional efforts to enhance coordination between financial planningand PRSP. Further improvements are needed as regards monitoring and evaluation o f policies, and the capacity and incentives for ministries to make use o f evaluation are still fairly weak. The revision o f the PRSP was a vii welcome opportunity to draw lessons fi-om the past three years in improving the monitoring and evaluation architecture, including better involvement ofcivil society and donors. Participatory nature of the budgeting process. Public availability of budget data is limitedalthough progress has beenmade inrecent years, includingthe adoption o fa new law on quality, availability, and timeliness o f budget data in January 2003. The budget process is formally transparent but reallocations made throughout the budget preparation process are not always based on clear criteria. Budget information is available to the government and donors during the year, but not to the wider public, which has to rely on PRSP progress reports and budget execution reports. Overall, hrther progress is needed inproviding timely information to the public on budget preparation and execution. The National Assembly plays a significant role in the budget process. However, it only discusses the input-based and does not receive the MTEF and program budgets. The National Assembly also does not discuss the PRSP or its progress reports and has shown limited ownership o f its contents inallocating additional resources to non-priority areas inthe past. Civil society participation remains need to be strengthened and the budget process does not include any entry points for its involvement. Information on the budget is not widely disseminated, and the capacity o f civil society to analyze the budget is weak. Moreover, the capacity o f local researchers (University) is under utilized compared with other PRSP countries. Budget allocations and execution under the PRSP. The PRSP process and the availability of HIPC funds have led to a substantial increase o f expenditure in priority areas (education, health, rural development) and similarly core priority areas (primary education, primary health, and rural development), with a doubling o f per-capita spending between 1998 and 2002. Expenditure increases were larger inpriority sectors than innon-priority sectors, with the education sector benefiting cumulatively from the largest share (50 percent) in additional expenditure in priority sectors. However, little o f the additional spending was for recurrent expenditure, raising concerns as regards the balance between investments and operational expenditure. The government has made considerable efforts since the adoption o f the PRSP to protect priority sectors from budget cuts in case o f resource shortfalls as execution rates o f priority sectors were systematically higher than those for non-priority sectors. As regards economic categories, the highest execution rates were for wages and salaries, and lowest for foreign- financed investment, Moreover, the share o f priority sectors in total discretionary spending increased over time. Concerning the PRSP costing, the allocation o f HIPC resources to priority sectors was less than originally programmed. Overall education spending in 2002 was close to original PRSP projections, while health spending showed large shortfalls. Budgetary spending and donor financing. The PRSP process has not generated significant additional resources in real terms. However, it has been accompanied by a shift from project to budget support. Nonetheless, the limited predictability o f budget support within fiscal years has caused difficulties for expenditure and cash management. Donor aid has been increasingly aligned with PRSP priorities as the increases in aid have benefited priority sectors. ... VI11 The govemment also increasedpriority spendingeven ifthose years inwhich foreign aid was not rising. IntegratingBudgetProcessesand the PRSPReview Programbudgetreform: measures, resultsand prospects.The govemment undertook a major reform in 1998 by adopting the program budget approach. Objectives, cost, duration and implementation strategy were set out in a project paper, and the reform was launched in the budget circular for the 1999 budget with emphasis on six pilot ministries. The guidelines were respected fairly well, and the exercise strengthened further with the integration o f the global MTEFinto budgetplanning in2000. However, after these initial accomplishments, the scope of the program budgeting remained limited, and little effort is afforded to program budgeting in view ofthe incentives producedbythe predominance oftraditionalbudgetingsystem. Existing program budgets in the health, basic education (MEBA), and secondary and higher education (MESSRS) ministries are evaluated with some standardized criteria, notably the existence o f a strategy with quantified objectives, the coherence o fprograms with objectives, and the existence o f a results measurement and information system. Inthe health ministry, program budgets build on an existing sector program, but are weak as regards financial plans, annual targets for physical output, and sector indicators. As regards MEBA, the onset of the ten-year basic education development program has improved the operational use of program budgets, keyed to the various program targets. The program budget is satisfactorily costed, and equipped with quantitative indicators. However, exposition o f the program budget using standard tables could be strengthened, and a fbrther regional break-down o f resources and objectives would be desirable. The secondary and higher education ministry(MESSRS) has one o f the best developed program budgets, including costing, indicators, and targets. Objectives are closely linked to past outcome and available resources. However, MESSRS staff notes the frequent postponements o f actions to future budget years as a result of fimding shortfalls or budget cuts. Several reasons can be identified for the mixed results inimplementing program budgets. These include the absence of a sustained effort and hasty introduction, lack o f computer applications and programmatic spending data, weak institutions, lack of resource predictability ad weak revenue forecasts, failure to systematically set objectives and adjust indicators, and lack of implication o f deconcentrated units. Links between program budget and the PRSP. The budget and PRSP process currently unfold completely separately. The budget is prepared over a period o f roughly seven months, starting with the distribution o f the budget circular in May. The circular i s based on the global MTEF and its allocations by ministry and title. Most ministries follow an imperfect and approximate mechanism to break down their MTEF allocations, and the objectives contained in program budgets reflect often ambitions unrelated to allocations. Budget arbitration focuses on procedural issues, and the PRSP and program budgets only play a marginal role. The National Assembly also does not consider program budgets. The PRSP review process is not connected with the budget process and does not follow a regular timetable. Progress reports, which are expected to present poverty analysis, macro i x results, reforms, budget execution, physical output, and sector results in relation to objectives, are generally not available before the development o f draft budgets. Subsequent budgets thus do not reflect the lessons leamed from implementation duringthe past year. A systematic evaluation of the implementation o f program budgets could serve as a base for preparing the progress reports, help to better link PRSPs and program budgets, and support the regional dimension of the PRSP process. IntegratingPRSPprogress reports and programbudgets. Better integratingthe PRSP review andthe budget process would require an alignment between budget andPRSP timetables. The progress report would need to be prepared earlier in the budget year to make key lessons available for the preparation o f the MTEF. To better integrate PRSP reviews and budget process and avoid duplicity, the progress report should rely on program budget execution reports. The latter would provide a financial execution statement, assessment o fphysical output, andprogress made in relation to sector indicators. The summary o f the program budget evaluation could therefore at the same time serve as annual progress report for the PRSP. As a first step, program budget execution should be evaluated on the regional level and be consolidated by line ministries to assess the global program budget needs for the next three years. Ina second step, the resource needs should serve as a basis for the exchanges on the global MTEF, which in tum establishes the ceilings for the budget circular taking into account macroeconomic constraints. Based on the ceilings, line ministries would prepare new budgets, including program budgets to break down resources by programs and regions, establishing physical output and objectives interms o f indicators. Ina fourth step national arbitration would evaluate the budget proposals with a view to PRSP objectives. Finally, after adoption o f the budgetby the National Assembly, line ministries would adjust their programbudgets inline with final budget allocations. The described process would fit into the recently adopted region-based PRSP monitoring andevaluation architecture, fill it with content, andclosely align it with budget procedures, thus ensuring tight links between PRSP targets and budgetprocesses. For the reform. to be successful, a long-term reform process would need to be anticipated. A strategic approach would need to be developed, and sufficient resources put inplace to follow through with the reform. To the extent possible, the reform process should place key advocates in line ministries, and strengthen the involvement o f decentralized units. To encourage reforms, it would also be important to avoid sharp and erratic cuts in budget allocations after adoption o f the budget. The further improvement o f the global MTEF, greater flexibility in allocating resources by sector, and further steps to improve sectoral MTEF approaches would also be useful. HealthSector Policies under the PRSP: Progress and Challenges Health sector in Burkina Faso: outcomes, resources, and utilization. Burkina Faso's health system at independence was characterized by limited hospital services in urban areas and almost absence of modem health facilities in rural areas. Health services were expanded in the 1970s and 1980s but suffered from chronic shortages. Inthe 199Os, the government introduced cost recovery for health services, and drug availability improved sharply, but costs remained X prohibitive for many. To improve accessibility and quality of care, the government is currently implementing a 10-year health sector development plan(PNDS). Communicable diseases, most importantly malaria, meningitis, and HN-AIDS, are the main causes of mortality and morbidity. Most key health indicators are below the average for other sub-Saharan Afncan countries, and worsened inthe 1990s, despite rising health spending, before recovering in the past 5 years. Moreover, health and nutritionindicators are significantly worse in rural areas than in urban areas, and large differences exist between the poor, and the wealthiest 20 percent of the population. The main reasons for the recent moderate decline in infant mortality are likely the improvedprenatal care and assisted births, the decline in fertility, and improved vaccination coverage. By contrast, the continuously high juvenile mortality can likely be attributed to worsening nutritiontrends and lack ofprogress in fighting malaria. Malnutrition and high fertility contribute both directly and indirectly to the highrates of child and maternal mortality. Moreover, community andhousehold factors play an importantrole inthe highrates of morbidity andmortality, particularly amongthe poor. Lack of access to safe water, low female literacy, lack of food security, inappropriate feeding and repeated illnesses all contribute to high mortality rates. Access, utilization, and quality o f health care also contribute and explain in part the regional and socio-economic differences inhealth outcomes, as prenatal care can significantly reduce the risk of maternal and child mortality, as can early treatment o f malaria and respiratory diseases. Traditional medicine continues to play an important role since the recent expansion of private health facilities took largely place inurban areas. The major source of modem health care for the poor and rural population is the public health system, organized in three levels; representing primary care, district hospitals, and regional and national hospitals. Although the number o f facilities has increased inthe past 10 years, access remains a constraint in many rural areas, with an average distance to the next health center o f 9 km, and many facilities are not accessible from remote areas during the rainy season. The number of health personnel has increased, but remains below the sub-Saharan African average, and personnel is heavily concentrated in urban and semi-urban areas. Medicine supply has improved through the establishment o f the purchasing center for essential generic drugs (CAMEG), but stockouts still occur. Maintenance and replacement o fmedical equipment is often inadequate because of lack of an efficient monitoring system and limited resources. Patterns of health services use vary by residence and income. Rural households have to rely mostly on nearest health centers, with urban residents having access to public hospitals and private clinics. Richer households are significantly more likely to visit modern health facilities. Overall vaccination rates have improved in the past five years, but large differences remain between richer and poorer households. Utilization of health services has improved only slowly, and the referral hospitals seem to be largely accessed by the local population, with limited use by population inremote areas. Vaccination coverage increased as a result of a concerted effort of government and donors, increased outreach by health districts, free provision of syringes and vaccination cards since 2002, and free vaccination since 2003. Reductions in prices o f prenatal consultation and xi essential drugs also contributed to improvements in consultations and prenatal care coverage. Statistical analysis for the 1998 Demographic and Health Survey confirms that households are more price sensitive ifthey have lower incomes, explaining part o f the different user patterns between richer and poorer households. Moreover, poverty status and educational attainment were important determinants for health center usage but less so for taking part in community outreach activities. Ethnicity and religion are closely correlated with use o f reproductive health services and vaccination, with significantly lower use among animists. Many women quote the perceived uselessness and high costs for not using prenatal care. Survey respondents quoted better infrastructure and equipment, reduction in costs, and better quality o f care as most important priorities for the health sector. Household health expenditure and financial access to health care. Average household expenditure for health increased moderately innominal terms between 1998 and 2003, reflecting increased spending by higher-income households while per-capita spending o f lower-income households declined, most significantly for the poorest quintile. The latter reflects the decline in the cost o f primary and preventive health services and explains inpart the increased utilization rates by the poorest. Most o f the health spending is on medication, especially among the poor rural population. The government has implemented several measures aiming at free prenatal care, standardization of consultation fees, and subsidization o f three drugs commonly used by children. The price o f drugs also declined through reductions on margins for the purchasing center for essential generic drugs. District surveys conducted annually confirm that these policies resulted inlower costs for vaccination and prenatal care, and reductions inthe cost o f emergency and surgical services. However, average costs of consultations increased slightly and prices continue to vary considerably among hospitals and healthcenters. Coverage with community insurance schemes remains limited, and the low monthly contributions are not sufficient to cover high-cost or urgent care. The use o f budgets to subsidize emergency care for hospitals is not closely monitored, and only very few patients seeking urgent surgical care were treated for free or being subsidized. Many patients report being asked to make unofficial side payments for care, with emergency services being considered the most corrupt. The poorest segments o f the population therefore avoid using hospital services altogether or present themselves too late. Health financing and expenditure. Despite low income levels, private households provide the largest source o f health financing in Burkina, followed by the government and donors. Most o f private health care spending is undertaken by the wealthiest fifth o f the population. Total public sector spending increased between 1998 and 2003 innominal terms and as a percentage of the budget. However, total resources for the sector stagnated owing to the closing o f World Bank-financed projects that were not replaced by equally high budget allocations, and the execution o f HIPC-related spending was slow. Information on overall extemal support to the health sector is generally poor as many donors bypass the government's budget. xii Allocations o f resources between the central and decentralized levels as well as operational and administrative functions have not changedsubstantially since 1998. Spending on regional hospitals rose faster than for national hospitals, but centrally allocated salaries rose even faster. Despite an overall positive trend inbudget allocations for health, the transition to budget support and the onset o f HIPC-related spending have resulted in a re-centralization of health expenditures. The government has recently begun to counter this trend through increasing delegatedappropriations. The allocation o f the health budget by broad economic categories has changed little over the years, with roughfy 30 percent allocated to salaries, 20 percent to goods and services, and 35 percent to transfers. However, much o f the transfer bill benefiting autonomous hospitals also supports salary payments, and taken thus about 40 percent of public health spending is for salaries. As regards priority programs, most o f them rely on donor funding with only central office functions being covered by the national budget. This is true to varying degrees for vaccination, malaria, reproductive health, and HIVIAIDS programs. The insufficient fimding o f nutritionprograms remains a preoccupation. Resource allocations vary considerably among districts and are inequitable, as regards both centrally managed personnel and infrastructure, as well as allocation o f delegated appropriations. Current information systems do not allow tracking overall resources available at the district level. Estimates identify cost recovery as largest own revenue source for districts, followed by donor fund and delegated appropriations. It appears that of these resources, an increasing share has been devoted to front-line services, and more attention has been given to maintenancespending. However, districts still lack integrated financing plans, and health centers use little o f the resourcesthey generate from drug sales for community health initiatives, despite often large cash reserves, raising questionsabout effective management oftheseresources. Within the public sector, resourcesare devoted largelyto provide services of government hospitals and health districts. Districts are using a performance-basedcontracting system with a limited set o f indicators, but there is no criterion for government budget allocations among districts. While hospitals enjoy some financial and administrative autonomy, they do not operate under a performance system, and personnel at referral hospitals is mostly paid by the university system with little incentive to supply services in hospitals. In general, government health workers are paidbasedon seniority with little rewards or sanctionsbasedon performance. Government budget allocation and execution process. The priority role o f the health sector inthe PRSPprocess has not yet been fully reflected inthe budgetprocess. The Ministryof Health has little influence on the MTEF preparation, and does not update programs and their costs to strengthen the ministry's role in the budget allocation process. The program budgeting process has had limited influence on actual budget allocations because often indicative allocations arrive late, costing is not rigorous, and there is no system to track program budget implementation. To improve budget procedures, the HealthMinistryissues a draft circular within the ministryto distribute MTEF resources, and the preparation o f a sectoral MTEF i s under way to improve costing o f programs. The sectoral MTEF promises to become the basis o f a results- basedfinancing system for the sector. ... X l l l Interms of personnel planningand execution, the Ministryof Healthhas limitedcontrol over personnel matters handled by the Ministry o f Civil Service. There are currently no consolidated statements of healthworker wages and no monitoring o f the location o f health staff. Recruitments buildon students inthe national health schools, who passed an entrance exam, but coordination o f recruitment policies between national schools and the health ministry are weak, and schools are overcrowded. Health personnel i s allocated among regions, but health workers are reluctant to accept positions inremote areas or leave their posts after their appointment, often without loss o f salary since the latter is not assigned to a post but to a person. As regards non-wage expenditure, the notification o f credits and execution o f the budget remain hampered, especially at the decentralized level, by late notification and cumbersome and centralized procedures. Regarding the transfers (operating subsidies) to autonomous agencies, which play a significant role for the health ministry, they often face problems with the release o f their quarterly tranches from the budget. Moreover, budget information and accounting systems for these agencies are insufficiently developed, and the Ministry o f Health does not prepare consolidated statements. Regarding HIPC expenditure, although the spending out of a special account usually allows some flexibility incarrying over allocations, this flexibility was curtailed as a result o f unpredictable debt relief. Moreover, HIPC resources followed special procedures, and processes were overly centralized. District staff acknowledges increased resources since the onset o f HIPC spending, but note the problems inmatchingHIPC spending with actual needs. Participation o f deconcentrated structures in the budget process i s limited. Although districts and regions prepare an annual plan, these plans are usually prepared after the budget process has been completed, and deconcentrated units have no influence on the arbitration process on the central level. Regions and districts receive delegated appropriations for materials and services, but the process has currently a number o f drawbacks associated with the late notification of budget allocations, the continuation o f centralized procurement for certain goods and services, and the difficulties inexecuting regular government procurement procedures inthe regions and districts. Cash advance accounts (rbgies d'avance) have been established to allow greater flexibility and faster response at the district level, but the pretty cash accounts encountered a number of difficulties, including late nomination o f account managers, misunderstanding o f procedures, and inflexibility o f rules goveming the accounts. For expenditure o f centralized budgets, deconcentrated units are not involved intheir execution and control, even ifthe regions and districts are their beneficiaries (personnel, infrastructure). On the district level, health committees manage the financed o f health centers. The latter prepare micro plans that set forth annual activities, and these are financed through own revenues o f the health management committees and in-kindtransfers from the health districts. A first audit o f health center management took place in2003. The EducationSector in BurkinaFaso: Performanceand Progress Education trends and outcomes. The number o f students in the BurkinabC education system has increased steadily during 1999-2002. The overall gross enrollment rate increased somewhat, with a larger increase at the primary level. However, despite some progress, more than half o f school-age children are out o f school, and Burkina Faso's enrollment rate remains among the lowest inthe region. Low enrollment is partially owing to the inadequacy o f supply at xiv all educational levels, with expansion of supply hampered by high unit costs of infrastructure. Despite strict access control (only one in four students passed the baccalaureat), even in higher education supply does not meet demand. The private sector is meeting some o fthe rising demand for education services, especially as regards secondary education. Its most importantcontribution is inthe area of vocational training. Large discrepancies remain in enrollment levels between urban and rural areas, by gender and by income. Urbanrural disparities are especially large on the primary level, with urban rates being three times higher than rural rates. Girls have systematically rower enrollment rates than boys, reaching on average 75 percent o f boys' enrollment, although this gap is much smaller in urban areas. There is a strong positive correlation between primary school enrollment and standard-of-living, with the poorest 20 percent o f households being half as likely to enroll children thanthe richest quintile, and widening disparities at the secondary level. Repetitionand drop-out rates remained highat the primary level, between 12percent for first grade and 36 percent for sixth grade, and worsened with the expansion of the education system in 1998-2002 before improving in 2003. Performance o n the secondary level was even poorer, with 55 to 60 percent o f students dropping out inthe last grade. The overall low quality can be partially explained byhighpupil teacher ratios, which increased sharply inrecent years on the secondary level. Teacher shortages on the secondary level are most pronounced for scientific disciplines. The high repetition rates and low completion rates imply high unit costs of education. Large efficiency gains could be made (a reduction by a factor o f 3) by reducing repetition at the primary level. The highest loss is at the university level with actual costs almost 8 times of average costs per student and cycle, and additional loss through grants and non-reimbursed student loans, The inefficiency weighs heavily on the education system and hampers a faster expansion o f the system. Moreover, the focus on formal theoretical training and relative neglect of industrial andago-pastoral studies or technical training does not allow matchingstudent skills with the country's Iabor market needs. Education financing and expenditure management. Education is financed by govemment and households, NGOs, and the private sector. Government resources from tax revenues and budget support are the main source of public funding, followed by extemally financed projects. The education sector also absorbed a large share of the HPC resources. Overall expenditure inthe education sector increased steadily from about 2.5 percent of GDP in 1998 to about 4 percent in 2002, and education spending as share o f the overall govemment budget increased, attributable to the start-up of the PDDEB and increasing use of HIPC resources. The distribution of education expenditure reflects the large weight of salaries and other recurrent spending, representing between 60 and 70 percent of total spending. With the onset of the investment program under the PDDEB, there has been a marked shift to higher investment. shares in education spending, which have now reached 40 percent of total spending. Among recurrent spending, the weight of transfers to autonomous bodies has also been growing (mostly subsidizing salaries and study grants). Spending on materials and services has been low, at 5 xv percent o f total spending, and parents of students hadto make up for these low levels o f spending with their own contributions to operating and maintenance costs. However, the introduction of delegated appropriations, recent waiver o f school fees, and the subsidization of schooling in the 20 most disadvantaged provinces has resulted in an increase in operating and maintenance spending in total spending to over 8 percent. It is also noteworthy that operational expenditures have increased significantly faster than administrativeexpenditure inthe past years. Education expenditures are concentrated on primary education, which absorbs about 65 percent o f education spending, followed by secondary (20 percent) and higher education (15 percent). These shares fluctuate between years as a function o f foreign-financed project spending. Overall unit costs o f education have risen over time in nominal terms; however they declined as a share of GDP. The unit costs also reveal the striking costs differences by level o f education, with educating two students in higher education costing about as much as teaching a whole class o f primary students, underscoring the need to reduce costs o f higher education in order to free more resources for basic education. Program budget allocation and execution. The Ministry o f Basic Education allocates a significant share o f its programmatic resources (excluding salaries) to increased supply of education and capacity building. Insecondary education, the focus is on vocational and technical training, whereas in higher education programs target improved quality. For the overall education sector, programs budget allocations targeting increases in supply absorb about 40 percent, whereas quality improvements receive about 33 percent, and the remainder going toward better administrative capacity. The absorptive capacity o f the education sector has been somewhat higher than for the rest of the budget, averaging 87-90 percent in 1998-2000 and about 95 percent in 2001-02. Absorption was highest at the basic education ministry, reflecting the important share o f salaries with commonly high execution rates in this ministry's budget. However, absorption rates of foreign-financed investment and HIPC resources remained low in the Ministry o f Basic Education at less than 50-60 percent in2001-02. The strong centralization, credit regulation, and less experience with foreign-financed spending are reasons for even lower absorption rates inthe Ministry for Secondary and Higher Education. The delegated appropriations at the Ministry o f Basic Education are a positive development to enhance responsibility o f local actors and improve transparency. Until 2004, delegated credits were managed only by regional departments (DREBA) and were earmarked for consumables, repairs, and a petty cash fund that can be replenished once a year. Execution rates for delegated credits were high, approaching 99 percent. The central budget continued to pay for fuel and equipment, construction, and teaching personnel. HIPC funds were also executed centrally, including the provision o f free schooling material for the 20 provinces with the lowest enrollment. Delegate appropriations do not exist inthe Ministry o f Secondary and Higher Education. Appropriations of the regional directorates are managed centrally, with many difficulties for local service delivery. The government has almost completely withdrawn from the construction of middle and high schools, which are built by local communities. The government generally xvi contributes materials and books. Food for cafeterias in middle and high schools is ordered centrally andthen distributed. For basic education, parent-teacher organizations play a significant role. Their participation is often hampered by illiteracy of members, and by their lack of experience in managing funds. At the secondary and higher education level, institutions dispose in addition to parent contributions of their own funds from school fees, 25 percent of which revert to the ministry. At the primary school level, and more importantlyfor secondary and higher education, schools are asked to transfer significant resources to the central administration (circumscription, provincial directorate, or regional directorate) to support their operating expenses since the centrally allocated government budget is not sufficient. The level of involvement of decentralized levels in budget preparation is low. However, with the onset of the basic education development program, the involvement of deconcentrated structures of the Ministry of Basic Education has increased. Provincial action plans and the development of a sectoral MTEF are steps in this direction. Further formal inclusion of these deconcentrated units inthe budget process would be desirable. At the Ministry of Secondary and Higher Education, most units report little involvement in the budget process, and some central structures are consulted only in the process of public investments. Budget arbitration takes place on the central level with very limited involvement of regional directorates. Middle and high schools, which rely much on their own resources, are not included inthe budgeting process. Key Recommendations Measure Proposed Timetable I Budgetplanning, execution, and monitoring I Systematically review the level o f recurrent and investment spending in Preparation o f the 2006-08 MTEF in priority areas basedon budget execution data for 1998-2003 February-March 2005 Implement recommendations o f the AFRITAC and EC technical assistance to December 2005 improve programming and monitoring o f foreign-financed investment spending Include identifier for priority spending inthe functional budget classification 2005 budget law 1Report systematically on execution of priority spending in the PRSP progress 2004 PRSP progress report I I reports 1 Improve the functioning of interministerial working groups to ensure Monitoring o fthe 2004-06PRSP consistency between PRSP and sectoral programs and open them for civil society, private sector, and donor participation Further deconcentratepayment orders to line ministries and regions 2005 budget law Improve efficiency o f ex ante control through adoption o f a price referential, a December 2004 decree on "morality" of expenditure, and a revision o f the nomenclature of required documentation xvii I Measure I Proposed Timetable Set up a public informationcenter inthe MFB and MEDEV carrying all main June 2006 documents related to the PRSP, sector policies, budget laws, budget execution reports, and donor-financedprograms andprojects Publish within year reports on budget execution 2006budget law Provide program budgets and the global MTEF to the National Assembly Debate of National Assembly on 2006 together with the budget draft law budget I Institutionalize exchanges with key donors on three-year projections o f Preparation o f the 2006-08 MTEF in program andproject aid February-March 2005 Seek to better synchronize budget aid with the budget cycle and reduce Discussions o f the PRSP timetable, uncertainty about amounts by accelerating provision o f PRSP indicators and budget support memorandum of pursuing new agreements linkingdisbursements to knownpast performance understanding and new budget support agreementsin2004 Within a sector, prepare programbudget on regional basis to be consolidated January 2006 for the preparation of into the national program budget 2007 budget Instruct line ministries to prepare program budget execution reports on the January 2005 regional and national level Prepare global MTEF with input from line ministries based on the results o f February-March 2005 programbudget implementation review Beginpreparingregional and national PRSPprogress report based on regional March 2005 and national program budget implementation report, conduct regional consultations Distribute budget circular and prepare budget proposals and new program May-June 2005 Improve allocations to front-line services and develop criteria to improve M ~ ~2005- J ~ ~ equity o fbudget allocations among regions and districts Strengthen monitoring by IGSS o f the application o fprices for preventive and ongoing primary care ... XVlll Measure 1 ProposedTimetable I Further reduce prices for essential care, continue efforts to reduce margins for Essential drugs, and develop mechanism for identifying and subsidizing 2006 budget law indigents Streardine procedures for the release o f delegated appropriations and closely 2005 budget monitor functioning of rkgies d'uvunce Strengthen accountability for funds at the district and community level by December2005 setting up a system to record receipts, expenses, and account balances, with regular audits by the Ministry Revise texts governing health committees to strengthen their role in public ~~~~~b~~2004 health Develop performance contracts for hospitals, including performance December 2004 indicators Define and implement measuresto increasethe demand for health services Revise the system o f user fees and clarify financial and organization December2004 framework for subsidizing the poor Establish system of annual reviews o f hospital performance, initiate global 2004-05 audits, strengthen oversight o f the Ministry o fHealth Improve distribution and motivation o f health personnel through better December2005 incentives in rural areas, decentralization o f budget` posts, and improved personnel management systems Increase involvement of regions and provinces in budget planning in the context o f aligning budget and PRSP processes and increase transparency o f November-December 2004 I budget discussions Simplify procurement processes and avoid overly stringent regulation o f 2005 budget budget lines inthe Ministryo f Secondary andHigher Education Extend the delegated credit system to the Ministry o f Secondary and Higher 2006 budget law Education Strengthen the accountability o f schools and autonomous institutions through December 2005 a system o f information recording and audits Engage in capacity building for regional and provincial levels to improve Ongoing starting in2005 budget planning, execution and monitoring Improve recruitment and deployment o f teachers and inspectors through Ongoing regional and provincial action plans, mapping o f vacancies to regions and provinces, and development o f a regional recruitment policy for the secondary ~1 level Improve teacher training inthe sciences On oin startin in2005 Establish procedures to recruit teachers at the local level Adopt incentives measures that could keep teachers inthe local areas Improve costs o f higher education through definition o f a viable and long- term strategy including cost recovery, notably o fFONER. Ongoing starting in2005 ~ xix I.INTRODUCTION 1. This Public Expenditure Review (PER) supports the government's public expenditure management reform agenda in the context o f the implementation o f the poverty reduction strategy paper (PRSP). Since 2000, the government at its own initiative has undertaken several sectoral PERs for health, basic education, infrastructure, rural development, public investments, higher education, andjustice. Inaddition, the Bankjointly with the IMF and bilateral donors has provided advice to the authorities through the Country Procurement Assessment Report (CPAR), Country Financial Accountability Assessment (CFAA), the report on the observation o f standards and codes (ROSC) and the Accountability Assessment and Action Plan (AAF') prepared in the context o f the Highly Indebted Poor Country (HPC) Initiative. The government followed up on this analytical work in 2001-02 with the preparation o f an action plan for the improvement o f budget management, known under its French acronym PRGB. Implementation o f the government's PER program and public expenditure management reform was supported by a series o f three poverty reduction support credits (PRSCs) in 2001-03 and by a variety of bilateral aid. 2. During three years o f PRSP implementation in Burkina Faso, the government has' undertaken reforms of budget procedures to strengthen the programmatic links between the budget and poverty reduction objectives. Recently, the government has prepared a revised PRSP for 2004-06, which i s accompanied by a Priority Action Plan (PAP). This PER responds to the need expressed by the government to analyze the developments of budget allocations and spending since 1998 and to take stock o f the achievements and shortcomings in linking the budget and budgetprocesses to the PRSP. The PER is organized infour main chapters, covering respectively: (a) an analysis o f budget allocations and budget execution in 1998-2002 and their linkwith PRSPpriorities; (b) aninstitutionalexaminationo fprogrambudgets and the MTEFand development o f possibilities for a closer integration o f the budget with the PRSP process and its annual reviews; (c) a review o fthe implementation of sectoral budgets inhealth; and (d) a review of the implementation of sectoral budgets ineducation. A. COUNTRY BACKGROUND 3. Burkina Faso is a poor landlocked country o f about 12 million inhabitants with a narrow natural resource base. In 2003 its GDP reached about $250 per capita and Burkina Faso was ranked 173rdout of 175 countries in the United Nations Development Program (UNDP) Human Development Index. About 20 percent o f the population live in urban areas (mainly the capital city Ouagadougou and Bobo-Dioulasso, the center o f the cotton-growing area) and most Burkinabt remain heavily dependent on agricultural activities. Macroeconomic performance and progress inreducing widespread poverty therefore remain vulnerable to exogenous shocks, such as climatic conditions and changes inworld market prices for cotton, Burkina Faso's main export good. 4. Burkina Faso was among the first countries to present a full PRSP. The PRSP is centered on four pillars: (a) accelerating equitable growth; (b) improving social services; (c) creating income for the poor; and (d) improving governance. Several long-term sectoral programs, such as the 10-year basic education development plan (PDDEB) and the national health development plan (PNDS) underpin the PRSP objectives. The government presented the progress made in implementing the strategy in three. annual progress reports (PR) in 2001, 2002, and 2003. However, the reports limited themselves largely to selected budget execution figures and did not provide a comprehensive picture on the link between the budget andPRSPobjectives. 5. Overall, Burkina Faso has made considerable progress in macroeconomic stabilization under three successive programs supported by the IMF's Poverty Reduction and Growth Facility. The growth rate o f gross domestic product (GDP) has averaged 5.7 percent between 1994 and 2003, against a population growth rate o f about 2.5 percent. Fiscal consolidation after the CFA devaluation in early 1994 resulted in an increasingly stable macroeconomic environment in the context o f the pegof the CFA to the French franc andthe euro. 6. According to preliminary analysis o f the 2003 household survey under a poverty assessment, poverty rates declined by 8-9 percentage points between 1998 and 2003 (from about 55 percent in 1998). However, key social indicators remain substantially below the sub-Saharan average. The figures demonstrate the challenge for Burkina Faso in making significant strides towards meeting the Millennium Development Goals (MDGs). These results also indicate the need for determined efforts to enhance the effectiveness of budgetary expenditure in achieving government growth andpoverty reduction objectives under the PRSP. 7. The country has remained highly vulnerable to external shocks. These shocks include a) rainfall conditions in the Sahel zone; b) international cotton price developments; c) political instability in neighboring countries; and d) lack o f predictability o f donor resources. To the extent that they influence budget execution in mid-course, these factors are susceptible to interrupting spending programs and can result in cut-backs in expenditure allocations, notably for spending on materials and services and investments financed with domestic resources. A limited credibility o f the budget process also undermined the incentive to seriously engage in program budgeting since the potential beneficiaries did perceive little connection between their budget requests, budget allocations, and resources eventually made available. 8. Budget implementation was affected in the past by the government's overly optimistic revenue forecasts. These forecasts were brought back to achievable levels in the context of IMF program discussion, which implied that the original budget could not be implemented as planned. The weak internal revenue base also resulted in a strong dependency on the level and timing o f donor resources. In 2002, budgetary expenditure, including foreign-financed investment spending, totaled CFAF 486 billion (about 22 percent o f GDP), o f which only CFAF 260 billion or 53 percent could be financed out of own revenues. Donors contributed grants (CFAF 119 billion), loans (CFAF 101billion), and HIPC debt relief (CFAF 21 billion) to finance the budget and reimburse some domestic debt. - 2 - B. PROGRESSBUDGET IN MANAGEMENT 9. The government has begun since the late 1990s to address some o f the weaknesses in budget management. In the past three years, revenue projections have become increasingly realistic and, with the full integration of the medium-term expenditure fi-amework (MTEF) into the budget cycle in2002, the budgethas become less prone to over programming o f expenditure. Less progress has been made in increasing the volume o f domestic resources as a share of GDP. Enhancing the efficiency of revenue collection remains the focus of the 2003-06 IMF-supported program, as well as an important component o f the govemment's budget management reform plan PRGB. 10. Three years o f PRSP implementation process have also seen shifts in the allocation of government spending toward priority sectors. The increase in spending allocations was accompanied by a variety o f analytical work by the World Bank, the IMF and other donors to strengthen the efficiency of public expenditure management systems (CFPLA, CPAR, ROSC, HIPC-AN). Among others, reforms undertaken by the government include full operational o f the computerized expenditure circuit and its stepwise extension to regions; improvements inthe timeliness o f budget execution laws and treasury balances; the revision of the budget classification; andwide-reaching changes to publicprocurementregulations. 11. Donor support has become more predictable since 2000 in the context o f increasing resource flows through programmatic budget support for PRSP implementation, reducing the frequently onerous procedures associated with foreign-financed projects. Direct budgetary aid from the EuropeanUnion (EU), the World Bank, the African Development Bank (AfDB), and several bilateral partners increased from 9 percent of total spending in 1999 to 15 percent of expenditure in 2002. However, little progress has been made in economic diversification and reducing vulnerabilities to climatic and political disturbances. In this respect, the 2000 drought and the 2002-03 crisis in neighboring C6te d'Ivoire demonstrate the negative impact of these events both on domestic revenue collection and expenditure with concomitant effects on the execution o fthe government's priority programs. 12. Since 2000, the government has sought to improve the budget execution in line with PRSP objectives. Chapter 2 analyzes how the government responded with fiscal management to the adoption o f the PRSP in 2000. It discusses limitations o f current budgetary practice, investigates whether the PRSP has fostered transparency in budgetary decision making, reviews overall budget allocations, and discusses whether the budget execution pattern is consistent with protecting priority programs. The chapter assembles for the first time comprehensive and consistent budget execution data from the budget execution reports (Eois de &gIement), and reviews the alignment o f donor resources with the budgetary processes. 13. The government also introduced programmatic and medium-term budgeting. A global MTEF was first developed in 2000, and the 2003-05 MTEF prepared in 2002 was fully integrated into the budget cycle. Program budgets were prepared for most ministries since 1998- 99. However, as noted during the third annual PRSP review inJuly 2003, the programmatic links between the budget and the PRSP have remained weak. PRSP progress reports have focused little on the connection between program objectives and budget execution to explain actual - 3 - outcomes o f PRSP implementation. Chapter 3 reviews the institutional settings for the MTEF and program budgets and develops recommendationshow the budget cycle and PRSP cycles can bealigned more closely inthe future. 14. Education and health play major roles in the national budget and the PRSP. The finalization o f the PRSP ledto an increasedfocus ofthe government anddonors on the budgetas primary expression of policies, with the attempt to change its composition over time and raise spendingin sectors linked to PRSP priorities. At the same time, resources were freed up under the HIPC Initiative and the government agreed that these resources would add to poverty- reducing expenditure. PERs were prepared in 2000 by consultants as part o f the government's PER program. Since then, significant reforms have been undertaken with the adoption o f the PDDEB and the first two years of its implementation; the work undertaken for the "Education For All Initiative;" as well as the adoption o fthe PNDS inthe health sector. These new programs reflect in part recommendations of past sectoral PERs, for example, the need to reduce the education unit personnel cost in order to fkee up additional resources for increasing schooling rates. However, challenges remain in these sectors to increase efficiency o f service delivery, as evidenced by slow progress in improving social indicators. Chapter 4 and 5 update the information available on these key sectors and recommend further improvements in the efficiency of expenditure allocations. 11. PRSP PRIORITIESAND THEBUDGET 1998-2002' 15, This chapter assesses to which extent public finance management reflects the principles and content o f the PRSP, and whether the PSRP process itself has fostered more accountable, efficient and pro-poor public finance management. To this end, the chapter reviews (a) the production of budget data and the results-orientation of the budget process; (b) the participatory and transparent nature o f the budgeting process; (c) resource allocation and extent to which budget allocations reflect PRSP priorities; and (d) partnership and donor alignment with the PRSP. 16. Strategic orientation before the adoption of the PRSP. Burkina Faso has a strong participatory tradition that is reflected, to this day, in the existence of a large number of NGOs and community,based organizations.* A strong focus on poverty reduction emerged for the first time under a "revolutionary" govemment in 1983-87. It emphasized a development process buildingmostly on internal resources, the redistribution of public resources from urban to rural areas through the expansion o f govemment-financed health and education services, support of the agricultural sector, and reduction inthe wages of civil servants located inthe capital. Popular participation inthe form of social mobilization, including as input into the elaboration ofnational development plans, was encouraged and a comprehensive development vision was elaborated and eventually expanded into popular development plans for the regional, provincial, and 'Core elements ofthis chapter were prepared by Rosa Alonso iTerme as part of a multi-counby case studyon fiscal aspects ofPRSP implementation, sponsoredby the World Bank Institute (WBI) and the PREManchor. In 2003, there were over 200 NGOs and around 14,000 community-based organizations in Burkina Faso. Evaluation de Cadre de DeveloppementIntegre. Etude de Cas :le Burkina Faso. - 4 - national level. However, the strict government control o f the development process and the forced income redistribution also fostered increasing dissatisfaction insome sectors of the population. 17. Inthe late 1980s, under President Blaise Compaork, the government sought to rekindle its relationship with donors and to retum to a more market-oriented approach to development, drawing more on external support while maintaining the participatory tradition. Two national meetings were held in 1990 and 1994 to design a new economic development strategy. These meetings resulted in the elaboration o f the Letter o f Intent on Sustainable Human Development Policies (LSHDP), which lays out a strategy for human resource development and poverty alleviation for 1995-2005. The LSHDP was discussed with donors at the 1995 round table meeting in Geneva and represented the first comprehensive effort by the government to formulate its own vision o f development. The letter sets out a vision for the country's development based on the concept o fhumansecurity, which entails economic security (access to training and employment), health security, food security, environmental security, and individual and political security. The letter also included key targets the government intended to pursue under the strategy, such as (a) increasing annual per capita GDP by at least three percentage points per year; (b) reducing the incidence o f poverty from 45 percent to 30 percent by the year 2015; (c) doubling the literacy rate; and (d) increasing life expectancy by at least ten years during the period. The LSHDP strongly influenced the PRSP developed in1999-2000. 18. Other government policies duringthe 1990s paved the way for PRSP preparation. These policies include: (a) successful macroeconomic stabilization under IMF-supported programs since 1991;(b) increased focus on poverty analysis, supported by two household surveys (1994 and 1998); (c) development of sector strategies inhealth, education and rural development3; and (d) increased efforts indonor coordination and alignment under the Conditionality Reform Test (1997-2000) in the context o f the Strategic Partnership with Africa (SPA). As noted above, the government also began a series o f reforms in the public finance area that were essential for increasing the efficiency and transparency o f the budget. These reforms included an integrated computerized expenditure recording system, program budgets for six pilot ministries, a global MTEF, and yearly government-led public expenditure reviews. 19. The 2000 PRSP. The PRSP emerged from a participatory and planning tradition and it built on the LSHDP and sectoral plans for rural development, health, education, and water.4 Its targets were closely linked to those included in the 1995 LSHDP. An inter-governmental body organized several consultative sessions in a first round o f consultations and two regional focus groups generated feedback on the draft document. The PRSP document itself, however, did not include a summary o f the main issues raised in the consultations and it is not clear how these The donor community had contributed to these strategies in a number o f ways, including by financing studies-the French government, GTZ and the World Bank for rural development; the WHO, the Swiss govemment and the World Bank for the health strategy; and Danida and the Netherlands for the water sector. Civil society was also involved intheir elaboration. The PRSP is considered a synthesis of the feedback provided in dialogues with a wide range o f stakeholders throughout the country. These included: forums for the design o f a new development strategy in 1990, annual seminars with NGOs during the Agricultural Workers' Days from 1993-2000, and high level meetings o f various stakeholders on education and health for the preparation o f long-term plans in 1994 and 1998, among others. Preliminary findings from household surveys and a Participatory Poverty Assessment that was finalized in 2001 were also used inPRSPPreparation. - 5 - views informed the documents' priorities. Althoughparticipation was not very deep, civil society organizations acknowledge that it was an improvement upon preceding consultative processes and that it has led to increased interaction between government and civil society in the PRSP implementation phase. 20. The poverty reduction strategy is grouped into four strategic pillars. These pillars are (a) accelerating equitable growth; (b) improving social services; (c) creating income for the poor; and (d) improving governance. The original PRSP includes three key priority sectors that are derived from the four pillars: basic education, basic health, and rural development (agriculture, rural roads; water supply; and animal resources). 21. Under the PRSP the government committed to enhance the dissemination of economic and social information as part o f a broader effort to publicize govemment efforts to combat poverty. To this effect, several PRSP dissemination workshops were conducted in 2001 by the Directorate General for the Economy andPlanning(DGEP) as well as several sectoral ministries. Furthermore, a group of national NGOs translated the PRSP into three national languages in order to increase its accessibility by the population. 22. Three years after the elaboration of its first document, Burkina Faso has just adopted a revised PRSP. The document was elaborated with a substantial degree o f participation-regional consultations were held, and civil society participation culminated in a national civil society forum. A number o f sectoral meetings at the national level involved all key ministries. The exact contribution o f civil society to the final document, however, is unclear as the draft does not discuss which feedback was reflected in the document. Overall, civil society organizations view their degree o f involvement inthe elaborationo fthe new PRSP as greater than init had beenin the original PRSP. Overall, they are also positive about the increasing level of dialogue they have attained owing to the PRSP process. The same assessment was made by line ministries. 23. As seems to be the case in revised PRSPs so far, the new PRSP for Burkina-Faso includes a largely-expanded list of priorities. The new priorities include security, environment, employment and informal sector, social exclusion, gender, small and medium-sized enterprises, small mines, rural electrification and migration. Adequately funding this expanded list of priorities will doubtless pose an enormous challenge, particularly given the current under- finding o f original priority sectors such as basic health and basic education (see section on resource allocations), serious constraints on the government's ability to mobilize additional domestic resource and limited possibility to increase borrowing without challenging debt sustainability. A. PRODUCTIONOFBUDGET DATAANDRESULTSORIENTATIONOFTHE BUDGET' 24. Comprehensiveness of budget coverage. The coverage o f the central government budget is comprehensive as budget execution in Burkina Faso is largely centralized and own ~ This section builds inpart on the HIPC-AAP evaluationundertaken in2004. Institutional issuesrelatedto program budgets and medium-termexpenditure are taken up inmore detail inthe following chapter. - 6 - revenue and spending o f local governments are small. However, the budget records only the transfers made to autonomous public institutions (such as hospitals and research facilities) but does not record their own revenue. For the health and education budgets, line items devolved to the regions and provinces can be identified easily. However, these reflect only a part o f the resources available to health centers and schools as both benefit from local semi-autonomous bodies (health management committees-COGES, and parent-teacher associations-APE) that collect and manage their own revenue and support the functioning o f local health centers and schools. For data reasons, the figures reported in this chapter therefore reflect only centrally- controlled resources financed with tax revenue and aid and do not give a full account o f local spending, some o f which may also support attainingPRSP objectives. 25. Inpart to overcome the information deficit as regards local spending inpriorityareas, the government has conducted tracking surveys and opinion polls on the quality of local services (health and education). Surveys have been carried out in2001 and 2002 and their results will be subject o f a more detailed discussion inchapters 4 and 5. They have provided important insights inthe resources made available bythe central government for materials and services onthe local level and the cost o f basic services. Institutionalizing service delivery surveys and other tools to assess the quality o f public services in the priority ministries and, in particular, in health and education, shouldbe an area o f focus for the Burkinabe government inthe future. 26. Budget classification. Despite recent progress toward their elaboration, neither detailed economic, functional, nor programmatic classifications are available for an extended period. The only budget classification that i s available for 1998-2002 and used consistently in the MTEF, budget guidelines, draft budget law, budget law and budget execution phases is the administrative/organizational economic line item classification. This classification is also the classification that is systematically employed by government, National Assembly and the donor community. The government has introduced a functional and economic classification with the 2004 budget but the non-availability o f historical functional and programmatic budget data hamper proper evaluation o f the impact o f policies on outcomes through evaluation o f sub- categories (for example, primary health care). Analyzing the link between the budget and PRSP priorities is affected by the absence o f specific markers for "poverty-reducing" spending. As a result, it is not possible to track expenditure inPRSP core priority areas with precision (see also section on resource allocation). 27. Burkina Faso follows a detailed system o f HIPC expenditure planning and accounting. HIPC resources are channeled through a special account approved by the National Assembly. In 2001 and 2002, the spending o f HIPC resources in priority sectors was decided by decree of the Minister o f Finance and they were not tracked in the computerized expenditure management system (0). This special procedure has been used to demonstrate the additionality of HIPC resources. According to this system, HIPC funds are deposited in a special Treasury account follow special budget procedures and are accounted for separately. However, this treatment of HIPC resources leads to a complex accounting system separating regular budget resources and HIPC resources. It also fosters a budgeting culture that separates the regular budget from HIPC spending and thus reduces the programmatic vision for the allocation o f resources. Since 2003, HIPC resources have been allocated by the budget law and have been followed with the CID. - 7 - 2%. Foreign-financed investment. The reporting on a significant part o f foreign financed investment spending is weak since budget coverage is incomplete, spending cannot be tracked through the computerized expenditure system, and some donors do not report their activities more than once a year.6 The latest report on foreign aid by the cooperation directorate (DGCOOP), which tracks foreign disbursements, covers the year 2001. Therefore, the results reported below rely only in part on estimations by DGCOOP, which is still finalizing the 2002 disbursement report. Similarly, large amounts o f expenditure are incurred by NGOs inthe social and other priority sectors while very little information on them i s available to the government. The availability of only partial information for government decision-makers renders the task o f overall resource allocation and planning difficult. 29. Budget procedures. An important inefficiency o f the budget process resides in the multiplicity o f a priori expenditure controls and the over-centralization in the issuance o f payment 0rde1-s.~The multiplicity o f a priori expenditure controls (at the commitment, verification and payment order stages) can lengthen the regular expenditure execution process without adding significant firewalls against improper use of funds. These difficulties are exacerbated by the understaffing o f the financial controller (DCCF) - only 12 controllers, 6 in the Ministry o f Finance and 6 in line ministries have the right to clear expenditure commitments and verifications. The centralization o f payment order issuance can also lead to backlogs and a general slowdown inthe budget execution process. Although credits dklkguks have been put in place to speed up expenditure delegated to the regions and provinces, these procedures only apply to specific expenditure categories (typically goods and services) in a small number of ministries, and the payment order issuance remains highly concentrated with only 5 out 13 regional capitals beginning to issue payment orders. The decentralization o f the payment order system to line ministries and to the regions seems essential to expedite budget execution in priority sectors and improve progress in PRSP implementation. It would also improve overall governance, as the concentration of all decision-making power on payment orders inthe Minister of Finance reduces the accountability o f line ministries. 30. As regards the spending out of the special HIPC account, many ministries lack the capacity to properly follow complex special HIPC procedures. Jointly with the belated adoption of the HIPC allocation decree in 2001 and 2002, this complexity has led to significant delays in the execution o f HIPC funds. A guide to the use o f HIPC resources and accompanying training activities for line ministries was elaborated in 2003 as a means to address capacity constraints. Despite the complications o f the HIPC system, however, staff interviewed at priority ministries expressed appreciation o f its resource targeting to priority sectors, as they believe these resources provided additional funding to implement their programs. The same conclusion was reached by the CDF reviewe8A HIPC spending audit by the govemment is currently under way to assess the Many donors do not provide any information to the government regarding their projects, inparticular those which 'areThis directly executed. problem has been stressed by every single document analyzing public expenditure management in Burkina Faso: 1997 Public Expenditure Review, "Revue des Depenses Publiques et des Budgets-Programme au Burkina Faso. Note d'Evaluation," 2001, Review of Standards and Codes, Country Financial Accountability Assessment, Programme de Reforme de la Gestion Budgetaire. "A major accomplishment of the PRSP is to have oriented HIPC resources to the social sectors up to the local level. 800 additional primary school teachers have been hired and equipment and material for health centers have been purchased." - 8 - inefficiencies encountered and to issue recommendations as regards the future use of HIPC resources. 31. Medium-term and programmatic budgeting. Medium-term expenditure frameworks (MTEF) have been elaborated since 2000. Originally, their main contribution was to provide overall expenditure ceilings for each ministry based on the basic macroeconomic framework and, hence, to discourage unrealistic funding requests, including for the program budgets. At first, the cycle for the preparation of the MTEF was not properly coordinated with the annual budget cycle, with requests for projections from line ministries for the elaboration o f the MTEF coming at a time in the year when they were not yet ready to provide them. This timing problem, which used to impair the linkages betweenthe two budgeting processes, has been solved in2002. 32. The macroeconomic projections underlying the MTEF have become more realistic over time. An important obstacle to the usefblness of the MTEF as a strategic resource allocation mechanism has been its systematic overestimation o f domestic revenue (by 20 percent in 2001 and 13 percent in 2002). This overestimation has led the IMF to make separate revenue projections for.use in its programs, hence leading to a separate planning process based on underlying parallel macroeconomic frameworks. The new MTEFs for 2004-2006 and 2005-07 make more realistic assumptions and include three case scenarios, allowing annual budget planningto be effectively based on the MTEF. According to Ministry of Finance officials, the MTEFhas also contributed to a transformationinbudgetingculture from allocating resources to achieving results. 33. Program budgets started to be elaborated in 1999 in six pilot ministries. There has been significant progress since then and, by now, all ministries prepare program budgets. A table that allows the translation of program budget allocations into administrative budgets accompanies program budgets. The quality o f these budgets, however, varies significantly and is substantially higher inthe six original pilot ministries (coinciding largely with priority sectors), Even inthese ministries, however, the program budgets that are currently elaborated are more accurately described by the term activity-based budgets. Fully developed program budgets would require an estimation ofthe true costs o fprograms, which is technically complex and information-intensive. 34. Program budgets are under-utilized, which limits their usefulness for fostering results- oriented policy-making. After their initial adoption, program budgets are not adjusted for final budget arbitration and executed program budgets do not serve as inputs into the policy monitoring process or for deciding program fbnding and resource allocations. This lack o f attention has acted as a disincentive for line ministries to devote the significant time and resources needed to elaborate program budgets. 35. Coordination of budget elaboration. The PRSP process is improving the budget dialogue within ministries. Traditionally, budget-fimding requests from line ministries to the Ministry of Finance were elaborated by financing departments (DAFs) without much input from planning departments (DEPs). This practice significantly impaired links between sector strategies and sector budgets. The focus of the PRSP process on programmatic aspects of budget - 9 - execution and medium-term planning, however, requires strong intra-ministerial coordination, This focus on results has led to a noticeable improvement inthe coordination between planning and financing departments, in particular in priority sectors, in order to ensure enhanced coherencebetween planning and financing strategies. According to the Ministry o f Finance, this improved coordination has increased the quality o f the financing requests prepared by line ministries and improvedthe strategic orientation ofbudgets. 36. Coordination between line ministries and centra1 ministries i s also improving. This improvement is a consequence of the process of elaboration of program budgets and MTEFs, as well as the PRSP process. The elaboration o f program budgets and MTEFs has led to the organization o f inter-ministerial workshops to discuss and, in principle, decide on proposed allocations for ministries. These inter-ministerial workshops include a key group o f ministries, the Ministries o f Economy and Development, Finance and Budget, Basic Education and Literacy, Health, and Agriculture. According to Ministry of Finance officials, this inter- ministerial dialogue has improved since the adoption o fthe PRSP and MTEFs.~ 37. Results orientation, monitoring and evaluation. In general, the culture o f evidence- based policy-making is a tenuous and recent one (as is the case in most countries). The PRSP process, with its focus on monitoring and evaluation and poverty results, however, is slowly fostering a learning process for results-oriented policy-making. In particular, the elaboration o f PRSP progress reports and the consultation process leading to these reports i s making an important contribution to strengthening this culture, especially in the MEDEV, the MFB, and priority sector ministries. Similarly, the orientation o f donor support toward results i s increasing incentives for using output of monitoring and evaluation in policy decisions. The delegation of credit for the purchase of goods and services in the health sector, for instance, was the direct consequence of a study, which showed that the over-centralization of expenditure management was impairing spending effectiveness. Similarly, survey-based information regarding the under- utilization of health services in rural areas has led to the effort to re-deploy qualified health personnelto these areas, 38. Overall, the PRSP process has had some positive effects in enhancing the results- orientation of the budget process. Inparticular, the results-orientation o f donor budget support programs is increasing the attention paid by the MFB and MEDEV to the links o f budget inputs with outputs and intermediate indicators. For example, the inability of the Ministry of Basic Education to produce routine data on school enrollments resulted in late disbursement o f the variable tranche of the EU's budget support program, causing strains in budget execution. This outcome has focused attention o f the monitoring teams on accelerating the production o f key indicators. 39. Technical personnel in central and decentralized government ministries interviewed for the CDF review believed the PRSP had had a large impact on the 10-year plans for education, health, rural development and water regarding their management, their results-orientation and The link between the 200412006 MTEF and the PRSP, however, is not clear as budget allocations for priority sectors are either flat or declining. - 10- their monitoring. On the other hand, they found the impact o f the PRSP process on other (non- priority) sectors to have been weak." 40. Even inthe priority sectors, however, there is room for improvement. Inthe health sector, for instance, the combination of administrative and financial autonomy o f public hospitals and primary health centers with the absence o f performance contracts or performance monitoring agreements orients these service delivery units toward the wrong "results" (improving their financial situation) rather than toward improving health outcomes, inparticular for the poor (see section on public expenditure in the health sector). The process o f establishing, reporting and monitoring outputs and other results-oriented indicators is somewhat more advanced in the ministries o f health and basic education due to the presence o f well-established donor-supported programs. Even in these sectors, however, there are difficulties and the execution o f program budgets is not properly monitored, in part because o f recurrent shortcomings in reporting in a timely manner on basic outputs and other indicators. For example, the health ministry acknowledges it is only able to produce basic statistics six months after year-end. Finally, the PRSP monitoring and evaluation system i s not sufficiently integrated with the public expenditure review process and the monitoring systems for sector programs in priority areas. This lack o f coordination leads to a multiplicity o f monitoring systems, and does not foster linkages between the overall strategy (PRSP), sector strategies and the budget. 41. The recent split o f the former Ministry o f the Economy and Finance into two separate entities-the Ministry o f Economy and Development (MEDEV) and the Ministry o f Finance and Budget (MFB) has also weakened the results-orientation o f the budget process. During the transition, strains appearedinboth the planningPRSP process controlled by the MEDEV and the budgeting processes controlled by the MFB, as personnel and expertise have been divided between bothministries, with limited coordination o f activities. 42. Further improvement is warranted regarding the technical and institutional capacity for monitoring and evaluation and for enhancing the capacity and incentives of ministries to make use o f results. Institutional mechanisms to support inter-ministerial dialogue are weak. Well- functioning inter-ministerial working groups are key to supporting the ongoing dialogue necessary for the coordination o f overall strategy (PRSP), sector strategies, MTEFs and budgets. At the moment, however, there are perhaps too many sets o f working groups with overlapping mandates, unclear terms o f reference, poor communication and tenuous links to the policy- makingprocess. At the beginning o f the PRSP process, six working groups with representation from govemment, civil society and donors were created in the following areas: macroeconomic and budgetary; health; education; governance; rural development; and competitiveness. However, neither their terms o f reference nor their expected outputs were clearly defined, nor no means were provided for their functioning. 43. Next to these exclusively governmental working groups, there is another set of working groups formed exclusively by donors. These are thematic working groups in the areas o f macroeconomic issues, statistics, gender, health and education. These working groups work well as a forum for the exchange o f views among donors, but, like the purely inter-ministerial lo Evaluation du Cadre de Developpement Integre (CDIj. Etude de Cas: le Burkina Faso. D. Lallement, D. McMillan, K.O'Sullivan, P. Plane andK. Savadogo. Ouagadougou :September 2002 :p. 11, - 11- working groups, they do not allow for much-neededdialogue between government, donors and civil society. 44. At present, the only functioning working groups including government, donor and civil society representation are those monitoring donor-supported programs in education and health. These groups, however, meet only twice or three times a year and focus on monitoring sector programs while not systematically tracking links between PRSP, sector strategy, MTEF and budget allocations. 45. The draft revised PRSP for 2004-2005 aims to draw lessons from the experience o f the past three years in designing a new monitoring and evaluation architecture. Two key lessons from the past, however, do not seem to be reflected in the proposed new structure. First, the focus of the new monitoring and evaluation structure continues to emphasize separate working groups for government and non-governmental actors, with only one instance o f joint membership. A key problem over the past three years has been weak communication, dialogue and informationexchangebetween govemment and donors. A second shortcoming has been the treatment of HPC resources, the PRSP and other donor support with a "project mentality." This approach has weakened the overall coherence and effectiveness of govemment policy. The current proposal to limit the monitoring role of donors to "the effectiveness of aid" entrenches this approach, which is untenableinacontext o fgeneralbudget support. B. PARTICIPATORYNATURE THE BUDGETINGPROCESS OF 46. Public availability of budget data is limited, though there has been some progress since the advent of the PRSP process and the move toward budget support. Budget documents are public and annual budget laws are publishedin the official gazette. However, there is no public information center where the public can obtain information on the PRSP process and related policies and budgets. Moreover, difficulties in obtaining budget data were identified as a problem inthe CFAA and the ROSC, which was confirmed by recent interviews with members ofthe donor community and civil society. 47. One of the key recommendations o f the 2002 CFAA was to "substantially improve the availability of information on government policy objectives, budget preparation and execution available to the National Assembly and the public." The National Assembly adopted a new budget law in January 2003 to enhance the quality, availability and timeliness o f the publication o f budget data. The law followed WAEMU guidelines and included provisions to ensure the comprehensivenessof budget laws, including multi-year projections of debt service obligations, lists of all special accounts and related expenditures, the government's balance sheet, monthly projections for budget implementation, and functional and economic classifications for the budget. Due to constitutional reasons-the Burkinabk constitution describes the applicability o f organic laws narrowly-the law was promulgated as an ordinary law.' Making further progress in applying the law and providing timely information to the public on budget preparation and " The constitutionadmits the principle of organic laws only for a limited number of institutions, such as the constitutional court and the economic and social council but has no general provision oforganic laws. - 12- execution to enhance the PRSP process is key to improving accountability and enhancing the efficiency inthe use o fpublic hnds. 48. The budgeting process is formally transparent but the reasoning behind resource allocations is not always clear. The budget process starts with the allocation o f resource envelopes for each ministry based on the macroeconomic projections o f the Ministry o f Finance. Subsequently, line ministries prepare budgeting requests using initial allocations and the respective sector strategy. Finally, a process o f arbitrage takes place with the MFB and the Council o f Ministers under the leadership o f the Prime Minister. Once the draft budget law is approved by the Council o f Ministers, it is submitted to the National Assembly, which usually amends it andthen approves it. However, the basis for decision-making inthe resource allocation process is not always clear to line ministries, local governments and non-governmental actors. This lack o f effective transparency is compounded by an annual process o f resource re-allocation inmid-year through a supplementary budget law; the lack of information inthe process and the criteria followed inthe allocation o f significant and rising "common inter-ministerial resources;" and the important role played by the freezing o f credit lines in the cash budget management system. 49. Information on budget execution i s available to the government and donors during the year, but not to the wider public. The MFB prepares detailed data on budget execution for the Council o f Ministers. Availability o f these reports to other ministries, notably the MEDEV and ministerial planning departments is important to allow them to closely follow budget execution from a programmatic perspective in the context of PRSP monitoring. Donors usually receive a monthly fiscal table (TOFE) within three weeks after the end o f the month, and recently this table has been accompanied by budget execution data by chapter and ministry. The broader public must rely on summary reporting in PRSP progress reports and budget execution reports publishedwith considerable delay inthe official gazette. For example, the final budget execution reports for 1995-2001have been adopted by the National Assembly in2003. 50. The Role of the National Assembly. The National Assembly has a significant role inthe budget process. This role is structured into five main phases. First, the National Assembly receives technical reports from line ministries explaining their sector programs and financing requests for the upcoming budget year. After these reports have been received, conversations are held between members o f the assembly and technicians at the various ministries. These interviews are followed by the elaboration o f studies by various committees on issues o f their. concern. An overall report on the draft budget law is elaborated for National Assembly's public finance committee. Finally, the National Assembly amends and approves the draft budget law. Halfway through the year, the National Assembly may intervene again by amending and endorsing the supplementary budget law. Although, according to WAEMU directives, the National Assembly should approve the MTEF, this has not yet been the practice in3urkina Faso. Similarly, program budgets are not submitted to National Assembly and, hence, the debate focuses on the traditional input-budget, which is essentially an administrative budget. 51. The National Assembly seems to have limited ownership of the PRSP process. A comparative analysis o f draft budget laws and budget laws over the past few years shows that National Assembly has so far increased budget allocations to non-priority sectors and, in - 1 3 - particular, for unallocated "common inter-ministerial funds" while systematically reducing allocations to priority sectors (Table 1). In fact, the largest percentage expenditure reductions enacted by the National Assembly between 1998 and 2002 havebeen inflicted on the Ministryof Basic Education and Literacy, the Ministry o f Agriculture and the Ministry o f Livestock. Therefore, defacto, the National Assembly's involvement in the budget process has weakened the link between the PRSP and the budget. This finding is, in part, due to the fact that the National Assembly's priorities do not coincide with those reflected in the PRSP.I2 It is also, however, due to the weak involvement of National Assembly inthe PRSP process. The National Assembly has not adoptedthe PRSP, has rarelybeen involved indebatesover the PRSP, andhas not approved or endorsed PRSP progress reports. Involvement of individuals, though, seems to be increasing as some memberso f National Assembly did take part in the forum to review PRS implementation for the second PRSP-PR. Table 1:NationaiAssembly's Priorities:4 PercentageChangefrom OriginalAllocations Average 1998 1999 2000 2001 200298102 Ministere de la Defense 0.0 -0.4 0.0 0.0 3.4 0.6 Ministkre de la SantC 0.9 -2.9 -1.7 -2.7 -2.6 -1.8 Ministere de 1'Enseignementde Base et de 1'Alphab -5.6 -5.6 -5.0 -4.7 -5.5 -5.3 Minde 1'EnseignementSec, Supet de laRechScient -3.8 1.2 -2.8 0.9 -2.5 -1.4 Ministhe de 1'Agriculture -5.2 -8.3 -5.2 -1.8 -2.6 -4.6 Ministere des RessourcesAnimales -9.3 -0.4 -6.3 1.3 -6.0 -4.1 Ministere de YEnvironnement et de 1'Eau -2.4 -2.5 -1.1 -0.3 9.5 0.6 MinInfrastruct, Habitatet Urbanisme. -5.9 -1.6 -0.3 0.4 -1.4 -1.8 Depenses Communes Interministerielles 14.8 11.9 11.0 16.8 21.4 15.2 Total budget 0.0 -0.6 -0.8 1.2 4.2 0.8 52. Implication of Civil Society. Burkina Faso has a long-standing tradition of social mobilization. The PRSP process builds on this tradition and has fostered debate on the country's overall development strategy and on its synergies with sector strategies among government, the private 'sector, and civil society-in particular NGOs. This debate has centered on the participatory processes for the elaboration o f the original and revised PRSPs as well as the workshops organized for the elaboration o f annual PRSP-PRs. Civil society has very much welcomed this opportunity for involvement and acknowledges that government openness to their participationhas deepened as the PRSP process has advanced. 53, Civil society participation, however, remains relatively shallow and the budgeting process currently does not include any entry points for its involvement. Civil society has very little information on the government's budget and, although the budget itself is published, there are no initiatives in place to publicize it. Similarly, there are no organized forums in which to share and discuss the annual budget guidelines with civil society. NGOs are particularly interested in improved availability o f budget information and are interested in increased l2 This lack o f ownership of PRS priorities is reflected in the view shared by a senior member o f the National Assembly during interviews with the PER team that primary education and primary health care are "the responsibility o fthe donor community and the NationalAssembly needs to watch out for other ministries." - 14- involvement in the budgeting process. The capacity o f civil society to effectively engage in budget analysis, however, is currently very weak. Civil society organizations are conscious of this shortcoming and have requested the support o f the World Bank Institute to develop and implement a medium-term capacity-building program focusing on strengthening their skills in the areas o fbudget andbasic economic analysis. 54. Existing capacity in Burkina Faso's research institutes is currently under-utilized. In some PRSP countries, such as Tanzania, research institutes have greatly contributedto the PRSP process. They have provided high quality, locally-rooted, independent poverty analysis, worked with the government in producing PRSPs, PRSP progress reports and other monitoring and evaluation documents and are key catalysts in forging links between government and non- governmental organizations. This dynamic is much weaker inBurkina Faso despite the existence of a number o f capable policy-oriented research institutes and their potential contribution to the above-mentioned areas. These institutes could also be key players in organizing and delivering the capacity-building programs that are essential to improved government ability to deliver, implement and monitor sector programs, program budgets and medium-term expenditure frameworks. c. BUDGETALLOCATIONSAND EXECUTIONUNDERTHE PRSP 55. This section assesses the degree o f coherence between PRSP objectives and resource allocation by the government through the budget process. To understand the relationship between the budget and the PRSP, the section discusses the evolution o f aggregate expenditure inpriority sectors; expenditure across and within priority areas and priority accorded within the budget to PRSP priorities; PRSP costing estimates and actual budget allocations; and execution rates o fpriority expenditures. 56. The analysis below draws on data from the draft budget laws, budget laws, the 2004-06 MTEF, and information on executed expenditure for 1998-2002 from the integrated computerized expenditure management system (CID). For 2003, we have usedthe budget law, as data on executed expenditures was not available when the report was drafted. Since the data regarding foreign investment is not comprehensive in the budget laws andnot available from the CID, it has been supplemented with information from the public investment program execution reports and from the cooperation directorate at the Ministryof Finance (DGCOOP). 57. The data processing work has been difficult for four key reasons. First and foremost, the non-inclusion o f most foreign financing in the integrated expenditure management system has required the piecing together o f information from a number o f other sources. Second, the lack o f clarity as to the distribution o f HIPC resources and their non-inclusion inthe budget law or inthe integrated expenditure management system in2001 and 2002 has required the use o f a number o f alternative sources and making some assumptions as to their actual use. Third, the absence o f a full functional classificationo f expenditure has forcedus to use various administrative units and - 1 5 - rs arid Figure3. - 17- - t x - ding - 20 - 170 1!,O 1.30 1I O pi: .IO '30 10 ~ lI i l ** . , ~ " 2000 2001 2002 a a Figure12. Bud ry f 2000 7 2002 - 2 3 - Flgurc 17. HlPC r CFA bn rt, realterms Figure 23$Budget Support: Quarteriy Dlsbursamsnts 7 2 3 4 Quarter Y E l External ftoanctng I 20002001 2 government ministries, civil society and the private sector, with the National Assembly still playing a rather marginal role. 87. The emphasis of the PRSP process on openness and transparency and the move o f some o f the donor community to budget support has increased pressure on the government to produce and publicize information, in particular budget data and basic statistics. Civil society is also increasingly requesting access to timely and user-friendly budget data in order to perform their role o f monitoring budget implementation and, in general, holding government accountable. There is also a greater demand for transparency inbudgeting, in particular regarding the role o f the inter-ministerial resources and the decision-making process to determine the freezing o f budget lines duringthe budget execution process. 88. The stress on participation has led to a still unfulfilled expectation for fbrther opening the budgetary monitoring processes to civil society. Although, as mentioned above, there i s greater debate about policy issues in large forums, true participation in monitoring and evaluation has not yet materialized. 89. There has been a noticeable improvement in the coordination between planning and financing departments, in particular in priority sectors, as well as between line ministries and central ministries. The fact that donor support i s based on results seems to be playing a crucial role in fostering this change. 90. The PRSP process and the availability o f HIPC funds have had a positive impact on resource allocation. Aggregate real expenditure inthe PRS priority sectors and core priority areas (primary education, primary health and rural development) has doubled between 1999 and 2002 both in absolute as well as inper capita terms. These expenditure increases, however, have been quite uneven across sectors. 91. Intra-sectoral prioritization o f budget allocations was strong in the education sector, but not in the health or rural development sectors. While the primary education sector received 80 percent o f the overall sectoral expenditure increase, only half o f the increase in health expenditures benefited the primary health sub-sector (and were all targeted HIPC resources). Almost the whole increase in priority sector spending over 2000-2002 is due to increases in developmentlinvestment spending, with only very small increases in recurrent spending, raising concems about the adequatebalance o f investment and recurrent spending. 92. Duringthe 2000-2002 period, the government has increasingly protected priority sectors from cuts during budget execution, Execution rates o f priority sectors have been significantly higher than those in non-priority sectors, indicating the strategic orientation toward PRSP objectives in implementing cutbacks o f budget allocations based on optimistic revenue projections. In economic terms, execution rates for wages and salaries have been highest and lowest for foreign-financed investment spending and HIPC. Mobilization and programming o f foreign aid and use o f HIPC resources could be improved to further increase the resources available for poverty-reducingprograms. - 3 1 - 93. Gaps between PRSP costingestimates and actual expenditures inpriority sectors vary, in large part, depending on the degree of donor involvement in the sector and on actual disbursements. Inthe health sector the gaps between PRSP projections and actual expenditures are large, reflecting the weak mobilization o f budgetary resources by the sector after the closing o f some foreign-financed projects in2001. 94. As is the case in other countries, the overall level o f foreign aid has not increased since the onset o f the PRSP process, but there is an increasing trend toward budget support. There has, indeed, been a change inaid modalities, with program grants and loans increasing from 2.6 to 4 percent o f GDP between 1998 and 2002. This increase is mainly due to greater availability of program loans (by 1.2 points o f GDP). 95. The predictability of aid continues to be limited and its variability high, which poses important problems for budgetmanagement, There is progress inthe area o f aid forecasting with a number o f donors now providing three-year aid disbursement estimates, which facilitates the government's task to elaborate and implement MTEFs. The gaps between aid forecasts and actual disbursements, however, are considerable and deserve serious attention by both government and donors to ensure that they diminish in future. Finally, aid tends to bunch heavily at the end o f the fiscal year, which poses great problems for proper budget management and execution. A better coordination of aid disbursements with the government's budget cycle i s essential for the government to deliver on PRSP objectives. 96. Both aid and the budget have become increasingly aligned with PRSP priorities. h 1998- 1999, the gap between foreign assistance and priority sector expenditure was roughly 6 points of GDP. Since the advent of the PRSP in 2000, the gap has progressively narrowed, reaching 2 points o f GDP in 2002. Overall, foreign financing has not increased very much since 1999 but significant resource re-allocation has taken place with most increases benefiting the priority sectors. 97. Key recommendations. Despite significant progress in budget management and increased focused on results, the government would need to adjust the budget planning process to better align it with the PRSP in the future. Inparticular, a revival o f the program budgets (or sectoral MTEFs) and muchcloser integration with the PRSP planningand monitoring process is warranted to avoid imbalances in allocations to priority sectors both across and within sectors. This subject is developed in more detail in the next chapter, as well as chapters 4 and 5, which deepen the analysis o f spending in the health and education sectors. Further opening of the programmatic discussions to participation and publication of budget data is needed to foster the participative nature o f the PRSP process. Finally, additional efforts by donors and the government are needed to improve the predictability ofaid disbursements. Inparticular: Improvebudgetplanning,execution, and monitoring e Strengthen the elaboration and monitoring o f program budgetsand MTEFs and feed back results o fprogram budget implementationinto the PRSPprocess (see chapter 3) to ensure consistency betweenPRSP, MTEFs, andprogram budgets - 32 - Based on the budget execution data for 1998-2002, systematically review the level o f recurrent and investment spending, the level of fbture expenditure increases in priority areas, and the definition o fpriority sectors linkedto keyPRSP objectives Improve budgetary programmingand monitoring o f foreign aid Identify core priority spending in the functional budget classification and report consistently on budget implementation inPRSPprogress reports Strengthen hnctioning o f inter-ministerial working groups, extend their tasks to ensuring consistency between PRSP, sector strategies, MTEF and annual budget and include donors, private sector and civil society intheir work program Enlarge the system o f delegated credits to all Ministries with deconcentrated bodies and further deconcentrate issuance ofpayment orders to line ministries and regionally Improve efficiency o f a priori expenditure control systems Improve the availabiiity of budget data 0 Set up a public information center inthe MFB and MEDEV carrying all main documents related to the PRSP, sector policies, budgets and donor programs andprojects 0 Publicize within-yearreports on budget execution 0 Include program budgets and MTEF in the annual budget submitted to National Assembly 0 Increase the capacity o f civil society and National Assembly in budget and sectoral policy analysis Improve the predictability of foreign aid 0 Agree with donors on three-year projections regardingproject andprogramaid 0 Better synchronize aid disbursementwith the govement's budget cycle 0 Base each year's disbursements on the performance from the previous year to increase predictability of aid. 111. INTEGRATINGBUDGETPROCESSESAND THE PRSPREVIEW 98. This chapter discusses practical ways o f incorporating the annual PRSP review into the budget process. Since the PRSP was adopted in 2000, the annual reviews have taken place during the third quarter of the year, at a time when the global MTEF and the preliminary draft program budgets have already been developed. The results o f these assessments have not really influenced development o f the MTEFs and the subsequent budgets. The purpose`of this chapter i s to propose a framework for better incorporating the PRSP review into the budget process. The first section analyzes the strengths and weaknesses o f the program approach adopted inBurkina Faso in 1998. It presents the objectives of the reform, describes the experience ofthree ministries (Health, MEBA and MESSRS) and analyzes the causes of the shortcomings observed with respect to application of this approach. The second section analyzes the links between the program budget and the PRSP. Based on this analysis, the third section then proposes a framework to incorporate PRSP monitoring into the budget process. The central hypothesis - 33 - developed in this chapter is that the results o f the annual PRSP review should be used as a key input for budget formulation. A. PROGRAMBUDGET REFORM: MEASURES, RESULTSAND PROSPECTS 99. The program budget reform project. In 1998, the government o f Burkina Faso undertook a major reform o f the system o f budget allocations by adopting the program budget approach. This decision was based on the acknowledgement that the traditional system o f allocation based on broad economic classifications and administrative units was not really grounded in rational choices. In this system, budget proceedings focused primarily on the resources to be made available to departments, without considering the main objectives pursued by these departments, namely the provision o f public services (education, health, outreach, control, protection, etc.). This approach had created incentives to artificially inflate funding requests inthe hopethat arbitration would ultimately result ina satisfactory level o f funding. The program budget reform sought to introduce major changes inthis regard, inline with the general objectives presented below. 100. To introduce program budgets, the then Ministry o f Economy and Finance (MEF) set forth the project's objectives, cost, duration and implementation strategy in a project paper entitled Development and Implementation o f a Program Budget in Burkina Faso. The stated general objectives were to improve budget management through gradual implementation of the program budget approach. With respect to revenue, the goal was to introduce the concept o f results-oriented management, i.e. to hold authorities accountable for tackling the objectives to be achieved in terms o f collection, quality of services and productivity and to give them a certain degree o f management autonomy in reaching these objectives. With respect to expenditure, the goal was to introduce objective-oriented management by training budget managers in the techniques o f program budget development, specifically through the following actions: (a) help government set multi-year objectives, make rational choices and establish priorities; (b) translate the government's intention to achieve its objectives through implementation o f consistent programs; (c) earmark allocations for agencies charged with implementing the activities; (d) strengthen the hierarchical oversight; and (e) train instructors inorder to ensure the sustainability of the project. The total project cost was estimated at CFAF 487 million, including an expected external contributiono f CFAF 366 million that was not obtained. 101. The project's implementation strategy was exclusively based on training the appropriate actors with assistance from an intemational firm spread over two phases. The first phase involved training for the Financial Affairs Directorates (DAFs) and Studies and Planning Directorates (DEPs) o f the sectoral ministries, while the second phase involved training of the instructors who were to ensure the sustainability of the project. The project was supposed to begin in 1998 and runfor three years. 102. Programbudget implementation.Inline with the timetable stated inthe project paper, the project was launched in 1998 for the 1999budget. Guidelines were issued to this effect inthe budget circular. Priority was given to the six pilot ministries to train the actors responsible for developing the program budget. However, all the ministries were invited to adopt the program - 3 4 - approach. The guidelines were followed fairly well. Furthermore, this effort was then supplemented in 2000 when a global MTEF was developed, permitting budget allocations for sectoral ministries to reflect macroeconomic constraints, particularly resource constraints. Since that time, budget allocations have been appended to the circular each year to guide the development o fprogram budgets by the sectoral ministries. 103. Despite these accomplishments, major shortcomings continue to limit the true scope of the program budgetingexercise. Infact, the objectives o fthe reform are far fiom being achieved. Inparticular, the very essence of the program budgets is subverted by the fact that budget proceedings continue to focus primarily on administrativebudgeting. As a result o f this situation, the sectoral ministries increasingly consider the program budget to be a simple exercise o f form rather than substance. Inthis regard, it is interesting to note that the quality o f the initial program budgets of some ministries was well above the quality of those prepared more recently. Many ministries now view the exercise o f program budget development as additional work with little benefit or consequence for their operation. In fact, the general opinion is that program budget reform inBurkina Faso has failed to change the traditional budgeting system, 104. Several criteria are available to assess the quality o f program budgets. A good program budget should include the following information: (a) existence o f a sector policy paper that presents the sector's current government o f development and main problems, on the one hand, and appropriate responses in the form o f objectives to be achieved, on the other hand; (b) whenever possible, programs should have quantitative objectives, and the methods used in setting the objectives should be transparent; (c) the programs should be consistent with the problems that have been identified; (d) a list o f indicators for measuring the results and progress achieved; (e) an adequate information system (statistics available on a timely basis; reliable data); and (f)proposal and implementationo f sector policy reforms. 105. Program budgets of the Ministry of Health. The program budgets developed by the Ministry o f Health remain incomplete and reveal a number o f shortcomings. The program budgets for 2000-2002,2002-2004,2003-2005 and 2004-2006were examined. These documents generally contain three parts: (a) a summary statement that repeats the sector policy objectives and lists programs by objective-in this regard, it is interesting to note that the program nomenclature has only recently been standardized, beginning with the preliminary draft budget for 2003 after the National Health Development Program (PNDS) was zdopted in 2001; for the years prior to 2003, there are variations in objectives and programs; (b) a table o f budget estimates, comprising seven columns that identify the list of programs and activities, the entities responsible for the activities (this column permits transition between the program budget and the administrative budget), the duration o f the programs, their total cost, cumulative prior expenditure, allocations for the year, and three-year projections; and (c) a transition table between presentation o fthe program budget and administrative budgeting. 106. The main strength o f this document is the existence o f a well-defined sector policy and a ten-year program, the PNDS. The main objectives and the list o f programs are drawn from the PNDS adopted in 2001. However, these advantages are not fully exploited. In fact, some important tools of budget programming are lacking in the program budget for the Ministry o f Health. First, these are budgets that contain no financial plans. Second, there are no annual - 35 - targets for indicators of physical progress. Third, there are no sector indicators. In addition, the method of updating the three-year budget programming is not sufficiently participatory and exact, The updating is done at the DAF in consultation with the central entities responsible for the programs and deconcentrated entities are not involved. The status o f budget execution by program is not available. The columns on cumulative prior expenditure and current year alIocations contain no data. Overall, the program budget for the Ministry of Health remains imprecise as a guide for sector policies and spending. 107. Program budget of the Ministry of Basic Education and Literacy (MEBA). Until 2002, MEBA's program budgets were largely not operational. Indeed, although the Ten-Year Basic Education Development Plan (PDDEB) was adopted in 1999, implementation did not really begin until 2002. Efforts were made in 2002 to present the ministry's program budget according to the nomenclature o f PDDEB, which includes 18 programs. In 2003, budget formulation expanded to deconcentrated units as a result o f the new PDDEB implementation arrangements, which call for a greater role for the Basic Education Provincial Directorates (DPEBAs). It was thus possible to develop a fairly satisfactory 2004 program budget, keyed to the various PDDEB components. The ministry's goals and missions are summarized in a brief. The programs are classified inaccordance with the eight PDDEB objectives, which are based on indicators of measurable results. The choice o f the eight objectives isjustified on the basis of an analysis of the current status (results achieved). This analysis is used to identify needs and propose actions to solve the statedproblems and advance toward objectives. 108, The costing o fprograms andresults measurement is satisfactory. Each program is briefly described and its relevance justified. The total cost i s given, along with budget estimates for the fiscal year in preparation. Quantitative indicators for measuring physical results are presented. However, such information is not summarized inthe standard tables designed for that purpose to facilitate programbudget analysis. Furthermore, it is not clear whether HlPC resources are taken into account inthe various programs. Table 3: Listof PDDEB ObjectivesandIndicatorsfor the Period2004-2006 0bjectives Indicators Target Objectives Objective 1:Expand educational Gross enrollment ratio (GER) Raise the GER from43% in2003 supply to 52% in2006 Obiective 2: Reducethe GER ofthe province with the Raise this province's GER from geographic disparities lowest ratio 18% to 23:4% Objective 3: Reducethe gender 1IGross enrollrnent ratio of girls Raise girls' GER to at least 70% disparities in2005 Objective 4: Improve the quality CEPE successrate: t5points of education Repeater rate: - 3 points Objective 5: Literacy Literacy rate Raise the literacy rate from 2%.4%to 32% in2006 109. The MEBA program budget exemplifies how results can be evaluated based on a fairly straightforward framework. Table 3 lists the objectives, indicators and target objectives set for the period 2004-2006. It shows that program budgets constitute in themselves a complete framework for policy assessment. An assessment based on budget performance, physical output - 36 - and analysis o f the indicators attached to the objectives listed above should indeed suffice to identify key changes that have taken place duringthe year. However, the global MEBA program budget still would need to broken down internally to be operational for regional and provincial monitoring. Among other information, a breakdown o f resources by region that shows each region's share of each program would be required with the ongoing deconcentration o f budget management. In this manner, regional objectives would be more visible and would be tied to available resources. The basis for such a breakdown already exists since the ministry has set minimum standards for classroom furniture (tables and benches, cabinets, desks, chairs), teaching materials (dictionaries, compendiums, world globes) and consumables (notebooks, chalk, slates). 110. The programming and budgeting exercise already undertaken at the DPEBA level (action plan development and implementation) augurs well for the future development o f program budgets at this ministry and for the incorporation o f PRSP monitoring in the budget process. DPEBA accountability provides an operational fkamework for the regional PRSPs. The sectoral MTEFwill be a useful instrument for consolidating regional and provincialplanning. 111, Program budgets at the Ministry of Secondary and Higher Education and Scientific Research (MESSRS). MESSRS has one o f the most highly developed program budgets due to the presence o f strong advocates within its ranks. This ministry is one of the few that have consistently maintained outstanding quality in their program budgets as can be confirmed by an analysis of the 2001-2003 preliminary draft budget. The ministry does not have its own sector policy paper, but the sector policy goals are based on the educational development plan prepared after the National Convention on Education, which also provided inspiration for the PDDEB. The sector policy goals, missions and objectives are enumerated and the sector policy implementation strategy is briefly described. A list o f indicators is provided, along with targets. Similarly, indicators for measuring physical results are available. A table o f budget estimates displays the total cost o f each program, prior expenditure, allocations for the year in progress, and three-year programming. A table projecting the expected results o f each program is also provided. 112. The DEP is charged with the development o f the MESSRS program budget. Inpractice, the program budget is developed by a small group o f people under the auspices of the studies, planning and projections unit. The development is facilitated by the fact that the ministry supervises one of the highest numbers of independent public agencies (EPAs), creating a situation that facilitates the development o f program budgets based on performance contracts with EPAs, which receive transfers from the ministry to conduct their work. 113. A central committee basesprogram budgeting inthe MESSRS on a breakdown o f global MTEF allocations. An internal ad hoc budget committee, chaired by the minister and composed of DAF, DEP and certain other central entities, performs a breakdown o f the MTEF envelopes and allocates resources to each entity based on spending priorities. However, it has been recognized that the minister is subject to many different pressures, andit sometimes happens that the arbitration moves forward at the expense o f certain spending priorities. Based on the allocations, each entity develops its own draft budget and defines the objectives to be achieved, - 3 7 - the programs and activities, andtheir costs. These data are centralized by DEP, which uses other sources of information, particularly sector statistics, to develop sector objectives. 114. Objectives are closely linked to past outcomes and available resources. With respect to setting the objectives, the initial data are taken from the statistical yearbooks and the objectives are projected on the basis of expected physical output, which depends on budgetary resources. The difficulties experienced in this area are due to the fact that the MTEF envelopes are not sufficient for achieving the sector objectives. Often a number o f activities are postponed fi-om one year to the next without ever receiving any funding, as the ministry has no additional resources beyond its MTEF envelope. . Inthe words of one ministry official, the program budget is said to be a "sliding" budget. Other problems encountered are the frequent change in staff and the lack o fprotection o f salaries for teaching staff. 115. Causesof the mixedresultsobtainedin the implementationof programbudgets. The main causes o f the shortcomings observed in the implementation o f program budgets are as follows: Absence of a sustained effort to follow through on the reform. Soon after this initiative was undertaken, its main developers were dispersed. Despite the efforts of their successors, progress has been very slow. Hasty pursuit of widespread adoption o f the approach without adequate resources to ensure success. Although a project paper was prepared, it did not receive adequate funding. Inthe project paper, the government was counting on external fundinginthe form of a grant (CFAF 365 million) representing 75 percent o f total project cost. This funding was not received, but the project was launched nonetheless with funding from the national budget, which was inadequate. Lack of computer applications to make it easier to review the figures and adjust the objectives at any time. One o f the shortcomings o f the conceptual document for the project i s that it underestimates the computer problems associated with this type o f approach, where iterations are necessary untilfinal adoption o f the budget by the National Assembly. Subsequent lack of data on budget execution consistent with program budget nomenclature. Information on budget performance is only available based on economic classification (title), ministry (section), administration (chapter) and purpose or nature o f the expenditure (paragraph and article). Program budgets thus become "relics" as soon as they are prepared since their execution cannot be checked. ~ n s t i t u t ~ ~ n a ~ No attention was given inthe project paper to composition and expertise issues. o f the budget arbitration committee, existence of relevant and competent entities for assessing the performance o f sector policies and programs, which would handle the arbitration. Incomplete accounting of funding availablefor programs, Some direct funding from donors to beneficiaries is not included inthe official statistics on external aid. - 38 - DifJiculties in setting quantitative objectives in a transparent manner. To date, with the exception o f a few ministries that possess embryonic models for setting target objectives (the agricultural module o f the Automated Projections Instrument, MEBA's Education for All model, etc.), inmost sectors objectives are set ina fairly approximate way. Failure to rigorously adhere to the budget schedule, which limits the amount of time given to sectoral ministries to develop their preliminaly draft budgets. The budget guidelines (circular) have not been published on time inrecent years. This situation does not permit true participation b y key ministry officials in the budget process.' The PRGB has already highlighted this problem andunderscored the importance o f adhering to the schedule. e ~eakparticipationby deconcentratedentities in budget development.This problem is related to the previous point. Deconcentrated departments, situated at some distance from the center, are the first victims o f the short time frame given to ministries to formulate their budget proposals. Some DAFs, for example at the Ministry of Health, try to compensate for this situation by starting to centralize all needs well before the circular has been issued. But, in any case, arbitration is performed without the participation o f deconcentrated entities and does not always reflect their priorities. Separation between the budget preparation process and the process of assessing and updating sector programs. At the present time, in many sectors there is not yet any mechanism that includes a schedule for systematically updating the costs o f existing programs and proposingnew programs. Unreliable revenue projections (see also the discussion in chapter 2 on MTEF and credit freezes). In2000-03, revenue projections included in the budget circular and the budget law were too optimistic. This situation, in combination with the difficulties encountered in mobilizing external budget support, has often led to adjust spending so as to contribute to macroeconomic stability. a Centralized system of budget execution. The centralization o f controls and payment order issuance has not been conducive for flexible and efficient budget management. Budget lines for deconcentrated entities have been created and are more or less functional, andprocedures for budget deconcentration have been adopted. However, funds are still being made available belatedly (April or May), deconcentrated credit allocations are small compared with the total budget, and the delegation o fpayment order issuance to the regions is still inits infancy. The recent extension o f the CID to five regional capitals augurs well, however, for further progress inthis area. Failure to update regularly and systemati~all~the list of indicators of sector results. Analysis o f the results based on the indicators is limited by the availability o f statistics. Most of the statistical yearbooks are produced six or seven months after year-end, at a time when the budgetprocess i s already under way. Absence of a marginal budgeting modelfor key sectors. -39- B. LINKS BETWEENPROGRAMBUDGETS AND THE PRsP 116. Ideally, analysis of the. progress made in implementing the PRSP should rely on the framework provided by program budgets. In practice, however, the two processes unfold completely separately at the present time. 117. Budget preparation process. Currently, the government's budget preparation process is , spread over roughly seven months and comprises five successive steps: (a) development and distribution of the budget circular; (b) development arid submission o fpreliminary draft budgets; (c) budget proceedings; (d) review by the Council o fMinisters; and(e) budget proceedings inthe Assembly. The first three steps are critical stages of the budget development process. 118. Development o f the budget begins officially in May. The internal regulations prescribe that during the first week o f May each year the MFB distributes the circular signed by the President o f Faso, which establishes the general guidelines for budget preparation. Since 2001, the global MTEFhas been appended to the circular to provide provisional ceilings for the budget proposals. The global MTEF is normally elaborated in March and adopted in, April by the Council of Ministers. It shows already committed resources and additional resources that can fund new programs or extensions o f existing programs. These additional resources are generally dividedamongthe priority sectors. The MTEFsets an envelope for each ministry and each title. 119. The global MTEF used for budget preparation i s relatively rigid. The envelopes are divided in advance by title, which limits the room to maneuver for the core ministries and the sectors as regards the strategic allocation o f additional resources. 120. Most ministries follow an imperfect and approximate mechanism to break down their budget allocations by administrative entities and programs. Ingeneral, the authorities of each ministry or sector should rely on the global budget circular to develop an internal circular that includes, as an annex, a breakdown o f the global resources received by the ministry's programs. This breakdownshould guide the developmento f the program budgets. The operational entities responsible for programimplementationshould intumbudget for the activities and set attainable objectives based on total available resources (own funds, government funds, donor funds). In reality, the process o f setting the sector objectives and the budget estimates in the program budgets is imperfect andapproximate. 121. Currently, most o f the objectives contained in the program budgets have no objective basis. They often reflect ambitions that are unrelated to the budget. Once formulated, these objectives are not adjusted in line with the final appropriations approved by National Assembly, with the exception o f the health sector, where districts develop action plans based on budget appropriations under appropriation law, projections of donor disbursements, andprojections as to the districts' own funds. In reality, the difficulty in setting realistic and readily adjustable objectives is also primarily due to the absence o f sector models that would permit marginal budgeting. This type o f model could be used to create automatic links between the sector objectives and the budgets and to adjust the sector objectives at any time, based on available resources. At the present time, only a few sectors such as education and agriculture have such models, but their integration into the PRSP review process is weak. The agricultural model is a - 40 - module of the Automated Projections Instrument, which is used by DPAM to develop the macroeconomic framework. It is operational and can be usedto set agricultural sector objectives, but often political consideration ovemde model projections. The Education for All (EFA) model is currently being appropriated by education authorities to prepare the education sector MTEF. 122. The arbitration procedures currently focus on procedural issues and are not based on systematic assessments of sector performance. Two committees meet each year to arbitrate budget appropriations: the budget committee chaired bythe Minister o f Finance and Budget and the committee responsible for arbitration o f domestically financed investment (Title 6). Neither of these committees assesses sector performance on a systematic and comprehensive basis when considering proposed appropriations. The budget committee chaired by the Minister o f Finance and Budget places its emphasis on raising the sectoral ministries' awareness o f the basic principles of budget management and ensuringcompliance with MTEFceilings. The discussions focus on the government's means o f operation (buildings, equipment, vehicles, etc.). Similarly, the Title 6 arbitration committee does not analyze the links between project expenditures and their possible impact, as the oversight it provides is purely administrative innature. It examines mainly compliance with MTEF ceilings and makes sure that priority is given to finalizing projects fiom previous years andthat investment proposals are complete (description of activities to be carried out, pro forma invoices available, preliminary estimates and architectural plans available etc.). 123. One o f the advantages of the MTEF for the arbitration process is the potential to use precautionary budget reserves. These are additional resources not allocated to programs during the preparation of the MTEF that should in principle be available for arbitration purposes. The review o f the MTEF by the government should not be a simple formality to inform the cabinet about macroeconomic constraints. It should lead to an allocation of some of these reserves, leavingthe remainder for the national assembly. Such a process would facilitate arbitration since the cabinet could still make additions to the MTEF proposed to the government by technicians. The MTEF2004-06 had significant precautionary reserves (upto CFAF 11billion in2006). 124. The final arbitration of the budget by the Council of Ministers is solely based on the administrative budget. The government reviews the preliminary draft budget crafted by the budget committee and proceeds to make final adjustments before submission ofthe budget to the national assembly. The budget reviewed by the Council o f Ministers is already in the format o f an administrative budget. Thus, there is no opportunity for the govemment to review the programs. 125. The National Assembly intervenes only at the end o f the budget approval process. The Assembly does not review the PRSP, the MTEF, or the program budgets, but it does take into account the MTEF ceilings that are incorporated into the draft budget submitted to it. Its field of action is limited to reviewing and adopting the govemment budget. To analyze the budget, the Assembly organizes itself in the form o f thematic and sectoral committees that prepare reports based on the draft budget received and responses to the budget questionnaire that ask each department to provide additional information on: (a) the status o f past budget execution; (b) sector priorities; and (c) the ministry's chief concerns. Budget proceedings then take place in three stages: subcommittee proceedings that involve ministry technicians; hearings with - 4 1 - department heads, for the purpose of defending budget estimates, thus allowing deputies to hear the political point of view o f government ministers and presidents o f institutions; and finally the proceedings of the arbitration committee. The Assembly does not analyze sector performance based on the budget received. 126. PRSP revision process. The government has not adopted an official schedule for the annual PRSP review. Until now, PRSP progress reports (PR) have not been available prior to development o f the preliminary draft budgets, and therefore the budgets cannot reflect lessons drawn from PRSP implementation during the past year. The joint evaluations organized in the past were conducted during the months o f September 2001 (first PR), September 2002 (second PR) andJuly 2003 (consolidated PR for 2000-02). The purpose o fthe annualPRs is to take stock of PRSP implementation and assess the progress made in meeting poverty reduction objectives. The assessment focuses on a number o f issues: (a) poverty analysis; (b) macro results; (c) reforms that have been instituted; (d) budget execution; (e) physical output; and (f)sector results inrelationto the stated indicators. 127. A systematic evaluation o f the implementation o fprogram budgets could serve as a base for preparing the PR and help better link PRSP and budgeting process. The theoretical framework for the presentation of program budgets encompasses all the elements for an annual assessment that could be used for the PRSP process. As noted above, program budgets list the missions o f each department and describe the sector policy objectives, the main reforms under way inthe sector, the programs and their budgetary costs, the expected physical results and the sector indicators, Inprinciple, the program budget execution reports should constitute a valid basis for a rational assessmento f the progress made inimplementing the PRSP. 128. Linkingprogram budgetingand PRSP reviews could also strengthen the regional PRSP process. For the time being, the regional PRSP lack an operational framework. In the proposal set forth below, the deconcentrated entities become a critical link in the PRSP process. Spatial analysis o f poverty clearly reveals the need to define more specific strategies and regional programs appropriate to the context of each region. Inthe medium term, these regional programs will need to bemanaged under the auspices ofthe future governors ofthe regions. C. INTEGRATINGPRSPPROGRESSREPORTSAND PROGRAMBUDGETS: PROPOSAL A 129. Better integrating the budget and PRSP process would require an acceleration of the production o f the PRSP-PR for it to be available for the budget preparation process. The ideal would be for the PR to be presented early in the year to draw lessons and set guidelines for the future budget that will be reflected in the global MTEF, to be prepared together with the PR. From this perspective, it seems feasible to work under a schedule that would make it possible to hold the annual PRSP review around April 15. This date takes into account the fact that effective participation inthe review requires that the draft PR be made available to participants by March 30. 130. Preparation of the PRSP-PR should rely on program budget execution reports. These reports should include several elements, including a financial execution statement for each - 42 - program and proposed appropriations for the next three years, as well as an assessment of physical output and progress made in relation to sector indicators. Ministerial program budget execution reports should be prepared at the beginning o f the year as input for both the annual PRSP-PR and the MTEF. 131. Ina first step, the program budget execution reports should be prepared on the regional and national level to feed into preliminary updates o f program budgets. Regional execution reports should be made available by the end of January and be reviewed in the context o f the PRSP process by regional workshops by the end o f February. Such a timetable appears feasible given that budget expenditure commitments end on November 20 since the budget execution data pertaining to the additional period of time could be estimated, if necessary. The regional reports should be consolidated into ministerial program budget execution reports by the DEP and DAF,under the auspices ofthe Secretary-General, who shouldpropose to the minister anupdate of the programbudget for the next three years, based on the results achieved and in light of the proposals made by the regions. This indicative programming is o f fimdamental importance to inform the arbitration procedures for the global MTEF andjustify the appropriations made at the level o f the global MTEF. At the same time, the budget execution report would serve as input for the PRSP PR. 132. As a second step, based on the program budget execution reports and the update o f the program budgets proposed by line ministries, the teams responsible for preparing the MTEF. should propose to the government an allocation o f additional resources by sector. The time constraint for the preparation o f the global MTEF is the legally set date for issuing the budget circular, notably the week following May 1, since the MTEF has become an important annex to this document. To avoid further changes to the budget framework at a later date, the macroeconomic framework underlying the MTEF should be consistent with the framework supported under the IMF's PRGF. Ifpossible, revisions to the macroeconomic framework should be discussed in late January and early February between the IMF and the government to ensure that the foundation of the MTEF i s consistent with the understandings under the program. 133. In a third step, the budget circular should be distributed in a timely fashion for the sectoral ministries to align their program budgets or sectoral MTEFs with the global MTEF adopted by the government. For most ministries, central ministry staff (DEP, DAF) could initially prepare the breakdown o f resources among the different programs, in consultation with regional directorates. Central and regional directorate should set objectives to be achieved, both interms of expected physical output and interms of sector results as measured with the help of indicators. This information should then be centralized and consolidated by DAF and DEP to produce the ministry's preliminary draft administrative and program budget that is consistent with the MTEFandbudget circular. 134. In a fourth step, a budget committee that focuses on results targeted under the PRSP would arbitrate the budget proposal on the national level. Budget arbitration proceedings o f line ministries with the MFB should primarily focus on an assessment o f the results achieved, effective use o f appropriations, physical output and the progress made. To improve the results focus o f the budget committee, its operating procedures would need to be revised. The budget committee should be formed o f members who have the experience to analyze the past -43 - performance and the budget proposalso f the sectoral ministries and base their evaluation also on the programbudgets submitted bythe ministries. Staffinchargeofthe PRSPprocesswould need to be associated to these proceedings to ensure that trade-offs for PRSP targets are being adequatelyconsidered. 135. Ina fifth step, the National Assembly would adopt the budget law andprogrambudgets, and line ministries would update their program budgets. Strengthening the results focus of arbitration would also carry over to the presentation o f the budget to the National Assembly by indicating the targets set for by the government and justifjmg allocations during technical discussions inthe parliamentary committees. Program budgetswould be finalized once the final version of the budgetlaw hasbeenapproved. 136. The procedure described above would give operational content to the PRSP monitoring mechanism and integrate it directly with budgetprocedures. The new institutionalmechanism for monitoring the PRSP (established by Decree 0 0 3 ~ ~ S ~ ~ E calls V D E for )two levels of coordination: a regional level, where the coordinating body is the Regional Development Advisory Council (CCRD), and the national level, where the coordinating body is the National Conference. Six thematic and sectoral committees covering the following areas support these bodies: (a) rural development and food security; (b) social sectors; (c) economic infrastructure; (d) institutional reform and decentralization; (e) private sector and competitiveness; and (0 public finance and resource allocation. By directly integrating the budget process with the regional structure o f the PRSP monitoring framework, the proposal described above would flesh out the regional PRSP monitoringframework that is currently purely theoretical, naturally create the necessary links between budgets and objectives at both regional and national levels, and closely integrate the programmatic (MEDEV) and financial (MFB) divisions o f govemment. I).LAYING THE FOUNDATION A SUCCESSFUL REFORM FOR 137. The framework outlined above would need to rely on a substantial strengtheningo f both program budgeting and the MTEF process. These sections presents a few ingredients and suggestions for undertaking these reforms.O 138. Elements of a strategy for improving program budgets. The government could consider a numbero f factors for improving the program budgeting. Develop a strategy paper for instituting program budgets. From the outset it i s important that the government set the contours o f the reform in writing (its objectives, scope and strategies, as well as the human, material and financial resourcesinvolved). 0 Anticipate a long-term reform process. I t is important to realize from the beginning that the reform entails a change of culture and that, as a result, it will take time for the expected results to appear. Inother words, it makes sense to count on a permanent entity to provide impetus to the reform. The members o f this entity should be recruited on the basis of their skills and should not necessarily all come from MFB or MEDEV. In this regard, the recruitment o f a numbero f specialists working inthe sectors i s important. - 44 - e Adopt a global approach. Burkina Faso already has a base o f experience and is not starting from zero. The program budget approach exists, although not all actors follow it rigorously. The entity charged with leadingthe reform should make the necessary support available to each sector at the time o fbudget development. ePlanfor andput inplacethe necessary resources. Itis important to do the programbudget o f the program budget itself to make sure that the activities will be funded. e Place advocates in line ministries. For a reform to succeed, it is often essential that the reformers themselves be highly enthusiastic about their efforts. This enthusiasm makes it possible to deal with discouragement and overcome the difficulties that are sure to appear. e Set up a system of motivation. The government should find incentives to sustain the reformers. e Decentralize budget execution. As much as possible, it is important that local actors be accountable for managingthe budget shares allocated to them. Some of the main issuesto be addressed in line with the reform are to increase the number o f managers, strengthen expust control and institute a solid accounting system. e Ensure the availability o f resources. For the approach to be credible, it is important that the government set in place mechanisms to make sure that resources are available to the entities responsible for implementing the programs. From this perspective, the timely availability o f revenues and donor contributions looms as a critical condition for the initiative to succeed. Improve the strategic allocation o f resources with the assistanceo f the MTEF -45 - Box 1:LessonsLearnedfrom the Use of Program Budgetsin OECD Countries The following conditions must be met to increase the chances o f successfbl implementation o f a reform instituting programbudgets: Those initiating the reform must understand that institutinga performance-based budget system i s a long-term reform andnotjust a limitedoperation. It is important that the need for reform is recognized and felt to be a necessity. There should be unanimous agreement that reform is neededso that managerswill subscribe to the reform. The draft reform should be incorporated into the government's general policy. Ownership o f the reform is necessary not only at the Ministry o f Finance and Budget, but also at the department level o f ministries. All managers o f public programs must assume their share o f responsibility for implementing this reform. It is not solely a matter to be handled by the central entities of MFB. I t is important and necessaryto have a draft reform document that describes: reformprogram implementation strategy teams responsible for the reform communication policy commitment to a sustained effort The main constraints to be addressedare: how to find agents o fchange (reformers) how to provide themwith incentives how to eliminate the constraints with which they are faced The risks o f undermining such a reform mostly revolve around the resistance of certain central entities, in this case the central entities of MFB, for the following reasons: fear o f the unknown the belief that budget decentralization will leadto waste and corruption loss o f certain financial benefits the weak capacity of MFB is often ignored, andthe reformers might not all come fiom MFB Factors that favor success include: encourage people to focus on the objectives rather than the means o f operation and, to that end, make sure that the objectives are published, which can actually motivate people to try to achieve them conduct performance-based evaluations, then reward the good managers who obtain results and punishthe bad managers the system should rely on computers to provide automation opportunities and other advantages such as gains in time and control 139. Global MTEF.The current approach for determiningthe volume o f additional resources that can be earmarked for fundingnew programs or extensions of existing programs needs to be reviewed, Under the current approach, the additional resources are obtained by making a distinction betweenexisting allocations and additional resources, which are distributedwithin the same title among the various priority sectors. This practice limits the government's flexibility for policymaking by excluding the strategic allocations of resources that would better reflect PRSP priorities. The justification for this approach is that the MTEF follows WAEMU convergence criteria and, as regards the split between recurrent and capital spending, the fiscal projections by the M F in order to maintain fiscal sustainability. The government should carefully review this approach, including during its exchanges on fiscal sustainability and resource allocation with development partners, to better align the MTEF allocations with PRSP needs. Of particular concern are the pressures on the wage bill and material and services spending caused by the l6Jack Diamond, Performance Budgeting: Managing the Reform Process,IMFWorking Paper WPf03133. - 46 - implementation of the education and heafth sector strategies that will need to be better reflected in the MTEF allocations and medium-term macroeconomic projections. Otherwise, the infrastructure investments financed with foreign resources may not be used optimally in increasing school enrollment and health services. 140. By limiting the overall fiscal deficit, the MTEF approach itself already provides a framework to limit the "mortgaging" o f future resources in the medium term. If a ministry decides to increase its wage bill and recurrent costs, more o f the future resources will be already committed and less MTEF resources can be distributed in fbture years for new investments or new programmatic initiatives. Table 4 below gives an indication of the volume of resources at stake. The total envelope of additional resources to be distributed comes to more than CFAF 125 billion in 2006. This envelope would decline considerably should the government assign more resources to the wage bill in2004 and2005. 141. Sectoral MTEFs. In order to improve medium-term planning o f program budgets, key ministries are currently buildingmedium-term sectoral MTEFs to give a three-year horizon to program implementation and improve costing. Based on the envelopes set in the global MTEF, the officials o f sectoral ministries will have to further break down these envelopes based on sector priorities. The prior existence o f a clearly defined sector policy will facilitate this work. In 2004, the government has included a budget line to fund the development of missing sector policy andprogram papers. Table 4: Additional ResourcesAvailable under the 2004-2006 MTEF Breakdownof envelopes 2004 20051 2006 Wages 0 0 0 Safety reserves 0.7 0.5 7.2 Equipment 4.9 8.7 12.3 Safety reserves 1.6 1.6 7.9 Current transfers 10 17.9 28 Safety reserves 7.6 2.2 2.5 Investmentsfrom own funds 18.7 32.3 61.a Safety reserves 2.6 4 5.5 Total additional resources 33.E 58.9 102.1 Totalreserves 6.5 8.3 1 7 . 7 Breakdownof HlPC resources Wages 1.7 1.7 1.4 Equipment 9.E 9.2 7.4 Current transfers 1.e 1.7 1.4 Investments 22.2 21.4 17.3 Total 35.2 34 27.5 Total additional resources (MTEF+HIPC) Wages 1.7 1.7 1.4 Equipment 14.4 17.9 19.7 Currenttransfers 11.E 19.6 29.4 Investments 41 53.7 79.1 Total 68.E 92.9 129.6 - 47 - IV. HEALTHSECTORPOLICIESUNDERTHE PRSP: PROGRESSAND CHALLENGES 142. This chapter begins with a description o f major health and nutritionoutcomes inBurkina Faso and of the health system, including the utilization of health services. It then reviews the financing of publicly provided health services, including trends in expenditures by category of expense and by level of the health system. The third section analyzes the strengths and weaknesses of the current budgeting and expenditure process for the sector. The chapter concludes with recommendations for better aligning resource allocations with sector priorities, and for improving the effectiveness and efficiency ofthe budgeting and expenditure process. 143. The main message of this chapter is that while it is important to hrther increase the resources allocated to the health sector, the equity and efficiency of existing health spending (both public and private) must be improved, as well as the quality and financial accessibility of health services, inorder to accelerate progress toward the health objectives laid out inthe PRSP andthe MillenniumDevelopment Goals (MDGs), A. HEALTH SECTORINBURKINAFASO: OUTCOMES, RESOURCES, AND UTILIZATION 144. At independence, Burkina Faso inherited a health system with limited hospital services in urban areas, and little or no modern health care for the rural majority, who either visited traditional healers or self-treated. The 1970s and 1980s saw an expansion of public sector health clinics in the rural areas. Services were provided free of charge, but suffered from severe shortages of medicines and basic supplies as well as qualified personnel. With the launching of the Bamako Initiative in the early 1990s, the govemment introduced cost recovery for health services, and introduced an essential drug policy based on least-cost generic drugs. Community management committees (COGES) were established to manage the proceeds of revolving drug funds, and to strengthen the participation of communities in healthcare. The national drug procurement agency (Centre d'Achat des Medicaments Essentiels Gentriques (CAMEG)) was established in 1992 to procure and distribute generic drugs, with a system of regional anddistrict level pharmaceutical depots. The cost recovery approach significantly improved the availability of drugs in health facilities, including in rural areas, but despite the relatively low prices for consultation and generic drugs, cost emerged as a bamer for the use of health services by the poor (see below). The Ministry o f Health's 10-year health strategy (Plan National de development sanitaire (PNDS), 2001-2010) identifies key priorities and program objectives for the health sector, including increased accessibility of health services, improved quality, increased financial access for the p00r.l~ 145. Disease profile, health indicators and determinants. The Burkina Faso disease profile is similar to that o f its neighbors in the Sahelien countries. Communicable diseases are the main "The eight priorityprograms for the PDNSare: increase geographic coverage ofhealth services; improvethe quality and utilization o f health services; strengthen the fight against communicable and non-communicable diseases; reduce HIV transmission; improve the quality and distribution o f human resources in the sector; increase financial accessibility for the population; increase health sector financing; and strengthen capacity in the sector. - 48 - causes of morbidity and mortality. Malaria alone accounts for one-third o f child deaths, followed by diarrheal disease (12 percent), and acute respiratory infections (11percent). Burkina Faso has one of the highest HIV prevalence rates in West Africa, with an estimated 4 percent of adults infected nationally in 2003, with higher rates in urban areas. HIVlAIDS has contributed to the reemergence of tuberculosis and other opportunistic infections. While vaccination coverage has improved since the late 1990s, the past few years have seen an resurgence of epidemics o f measles and meningitis, with the latter second only to malaria as the leading cause of child deaths inhealth facilities in2002 due to emergence of a new meningitis strain. Table 5: Health Indicatorsin BurkinaFasoComparedto Sub-saharanAfrican Countries Life Infant Child Child Maternal Fertility HIV Doctors Expectancy Mortality * mortality* malnutriti mortality* *** preva- per capita on (weight * lence for age) Burkina 46 105 219 32.7 484 6.8 4 22,985 Faso Afrique 52 92 151 32 822 5.6 8 Cote 55 88 138 24 597 596 10 d'Ivoire Ghana 60 71 110 27 740 590 Guinea 46 122 220 24 880 577 2 Mali 50 120 192 31 577 657 1,s Source: Annuaires statistiques (2002) DEP/Sante`, WorldBank 2003. 146. Most health indicators inBurkina Faso are below the average for sub-Saharan Africa, and lower than might be expected given Burkina's GDP per capita (Table 5). Most key indicators either stagnated or worsened during the 1990s, but with improvements in the past five years (Table 6). Infant mortality (childrenunder age 1) remained essentially unchanged duringthe 1990s, with about 10percent of children dying before age 1, but the 2003 DHS shows a decline to 83 deaths per thousand in2003. ChiId mortality worsened during the 1990s due to an increase of child deaths between ages 1and 5. Child mortality has since recovered (184 deaths per thousand), butjuvenile mortality in2003 is essentially the same as in 1993. Child~aIn~tritzon worsened duringthe 1990s, and a third of children were malnourished (height for age) in 1998. Although 2003 D H S data are not yet available, other surveys suggest that this worsening trend has continued, with 44 percent o f children stunted in 2003.18 The 2003 EnquEtePrioritaire(INSD2003) found that malnutritionrateshaveworsened steadily inthe past decade, from 33 % of childrenstunted in 1993,to 37 % in 1998199,andmorethan 44 % in 2003. Child nutrition datafrom the 2003 DHSare not yet available. - 49 - Fertility rates are among the highest in Africa, but recent data show evidence o f a slow decline. Fertility has declined more rapidly inurbanareas, and is beginning to decline in rural areas. Knowledge and use o f contraceptives has steadily increased in the past decade (Table 7), but rural contraceptive prevalence remains low (only 5 percent of mamed women use a modem method) (DHS 2003). e Maternal mortality is difficult to measure accurately, but survey data suggest an improvement duringthe 1990s. Rates remain unacceptably high(484 deaths per thousand during the 1990.9, however; direct causes include hemorrhage, obstructed labor, infections, and anemia, While most maternal mortality is preventable, inadequate pregnancy monitoring, limited access to emergency obstetrical services, and induced abortions, contribute to unnecessary deaths. Table 6: Trends inMortality, Malnutrition and Fertility Source: EDS 1993, EDS 1998199, EDS 2003; World Bank 2003. * 2003 DHSchildnutritiondataarenot yet available. 147, National averages mask significant variations within the country. Health, nutrition, and reproductive health indicators are worse inrural areas compared with urban areas, and as well as for the poor majority o f the population compared with the wealthiest 20 percent (Table 8). While child mortality remained unchanged during the 1990s inurban areas, it increased significantly in rural areas. Survey data show relatively little difference in child health outcomes among the poorest 80 percent of the population, apparently a reflection o f the poor health status in rural areas, where 85 percent o f the population lives, There are important differences in health status and access to services within urban and rural settings. Large numbers o f poor people live on the outskirts o f major cities, where they have limited access to health and other services. Health and nutrition status also varies by regions; child mortality i s higher North and CenterlSouth compared to the rest ofthe country, and worsened between 1993 and 1998 (Table 8). Have knowledge of Have ever used Are currently using modemcontraceptives modem modem contraceptives contraceptives 1992193 63.3 10.0 4.2 1998199 77.2 12.2 4.8 2003 90.5 19.5 8.6 148. Key questions arising fiom the data are: Why did key health sector outcomes stagnate or worsen during the 1990s, particularly mortality among children aged 1-5, and those living in rural areas? Why has infant mortality declined, but juvenile mortality stagnated in the past - 50- decade? And what factors might explain the differing health outcomes in rural compared to urban areas? 149. Available evidence and the literature on determinants of child and infant mortality suggest the following broad conclusions. First, the modest declines in infant mortality are probably attributable to: improved coverage of prenatal care; slow but gradual progress in assisted births; and possibly declines in fertility. Second, the highlevels and worsening trends in juvenile mortality in the 1990s were probably a result of worsening nutrition status, declining vaccination coverage, and continued high incidence and mortality from malaria. Decreased vaccination coverage, for example, contributed to a resurgence of meningitis epidemics in the latter half o f the 1990s(including a major outbreak in 1996), which was the second to malaria as the leading cause of death among children aged 1-5, While subsequent improvements in vaccination coverage has probably helped reduce child deaths, juvenile death rates remain high because there has been little progress in reducing malaria and child malnutrition. Moreover, the deteriorating trends in the 1990s occurred despite increased per capital health spending, and increased availability o f health infrastructure. As will be argued below, this was a resuit of decreased financial access to health services by the poor, and lack o f equity and targeting of . health expenditures. Table 8: Evolutionof ChildMortalitybyIncomeandResidence Poor 192 214 Poor-medium 196 222 Medium 183 223 Medium-rich 173 212 Rich 109 139 Ensemble 1 180 206 150. Variations inhealth outcomes between urban and rural areas reflect, among other factors, higher incomes on average in urban areas, lower access to health services, clean water, and sanitation in rural areas. Highrates of illiteracy, particularly among women and continuation of traditional practices such as female circumcision or early weaning of infants also play a role. - 51 - 151. AIDS has clearly contributed to adult mortality, but data are limited. In terms of child mortality, mother to child transmission o f HIVIAIDS undoubtedly contributed to increased child deaths, but probably only on the order of 1-2 percentage points (assumes 5 percent prevalence among pregnant women in the late-l990s, and that a third of these women transmitted HIV to their children.) 152. Malnutrition and high fertility contribute both directly and indirectly to the high rates of child and maternal mortality. Studies in Burkina Faso and elsewhere in Africa have shown that malnutrition contributes to between a quarter and one-half of child mortality. Malnourished children get sick about as often as other children, but are more likely to die because of weakened immune systems. Thus malnutrition probably was an important factor in increased child mortality during the 1990s. Numerous studies have also shown that children with vitamin A deficiencies are more likely to die from measles, and women with anemia are more likely to die fi-omhemorrhage duringchildbirth. Highfertility andinadequate birthspacing contributes to low birth weight, and increases the risks of child and maternal mortality. Controlling for other factors, children born less than two years apart in Burkina Faso are a third more likely to die before the age of five than children bornmore than two years apart (World Bank 2003). 153. Community and household factors play an important role inthe high rates of morbidity and mortality, particularly among the poor. Inadequate access to clean water and sanitation, low female literacy, and harsh climate all contribute to poor health and nutrition outcomes. Child malnutrition is a consequence of lack of food security, inappropriate child feeding and weaning practices, as well as repeated illnesses (such as frequent diarrhea andor malaria episodes). Only 18 percent o f infants 0-6 months were exclusively breastfed in2003, which is an important factor explaining continued high rates of diarrhea and infant mortality, despite recent improvements. While the incidence of diarrhea among children is about the same among wealthier and poorer household, one-third of the wealthiest households provide oral rehydration therapy to children with diarrhea, compared to less than 20 percent o f poorest households. Although insecticide treated mosquito bed nets can significantly reduce malaria incidence and mortality, particularly among children, currently only about 8 percent on households have a bednet, and only about 2 percent are treated. 154. Access, utilization, and quality of health services also influence health outcomes, and partly explain differences inhealth outcomes amongregions and between income groups. Access and utilization o f health services is lower in rural areas and among the poor - including for preventive service such as vaccination and prenatal care, for basic primary care services such as treatment of malaria and acute respiratory infections, and for emergency and hospital care, such as emergency obstetrical services. Children of women who undergo at least one antenatal visit duringpregnancy are 20-30 percent less likely to die before the age o f five (World Bank 2003). Prenatal and postnatal care, skilled birth attendance, and access to quality emergency obstetrical services are all key factors influencing maternal mortality. Lower fertility also can reduce risk of maternal death. Correct diagnosis and early treatment of severe malaria and pneumonia are essential to preventing child deaths." l9 A recent survey in the rural health district o f Nouna found that the majority of children with severe malaria at health clinics were not treated accordingto recommended protocol. - 52 - 155. The organization and financing of essential public health and health education services remains inadequate, however, at both the national and community levels. Although mandate o f the COGES includes community health activities, with the possible exception o f vaccination, most have focused on curative services and management o f drug cost recovery hnds (see below). Moreover, among rural women who visit their CSPS, only 20 percent were aware o f the COGES existence (INSD2003), anindication o fbothweak accountability to the community and inadequate engagement incommunity healthpromotion. 156. Health system organization and resources. Traditional medicine continues play an important role in health care and treatment, particularly for the rural majority, for reasons o f culture, cost, and ease of access (see below). The past 10 years also has seen a rapid growth of the private pharmacies and clinics, but these are concentrated in urban areas, with nearly 80 percent of private health clinics and 87 percent o f private pharmacies are concentrated in the major cities of Ouagadougou andBobo Dioulasso. Most public sector doctors inurban areas also work in private clinics, which constitute their major source o f income. Nongovernmental and religious organizations operate health centers and district hospitals (Paul VI and Nanoro), under a framework agreement with the ministry o fhealth. 157. Infrastructure and geographic access to health services. The major source o f modem health care for the rural and poor majority o f Burkina Faso remains the public health system, which is organized into primary, secondary, and tertiary levels. Health districts, primarily through a network of Centre de S a d et de Promotion Sociale (CSPS) provide primary care. The district hospitals (centres medicaux avec antenne chirugircale-CMA) serve as referral centers for the CSPS. Nine regional hospitals (CHR) and three national hospitals (CHN) in Ouagadougou and Bobo Dioulasso provide the second and third levels o f care. Administratively, the ministry o f health is organized into three levels: the central ministry o f health, 13 Regional Health Directorates (DRS), and 55 health districts (districts sanitaires), which are administered by a district health team (equipe cadre de district (ECD)). The University and regional hospitals all enjoy a degree of managerial autonomy, and receive their budgets through a direct transfer from the central government. 158. The number o f government health facilities has increased over the past decade (Table 9), but geographic access remains a constraint in many rural areas. About 58 percent o f the population lives within 5 km o f a health facility, and 19 percent live 10 km or more-with an average distance to a health center of nine kilometers. These ratios conceal differences among regions. Inthe central region, about 70 percent of the population live at least 5 km from a health facility, and only 11 percent live 10 km or more. Inthe eastern health region, these percentages are respectively 41 percent and 32 percent. The disparities also exist among districts; in the health district o f Paul VI, 80 percent of the population lives less than 5 km and only 4 percent more than 15 km. from a health facility, but these ratios are respectively 35 percent et 39 percent inthe healthdistrict o fBogandt. Physical access is determined notjust by distance to the health facility but time, including availability o f roads and transport. Health facilities are inaccessible for many villages during rainy season, and there is a shortage o f reliable transport both at rural health facilities and within villages themselves. Ina recent survey, rural residents estimated that on average the time to reach the nearest health facility was faster on foot (20 minutes) or by donkey cart (25 minutes) than by ambulance (38 minutes) (INSD 2003~). Finally, despite - 53 - increases inthe number o frural health clinics, pressure from influential persons together with the absence o f a "health map" ("carte sanitaire") has contributed to construction o f some new clinics near existing facilities (a national healthmap is currently under development). Table 9: Evolution of HealthInfrastructure Source: Annuaires statistiquesDEPlSante 2002 ;INSD 1993, 159. Number and distribution of health personnel, While the total number o f doctors and nurses has increased, the number o f doctors per capita is still below African average (Table 10). Moreover, there is an over-concentration o f trained medical personnel (particularly doctors and sage-femmes) inthe major cities o f Ouagadougou andBobo Dioulasso. According to Ministry o f Health data, just above three-quarters o f CSPS fulfill the norms interms of personnel (only half of CSPS fulfill the norms inthe northern region). But a recent INSD survey based on a sample o f CSPS found that only 45 percent fulfilled the minimalpersonnel norms in2002, compared to 59 percent in 2001 (INSD 2003). Despite recent efforts to improve staffing levels, it remains difficult to attract doctors, nurses and other paramedical staff to rural zones, due to more difficult living and working conditions, lack o f financial incentives to work in remote areas, lack o f opportunities for private practice in rural areas, and limited employment opportunities for spouses. Finally, although there are no surveys on levels o f absenteeism, anecdotal evidence suggests that many doctors work less than the required hours at public hospitals in favor o f their private practices, and in rural areas, nurses and doctors are often absent for trainings, meetings, or for personal or administrative reasons (WHO 2003). 160. Medicines and supplies. The establishment o f CAMEG in 1992, the adoption of an essential drugs policy, and locally managed revolving drug funds at the health facility and community (COGES) level has helped improve the availability o f medicines at health facilities. While stockouts o f essential drugs remain low at CAMEG, shortcomings indrugmanagement in some districts still lead to stockouts at the district depot or CSPS level (27 percent o f CSPS experienced stockouts in 2002). Non-generic medicines are widely available through private pharmacies in urban areas, though generally at 2-3 times the costs o f generics (private pharmacies have been allowed to sell generics purchased through CAMEG since 2001, which has helped hrther increase the market share o f generics). Most CSPS have the minimal level of medical equipment, but are often constrained interms o ftransport. 161. The average CSPS also has a functioning refiigerator (for vaccine and medicine storage), but inadequate financing o f maintenance and replacement costs have contributed to disruptions inthe vaccine cold chain. The availability and maintenance ofmedical equipment at the referral levels (CMA, CHR, CHN) is more problematic, due to lack o f standardization (with much donated equipment), inadequate allowances for maintenance andplanning for replacement costs. Limited in-country representation o f international medical equipment suppliers further -54- complicates maintenance o f complex equipment. Rigid procurement rules, including monopoly contracts for some crucial supplies (including oxygen gas), has contributed to shortages and stock-outs. The General Directorate o f Infiatructures, Equipment and Maintenance (DGIEM) is in the process o f surveying of the quantity and govement of repair of health equipment, to strengthen planning for maintenance and replacement. Table 10: Evolution of Health Sector Personnel Years Doctors Surgeons1 Pharmacists State Nurses IB Total Percent o f Dentists Midwife (IDE) CSPS fulfilling personnel norms 1990 314 19 112 325 I 800 1239 2809 1995 392 29 69 349 1086 1286 3211 2000 251 17 77 280 1226 138 1989 2002 490 36 60 575 1698 1492 4252 77 162. Utilization of preventive and curative services. Patterns o f health system usage (and non-use) vary according to residence and income levels, as well as by the type o f service. Table 11 shows that for the vast majority o f rural residents, the primary source o f health services is the nearest CSPS. Inurban areas, the pattern is more varied, with wealthier urban residents seeking care primary through private clinics or public hospitals, while sources o f care for the urban poor are more evenly distributed, with urban CSPS, Centre Medicale, and public hospital the most common, but with some use of private practitioners even among the poorer quintiles. Even though the poor are more likely to suffer disease, inrural areas they were three times less likely to have sought care in the past four weeks compared to wealthier residents. The richipoor difference i s even greater in urban areas, with the wealthiest 20 percent of the urban populace visiting health facilities 10times more often thanthe poorest 40percent combined. 163. Vuccinution coverage has improved. Following a decline in the late 1990s, child vaccination coverage has improved in the past five years (Table 12). Dropout rates remain high, however, as evidenced by the slow improvements in DTC3 coverage. And coverage varies considerably among health districts and among richer an# poorer households (31 percent o f the poorest quintile children received a measles vaccine, compared to 70 of the richest fifth (DHS 1998). In2003, 61 percent of urban children were fully vaccinated, compared to 41 percent in rural areas (DHS 2003). Continued outbreaks o f meningitis and measles suggest coverage gaps, the emergence o f a new strain of meningitis, and possibly problems in vaccine efficacy due to difficulties inthe cold chain. National coverage data for 2003 are not yet available, but stockouts of some vaccines (including BCG) occurred earlier in the year due to delays by the Ministry o f Finance inreleasing government funds to purchase vaccines (through UNICEF). - 55 - Table 11: Percentage of Individuafs UsingHealth Services in past Poorest 2nd Richest quintile quintile 3rd quintile 4th quintile quintile Total 1998 2003 1998 2003 1998 2003 1998 2003 1998 2003 1998 2003 CSPS 76.2 42.7 72.1 59.5 73.1 55.4 68.2 64.1 71.6 58.8 71.5 57.9 CMMCM 8.5 4.9 8.6 8.0 10.0 4.5 10.6 5.4 9.8 10.6 9.8 7.0 hospital 1.4 7.1 2.4 2.7 1.0 6.4 3.0 6.0 5.4 10.6 3.1 6.8 Private 0.6 9.5 3.3 11.3 3.4 7.2 2.5 8.5 1.5 5.9 2.3 8.1 Traditionallother 13.3 35.8 13.6 18.5 12.5 26.5 15.6 16.0 11.6 14.1 13.3 20.1 CSPS * * * * 34.7 32.5 37.1 28.7 23.0 20.0 25.9 22.5 CMAKM * * * * 22.1 5.2 33.9 7.5 18.0 12.5 21.5 11.8 Regionallnational * * * * Urban hospital 20.2 28.9 17.4 22.6 25.3 28.8 22.9 21.2 Private * * * 19.5 15.6 7.0 31.8 24.6 33.2 20.8 30.2 TraditionaYother * ** * * 3.5 17.9 4.7 9.4 9.1 5.5 9.0 8.4 1 1 1 I 1 I CSPS ::: I 71.3 40.7 67.9 56.6 67.2 53.5 62.4 57.3 47.1 41.6 58.0 49.4 CMAlCM 10.1 6.1 10.7 9.2 11.9 4.5 15.0 5.8 13.9 11.4 13.2 8.1 Regionallnational hospital 8.0 3.2 4.4 3.9 %: 9.2 5.7 15.5 18.7 9.0 ` 11.8 Private 8.6 3.8 11.9 5.9 3.4 13.0 13.2 18.0 7.8 13.4 Traditionallother 115.0 36.6 14.5 17.9 I I11.1 25.7 113.6 14.7 10.3 10.3 12.0 17.3 1 I I Note: *=the sample size is under 30, and therefore the averagesare not reported. Source: EP 1998 and EP 2003. Table 12: Evolution of Health Indicators Source: Annuaires sfatistiques(2002) DEPISante; Note: This apparent `decline' may be due to changes * inpopulation denominatorsfor calculatingANC 164. Use of other curative and preventive health services has increased more slowly. Utilization of other health services, including consultations, prenatal care, assisted births, have shown slower progress, however, despite the growing number of CSPS being constructed over - 56 - the past decade. The per capita consultation rate remains low at just over 0.2 visits per capita per year, with modest increase to 0.26 new contacts per year in 2002. The percentage of women undergoing at least one prenatal visits has increased gradually to about 60 percent, but much fewer participate in the recommended three visits. (Although national data indicate a decline in antenatal coverage in 2002, available district data suggest a steady increase in the number o f consultations from 2001 to 2003 inmost districts). The percentages o fbirths attended bymedical personnel have also increased slowly, but remain relatively low, particularly in rural areas (32 percent inrural areas, compared to 88 percent urban) (DHS 2003). 165. Hospital services and the referral system. Nationally, in2002 the referral hospitals (CHR and CHN) registered 114,000 new extemal consultations (compared to 1.3 million consultations at the district level), with 55,000 persons hospitalized o f which 3,100 died. Despite the referral roles o f the hospitals, about three-quarters o f those hospitalized came from the city where the hospital was located. Only 13,600 births took place at all the referral hospitals in2002, with 612 maternal deaths (5 percent). At the districtlCMA level, 103,000 persons were hospitalized in 2002, with 7 900 deaths. Overall, C H N tend to be overutilized as primary care facilities, particularly for the middle class, while the CHRs and CMAs are generally underutilized. Because of the varied levels of trained personnel and equipment at the C M A and CHR level, cases are often referred directly to the CHN. Hospital officials report that the referral system works adequately overall, although referred patients pay the same as those who go directly to tertiary hospitals. The counter-referral system is not yet well developed (follow up at the local level for referred patients). Table 13: Incidenceof IIlnesslInjury and Health Care Utilization (past month) I 11998 12003 Poorest Richest Total Poorest Richest quintile2nd 3rd 4th uintile quintile2nd 3rd 4th Total care _. Rural 27.9 28.1 35.5 44.3 52.4 39.0 43.5 56.7 160.8 67.5 71.1 61.6 Urban 23.8 35.3 43.1 44.9 53.5 49.3 58.8 50.9 149.9 71.0 77.6 71.4 Total 27.6 28.6 36.4 44.4 53.0 41.6 44.6 56.2 159.6 68.2 73.8 63.7 166. Patterns and Determinants of health service utilization. A number o f supply and demand factors influence utilization o f health services. These include geographic access, cost (both direct cost and the opportunity cost o f lost time in seeking care), perceived quality of care, mother's education, cultural or personnel preferences. Increased vaccination coverage appears to be a result o f a concerted effort by government and partners at the national and international level, strengthened vaccination outreach by health districts and communities at the local level, and the govemment's decision to provide syringes and vaccination cards free o f charge in 2002, and to make vaccination free o f charge in 2003. Reductions in the prices o f prenatal consultation and - 57 - essential drugs may have contributed to the recent improvements in consultations and prenatal coverage. The opening of new CSPSs could also have contributed, although increased number of CSPSs actually coincided with declining utilization during much o f the 1990s. Table 13 shows that compared to 1998, the poor individuals were less likely to report illness or injuryinthe four weeks, but more likely to seek care in2003. 167. A statistical analysis based on the 1998 DHS estimated the determinants of use of key preventive and curative health services, controlling for a variety o f factors including poverty, mothers' education, urbdrural, religion and ethnicity (World Bank 2003). First, not surprisingly, the higher the cost of the service, the greater the difference between coverage and the richest and poorest households (e.g., delivery at a health center compared to child vaccination). Second, the influence o f poverty, mothers' education, and place of residence was greater for services delivered at fixed health facilities than those that involved "advanced strategies" and community outreach. Thus, for example, controlling for other factors, there was no significant correlation between poverty and the likelihood o f children receiving their first dose o f DTP (which often involves community outreach campaigns), but the poorest quintile o f women were only a third as likely to have had at least one prenatal visit, and one-fourth as likely of delivering in a health facility compared to women in the richest fifth of the population rate. Third, the ethnicity and religion o f the household was significantly correlated to the use o f reproductive health services and vaccination coverage (2-3 times higher among Catholic, Protestants, and Muslims compared to "other religions"), but not the use of curative child health services. Finally, although CSPS are supposed to serve several villages, a recent survey ina rural sector of Yak0 found that while two-third of those in the same village use the CSPS, less than a third of those insurrounding villages do (Konate 2002). 168. A variety of qualitative surveys also have sought to assess the reasons for use or non-use of health services. According to one recent survey, among the women who had not attended prenatal consultations (38 percent), the reasons most often invoked was perceived uselessness of prenatal consultation (35 percent), followed by the cost o f consultations (TNSD 2003). Three quarters of those surveyed also reported using traditional medicine as well, for reasons of proximity, lower cost, and perceived "complementarity" with modern medicine. Among those not visiting health clinics (only 10 percent in this survey), 57 percent cited cost of the primary reason. In terms of perceived quality o f care, half of rural residence surveyed cited lack o f medical material, and 37 percent through that the waiting time was "long" or "very long". The majority of rural health service users surveyed thought that the cost o f services was acceptable, but 35 percent stated that the costs were too high (INSD 2003). When asked for suggestions to improve health services, rural resident cited most frequently improved infrastructure and equipment (19 percent), reduction in costs (14 percent), improved service quality (14 percent), and improvedqualificationso fpersonnel (10 percent). B. HOUSEHOLDHEALTH EXPENDITURE FINANCIALACCESS HEALTH AND TO CARE 169. Household health expenditures. Average household health expenditures increased modestly innominal terms fiom 1998 to 2003, but this was due to increased health spending by the wealthier portion o f the population. Per capita health expenditures for the poorest three-fifths - 58 - o f the population declined innominal terms (Table 14). This decline was even more significant inreal terms -for example, from CFAF630 in 1998to CFAF380 in2003 for the poorest quintile (2000 prices). Declining health spending bythe poorer segments o f the population appear to be a result o f declines inthe cost of primary and preventivehealth care services - particularly among the rural population -and would explain inpart the increased utilizationof these services by the poor, despite continued difficulties with financial access. In 2003, rural households spent an average o f CFAF1980 a month compared to CFAF1800 in 1998, and urban households spent CFAF6400 per month on average in 2003, compared to CFAF 6100 per month in 1998. Urban health spending remains three times higher than inrural areas, but this is due almost entirely to higher spending by the wealthier 40 percent o f the urban population. In both urban and rural areas, better-off households spend considerable more on healththan the poorer segments, bothin absolute terms and as-apercentage o f income (5-6 percent o f income for the wealthiest quintile, compared to just over 1 percent for the poorest). Spending on medicines dominates household health spending (about three-quarters in both urban and rural areas) with medicines accounting for nearly 90 percent o f spending for the poorest 40 percent o fthe population. Table 14: AverageHouseholdOut-of-pocket Paymentson Health, 1998,2003 (CFAFper month,nominalterms) 1998 I 2003 I Source:World Bank Poverty Study (forthcoming), EnquetePrioritaire2003. 170. Evolution of the cost of services. Despite the govemment's longstanding commitment to the principle o f cost sharing under the Bamako Initiative, there is now widespread recognition that financial accessibility has emerged as an important bamer to health service by the poor. To address this situation, the MOH inFebruary 2002 sent a circular notifjhg all health regions that consultations for prenatal and infant care should be provided free o f charge, as well as well as syringes, vaccination cards, and iron supplements for mothers, choloroquine for children and health cards (the direct cost of vaccines hadbeen financed through the PPTEsince 2001). A year later, a similar circular letter instructed regional health directorates to standardize consultation fees for C M A and CSPS. This letter also called for providing prenatal consultation for free, as well as a 10 percent reduction in the price o f three drugs commonly used by children (omoxicilline, paracetamol, and uniprophene for children 0-5 years old. All o f these measures were integrated into the measures o f the PRSP andthe Poverty Reduction Support Credit (PRSC) of the World Bank. - 59 - 171. Equally important, since 2001 CAMEG has steadily reduced the costs and margins on essential drugs, both margins charged at the central level and the margins to be charged by district drug depots (KITS 2002). This has helped further reduce prices of essential drugs in the public sector (although some facilities have continued to maintain the original margins, essentially capturing additional profits). Despite uneven application, these measures have reduced drugprices andthe average costs o fprescriptions (Table 15) Actes Medicaux CSPS 2000 2001 2002 variation InfantVaccination 171 419 232 -44% Vaccination card 53 102 80 -21% Child Consultation 86 86 94 +9% Consultation adult 96 103 113 +lo% Pregnancy monitoring 1315 1225 1007 -18% Normal delivery (eutocique) 986 905 897 -1% . Medicines for normaldelivery 3368 3150 2364 -25% 172. Annual district surveys show that these policy changes have led to reductions inthe costs of essential vaccination and prenatal care, although the average cost of adult and child consultations actually increased slightly (Table 15). Results also show progress in reducing the cost of emergency and surgical services, at least at the CMA level, primary due to reductions in averaged cost of drugs prescribed before and after surgery." Prices still vary considerable among hospitals and CSPS, however, due both to a tradition of autonomy as well as limited enforcement capacity by the Ministry of Health. Costs for a caesarean at a CMA, for example, range from 27,000 F C A in the Nouna to 61 000 CFA in Yako. Even with the price reductions, the costs of these surgical acts represent a substantial portion o f annual income for the poorest households (only 360,000 CFA per year for the poorest one-fifth ofthe populace). 1Cost of medical acts at CMA (including- 1 2000 1 2001 2002 1 Variation I prescriptions) (2000-02) N o m 1delivery (eutocique) 5,438 5,897 4,437 -20% Obstructed delivery (dystocique) 10,220 9,298 5,744 -34% Cesarien 46 654 46 496 38 620 -17% Hemia 38,254 35 612 35 651 -7% 173. Access to emergency care. Public financing of hospitals in principle can help increase access by the poor to high-cost hospital and emergency services and reduce the impoverishing effect of illness, particularly in the absence of social health insurance. Despite growing interest, coverage of community micro-insurance (mutuelles) remains limited, and the low monthly contributions are insufficient to cover high-cost or urgent care. Inthe late 1990~~ government the added a specific budget allocation for emergency care in the financing for all referral hospitals, 2o Accordingto the survey, the costof medicinesprescribed for a cesarean (before and after the procedure)represent on average 75 percentof the total medicalcosts (ENSP 2003). - 60 - but the use of these funds were not comprehensively monitored (some hospitals are currently attempting to establish monitoring systems). Thus the continued relative high cost of services and the lack o f a adequate exemption mechanisms means that despite substantial government subsidies, hospitals largely do not play a social safety net role for the poor. A recent survey found that only 2 percent of patients seeking urgent surgical care (including cesareans) were provided free or subsidized treatment (Bicaba 2003). This lack o f financial access is clearly a contributing factor to maternal mortality. 174. Exemptions systems have limited coverage of the poorest for reasons o f organization, financing, and the widespread belief among health workers that "everyone is poor" and that "someone has to pay, so everyone should pay." In addition to official payments, clients are sometimes asked to make various "unofficial" payments in order to receive emergency care, particularly in urban hospitals. In a recent NGO survey in covering urban areas, half of those interviewed said they had direct experience of corruption in the provision o f government health services, and emergency services were considered by far to have the greatest problem with corruption (46 percent thought emergency services were most corrupt, compared to 6 percent for maternity services andblood banks) (RENLAC 2003). The poor often thus either avoid hospital services entirely or present too late. Late arrival (either during or after difficult deliveries) explains inpart why there were 450 maternal deaths at Yalgado hospital in 2003, equal to about 10 percent of all births at the same hospital (MOH 2002). Althoughthe govemment policy i s that hospitals should provide urgent care in life-threatening situation, anecdotal reports suggests that critically illpatients (including women with problem deliveries) are sometimes kept waiting while the families searchfor money. c. HEALTH FINANCINGANDEXPENDITURES 175. Despite high levels ofpoverty inBurkinaFaso, householdsrepresent the largest source of health financing, followed by govemment financing and donor contributions (Tables 17). Total private health spending by households is estimated at about 62 billion CFA in 2003 (not including direct spending by private companies); the wealthiest fifth of households are responsible for 70 percent of this private spending. The next most importance source i s the govemment budget, which allocated 36 billion CFA for the health sector in 2002, including 10 billion CFA of HPC resources. Donor and external financing was approximately 5.7 billion in 2002, but this is undoubtedly an underestimate, since it does not capture all donor and NGO contributions. 176. Total public sector health financing increased innominal terms between 1998 and 2003, andalso increasedas apercentageofthe nationalbudget(including PPTE). Most ofthis increase occurred since 2001, due primarily to the arrival PPTE resources (Table 18). Total resources for the health sector stagnatedfrom 1999to 2001, however, for severalreasons. First, the closures of two large World Bank-financed health projects in 2001 led to a sharp drop in external project financing for the health sector. Second, despite the transition by the World Bank and EU to supporting the health sector throughthe central budget, the govemment initially respondedto the availability of PPTE resources by decreasingthe percentage of the government budget allocated to health from 11 percent in 2000 to 10.3 percent in 2002. This occurred even though PPTE - 61 - resources were supposed to be additional to the health.*' Third, the execution rate o f PPTE was low in2001 and2002, due to centralized managementofthese funds, among other factors. Table 17: Total Private Spendingon Health, millions of CFAF Poorest 41 42 43 Richest Total 1998 1,233 2,257 3,753 6,575 32,216 46,035 2003 1,043 3,028 4,208 9,052 44,445 61,776 177. The government currently tracks only a portion o f donor h d s for the health sector through the PIP; the authors sought additional financing information from donors, but the estimated external contributions represents an underestimate. Although cost recovery funds typically exceed government financing for goods and services (credits delegues) at district level (see below), there is currently no system inplace to track these funds at the central or regional level. The Ministry o f Health is seeking financing for preparation o f National Health Accounts, which will provide a more comprehensive picture o f public and private health financing and expenditures. Table 18: Public Expenditures for Health, millions of CFAF Total Public Financing Sources: Circuit informatist de la depenses, Loi de finances; *Budgeted amount Notes.Householdspendingestimatedusingdatafrom 1998 and 2003 Enquete Priorilaire (INSD 1998and 2003). Table does not includeestimate ofnon-householdprivatesector spendingfor health. 178. Allocation of health financing by level. Although the Ministry o f Health has made an effort to increase overall funding at the district level, the percentage allocations between central and decentralizedlevels, andbetweenoperational and non-operational arms o fthe Ministry, have not changed substantially since 1998 (Table 19). Spending for tertiary hospitals has only increased by about 50 percent since 1998, while expenditures by regional hospitals have more than doubled. Yet spendingby the central directorates ofMinistryof Healthincreased three-fold, primarily due to increased salary expenditures attributed to the central level. This is due to 21As noted in chapter 2, total external resources stayed about constant after the onset of the HIPC initiative, signifying that the HIPCresources didnot addto the overall resource envelope. - 62 - weaknesses in the personnel management system which attributes all staff on training or awaiting assignment to central directonate, at the exclusion ofpublic health centers). 179. Despite the positive trend in overall public sector financing since 2001, the increased reliance on HIPC funds (which have been centrally managed) and the transition to budget support resulted at least initially in a re-centralization of health expenditures and a reduction in the availability of funds managed at the district level. Following a sharp increase in 1999, the total resources allocated to health districts actually declined in nominal terms through 2001. District financing subsequently than doubled between 2001 and 2d03 due to the addition of PPTE funds, but as will be discussed below, the management o fthese funds was initially highly centralized and often not consistent with district priorities. Since 2001, the Ministry of Health also has made a particular effort to increase delegated appropriations (for goods and services) faster than the overall budget. Although delegated credits represented only 5 percent of total spending in2002, this trend has been greatly appreciated by districts. Table 19: BudgetAllocationsbyLevel, millionsof CFAF (includes governmentbudget, HIPC, externalfinancingincludedinbudget) 1998 1999 2000 2001 2002 2003* I.Non-operational level 2,582 2,637 3,086 4,802 5,020 6,375 Central Directorates 1,561 111,347 1,973 2,943 3,935 5,081 National Hospitals I 3,216 1 2,380 14,152 14,194 I 4,607 14,841 Other National expenses 2,363 1,362 1,198 2,100 2,441 2,694 2. RegionalHospitals 1,608 1,583 2,719 2,813 2,916 3,176 3. Primary Care (CMA et CSPS) 11,230 16,669 17,858 15,025 13,786 27,095 4. Common operational expenses 4,291 4,92 1 2,992 3,297 12,816 5,676 Total Health Expenditures 25,290 29,551 32,005 32,231 41,586 49,858 subsidizedhealthcare for students. 180. Allocation by category of expenditure. The part o f the government's budget allocated to health i s divided amongpersonnel (title 2), goods and services (title 3), current transfers (title 4) and investments (title 6). The percentage allocation of the government budget has changed little among these categories from 1998 to 2003. The percentage of the government budget for investments declined from 18 percent in 1998 to 8 percent in 2002, but with 15 percent budgeted for 2003. Goods and services and transfers stayed relatively constant as a percentage of the budget (around 16 percent and 36 percent respectively). These expenditure categories do not provide an accurate picture of the actual allocations among salaries, non-wage goods and services, and investments, however. Current transfers to public hospitals, training colleges, and related activities include finding for salaries, goods and services, and some investment and - 63 - 0 Vaccination, the Ministry has made a special effort to increasingly finance vaccine purchases through the budget and through HPC, but intemational partners continue to provide the majority of financing. The ministry has prepared a long-termplan for vaccine sustainability, although partners including GAVI, UNICEF, and WHO continue to provide substantial financial and technical support. 0 Malaria. Eventhough malaria is the major killer inthe country, malaria control programs have been systematically underfinanced. There is no budget line for malaria control, and with the exceptiono frelatively small allocations through the Roll Back Malaria program, donor support has been limited. The Global Fund recently approved a USD$7.5 million global fbnd proposal for malaria (for 2004 and 2005), which seeks to improve case management, increase coverage of insecticide-impregnated bednets, and improve coverage of malaria prophylaxis for pregnant women. But complementary financing will be essential ifthe malaria incidence and mortality is to be significantly reduced, including the assurancethat funds are available to finance malaria activities at district level. e Reproductive health. Partners had financed procurement o f contraceptives until recent years (first by W P A , then through the Bank's PPLS project), but with the end o f these extemal funds, there is a financing crisis for contraceptives, with risks o f impending stockouts. The majority o f fbnding for DSF programs in reducing matemal mortality, adolescent healthcomes &om extemal partners including UNICEF, WHO and others. 0 ~ ~ t r i t jThe profile of nutrition in the Ministry has been increased somewhat by the u ~ . elevating the former department of nutrition to a Directorate o f Nutrition. Most activities and fundingto date focus on micronutrients, including vitamin A. While important, more needs to be done to combat protein-energy malnutrition, including through community- level interventions. 0 HIY/AIDS. Financing for HIViAIDS prevention and treatment has increased rapidly in the past decade, reaching an estimated 14 billion CFA in 2003 - including 7.4 billion in grant financing and 3.6 billion fkom the World Bank M A P credit. While initially entirely financed by extemal partners, the government's contributionhas increased to an estimate 1.1billion in2003, although this represents above half of estimated household spending on HIVIAIDS care and treatment.23The initiation of anti-retroviral treatment promises to reduce AIDS related mortality, but remains costly despite decliningprices for ARVs. 183. Resource allocation and trends in health districts. Total resources vary considerably among districts, and allocations are inequitable (Figure 27). First, as noted earlier, personnel and infrastructure make up the bulk o f government resources allocated to districts, but these are not equitably distributed, and are not directly managed by districts. Second, the allocation o f the delegate appropriations (goods and services) among health districts is inequitable and based and not basedon a clear criteria (population, poverty, or specific district needs), see Figure 28. Third, proceeds from cost recovery ("fonds propres") also vary several fold among districts, often with those inhigher average incomes or greater availability o f health facilities and personnel able to generate more income. Fourth, the number o f donor partners and level o f donor financing also varies substantially, but are not tracked systematically at the central ministry. ~ 23 Source: UNAIDSISIDALAC, ((Niveau et Flux des Ressources Allouees a la Lutte contre le VIHiSIDAn,, summary presentationofHIVlAIDS nationalaccounts, Ouagadougou, April 2003. -65 - Table 20: Financing of Vaccination, Malaria, Reproductive Health, and Nutrition Programs in 2003 (billions CFAF) Government Partners Total Vaccination Malaria Reproductive Health Nutrition HIVlAIDS 1.1 11.0 112.1 184. Current information systems do not allow systematic tracking o f overall resource available to districts - including human resources, investment, partner contributions, and cost recovery, But the average salary allocation for each district would be around 100 million CFA, makingthis the major sourceo f government financing for districts. Interms o fresource directly available for the functioning of health districts, the most important source o f financing for most districts is cost recovery, followedby donor financing, and finally the delegated appropriations. The averagedelegatedcredit for 2002 was about 33 millionCFA in2002, which increasedto 36 million in 2003. The average estimated contribution of partners to health districts was CFAF 32,800 in2002, a significant drop from CFAF 45,000 in2001 (attributed primarilyto the close in 2001 o f the World Bank's PDSNand PPLS projects) (INSD 2003). These findings are consistent withthe (incomplete) data gatheredfrom health district actionpla+ns,which show a drop ofdonor resourcesat district level in2002, but with signs ofrecovery in2003. Total revenues from cost recovery (including consultation fees and sale o f drugs) vary considerably among health districts, and data are incomplete and not recorded in a standardized manner. Among districts surveyedby INSD, the average cost recovery revenue for the CMA was 39 million CFA in2002 (compared to 33 million CFA in2001), varying from 117 million CFA at the CMA de Paul VI to only 7.6 million CFA at the CMA in Yak0 district. The average revenue of a COGES was approximately 6.6 million CFA in2002, which would suggest that'the COGES inan average district generated I30million CFA in2002. - 66 - I;xf-tfc21f Table 21: Allocation ofdelegated appropriationswithin districts (thousands of CFA) 2000 2001 2002 Average budget transfer to districts 14,506 16,860 32,977 District management (ECD) 25% 20% 27% CMA 33% 34% 16% CSPS 42% 46% 57% 186. Second, the 2001 PER also highlighted importanceof increasing financial allocations and planning for maintenance. The total allocation o f delegated credits devoted to maintenance has increased in the past three years, fi-om a total of only 161 million CFA in 1999 to nearly 460 million in 2003. Lnaddition, beginning in2001, districts have been allowed to use their delegated credits for maintenance of buildings, which previously hadbeen managed centrally (Table 22). Table 22: Total DelegatedAppropriations Allocated for Maintenance (millions of CFAF) Buildings Materialltransport Other materiel Total 2003 129 254 74 458 2001 103 130 26 260 1999 __ ' 133 30 161 187. Third, although districts are encouraged to develop an integrated plan for financinghealth services, the deconcentrated appropriations are mostly used to finance routine functioning, while partner resources are relied on for financing programs. In effect, this means that priority programs that lack specific partners are often simply dropped from district plans. The relative absence of specific malaria activities inthe 2003 various district action plans was striking inthis regard, and was attributed to the limited extemal financing for malaria. Although this situation will change in 2004 with the arrival o f Global Fond moneys, it highlights the need to ensure adequate budgetary funding for priority activities at the district level. 188. Finally, at the level o f CSPS, the sale and purchase o f medicines dominate the finances and activities o f the CSPS (see Table 23). Notably absent are expenditures for community health education or other public health activities. Moreover, although CSPS are allowed a 50 percent markup on the sale of generic drugs, on average about 30 percent o f total revenue (1.5 million) remained unspent, in addition to an equal amount carried over from the previous year. There are no systemically maintained records at the central or regional level regarding total revenues, expenditures, or balances inthe accounts of the CSPS. A survey in 2002 found that COGES had on average about 1.5 million CFA of idle resources in their accounts. While the situation varies among COGES, evidence from both financial data and the recent global audits o f the health districts suggests that (a) in some COGES money is simply accumulating in the local bank account; and (b) in other, either money, medicines, or both are being diverted to unauthorized uses. - 68 - Table 23: Average Income and Expendituresof CSPS by Category, 2002 Source: INSD2003. 189. These lost or unused resources are substantial when added up nationally - an average unused balance o f 1.5 million times 1051 COGES equals nearly 1.6 billion CFA annually, which i s only somewhat less than the total delegated credits allocated to health districts in 2002.24This finding highlights the need both for strengthening the directives for the COGES, to emphasize their obligations for community public health activities andor for subsidizing priority or emergency services, strengthened accountability and oversight, and further training for COGES members. While ministry officials and partners initially expressed concem that free vaccinations and antenatal care, and reduced margins on certain drugs would undermine the financial sustainability o f the COGES, available data for 2003 show that most districts and COGES continuedto runa surplus. 190. "Purchasing" of health services and provider payments system. Individual households, government, or private sector entities can purchase curative or preventive health services. In the private sector, most payments are done on a fee-for-service basis, including the purchase o f medicines. This contributes to a bias toward over prescription o f medicines, including higher cost non-generic drugs. Private insurance, which limited to the small formal sector, is primarily based on paying fees for services. 191. Within the public sector, most government and donor financing is used to 44purchase" services provided by government hospitals and health districts. Beginning in 1999, the Ministry of Health introduced a performance-based contracting system (initially piloted through the PDSNproject), inwhich district were given money directly inlocal bank accounts, and prepared annual action plans based on a limited number o f key indicators, Although there are neither rewards or sanctions for good or poor performance, it was the first time districts prepared annual plans based on agreed indicators, and were given liquid financing to implement the plans. The 24 A review of financial reports from health districts also confirms that unspent resources from cost recovery is often nearly equal to the deconcentrated allocations. Conversely, a number of districts consistently report that expenditures from own are precisely equal to revenues, which does not seem credible. The recent global audit o f health districts underlined a variety o f shortcomings in accountability for cost recovery funds at the district and COGES level. - 69 - system was initially piloted using hnds from the World Bank Health and Nutrition project, and is being carried forward through the Dutch-sponsored PADS project. In the past two years, districts have been encouragedto incorporate a wider rangeo f indicators, and to integrate donor, government, and cost recovery funds into their plans. While the allocation o f funds among districts through the PADS project is based on standard criteria (district population, number of health centers, sound supervision mechanism, poverty, etc.), there is no criterion - either population, poverty, disease burden- for government budgetallocations among districts, 192. Conversely, while public hospitals enjoy a degree o f administrative autonomy with regard to management o f financial resources, there is no performance contract or perfomance monitoring agreement between the central ministries and hospitals. Transfers from central government to referral hospitals vary significantly among referral hospitals, but are not based on anyclear criteria. Government hospitalsdistribute aproportion ofproceedsfrom cost recovery to health personnel (e.g., about 10 percent on average, but up to 20 percent at many regional hospitals), but this creates incentives to maximize cost recovery, and disincentives to provide exemptions to the poor. Faculty from the medical school representsthe majority o f specialists at the referral hospitals, but they are entirely paid by the Ministry of higher education (MESSRS), and receive nothing directly for their hospital service. This reduces accountability to hospital administrators and contributing to highrates o f absenteeism among senior specialists, reducing the quality ofhospital care. 193. Hospitals also rely substantially on cost recovery for their routine operation, with total receipts for all hospitals equal to 1.9 billion CFA in 2002, varying from 731 million CGA at Yalgado national hospital inOuagadougou, to 37 million at the regional hospital inDori. Pricing and cost recovery policies differ substantially among hospitals, as does the applications of government regulations require free or subsidized care for civil servants, students, and indigents (the latter category not being clearly defined by current texts). Hospitals receive a line item in their global budget allocation for emergency care ("secours particuliurs"), equal to 230 million at Yalgado andranging from 15-30 million CFA at regional hospitals in 2003, but accountability for these funds has been loosely monitored. A recent study found that a small percentage of indigents benefited from these funds, and many who received subsidies were not indigents (Abica et a1 2003). Some hospitals are in the process o f establishing tracking systems, but the entire institutional and financial framework for urgent care and exemptions needs to be revised, 194. Health worker payments and motivation. Government health workers are paid based on seniority, with limited rewards or sanctions based on performance, or for working in underservedareas. This contributes to the lack of personnel inunderserved areas, and reduces the incentives to provide quality care inthe public sector. D. GOVERNMENT BUDGETALLOCATIONAND EXECUTION PROCESSINTHE MINISTRY HEALTHOF 195. This section analyzes the budget allocation and execution process to identify problems and identify solutions. The current process for the allocation and execution of health sector resourcesmakes it difficult ensure that health sector expenditures are aligned with the objectives - 70 - PRSP and the national health strategy (PNDS), but reforms currently underway, including preparation o f sectoral MTEF, could help address this issue. 196. The Budgetprocess. Although health is a considered a priority sector inthe PRSP, this fact has not been fully reflected in the budgeting process.25While the Ministry o f Health is involved inoverall MTEFpreparations (the DAF andthe DEPparticipate), the MOHhas limited influence on the process. Due to the absence o f an institutional mechanism for systematically updating the list of programs and annually updating their cost, the MOH is not able to present a strong argument regarding the budgetary requirements to achieve health objectives under the PRSP, or to provide detailed information on the number and the cost o f new programs to be proposed. 197. In 2001, the Ministry o f Health began preparing annual 3-year program budgets and a three-year investment plan, consistent with recommendations from the 2000 public expenditure review. But this program budgeting process has had limited influence on actual budget allocations for several reasons. First, the Ministry often receives its indicative allocation late in the budget cycle, thus central departments sometimes have only a week or two to prepare their requests. The tendency is to overestimate needs, anticipating cuts from the central level, underminingthe credibility o f the process and contributing to a lack o fprioritization. Second, the Ministry has not establish a rigorous process for estimating the cost o f new programs, or more importantly, the cost o f achieving objectives stated under the PNDS and PRSP. The program budget i s thus more a summary o f proposed activities than a costing o fpriority actions to achieve sector objectives. Third, there i s no system for tracking budget execution according to the program categories in the program budget or the PRSP. Once the budget allocations among expenditure categories are determined for the annual budget law, the program budget is essentially set aside and not referred to again. Thus for the health round table heald in April 2003, the Ministry was not able provide partners with estimates regarding expenditures according to PNDS programs. Moreover, the health sector budget presented in the 2004 PRSP Action Plan was not consistent with that in the 2004-2006 program budget. Finally, funds from cost recovery and partner resources are not adequately integrated into the budgeting process, nor is their execution adequately monitored. 198. At present, responsibilities for preparing the preliminary draft budget o f the Ministry o f Health are split among a number of different structures. DAF is responsible for the government budget (Titles 3, 4, and 6), deconcentrated appropriations, and HIPC resources. Wage forecasts are now entrusted to the HumanResources Directorate (DRH).DEP is responsible for planning and monitoring project expenditure financed by external partners. The health districts' own resources from cost recovery and direct financing from technical and financial partners (PTFs) are planned and managed directly by the districts. Information on this financing is available in the annual action plans prepared by all districts, but is not centralized or routinely monitored by DAFor DEP. 25 Sectoral ministries begin preparation of draft budgetsuponreceipt of the BudgetCirculargiving instructionsto be followed by all departments, which includes an annex giving the overall Medium Term ExpenditureFramework (MTEF) setting forth the ceilings for the various ministriesand institutions. These ceilings pertainboth to approved items and to the additional resources that can bedevotedto new measures(e.g., new programsand additional staff). - 71 - 199. In2004, DAF introduced for the first time a draft circular within the ministrybefore issuing the budget circular for preparation of the 2005 budget. Health's circular is basedon the indicative ceilings set forth in the 2005-2007 MTEF, which are already known, and proposes a distribution o f the MTEF ceiling among the various ministry's structures (chapter level in the nomenclature), using the following criteria: (a) deconcentrated structures: (+ 10 percent); @) central structures (+ 5 percent). I t is up to each structure to allocate these amounts among the various expenditure headings in accordance with its own priorities. Frequently, the structures mechanistically apply the authorized increase to every budget line, without really engaging in tradeoffs, and approach DAF for additional appropriations to finance new activities. The health budget circular emphasizes this issue, and it i s likely that budgets will better reflect priorities of different ministerial structuresthis year. 200. The M O H is among the ministries that are reinforcing the programmatic approach through a sectoral MTEF. The government recently began asking priority ministries to prepare sectoral MTEF, which would indicate allocations among the various ministry structures and programs, based on envelopes establishedinthe global MTEF. The Ministry o f Basic Education (MEBA) prepared an "experimental" sectoral MTEF in 2003. The Ministry of Health has embarkedon the MTEF process inlate 2003, establishing a working group ledby the DAF with representatives from DEP and the public health directorates, and drawing on experiences with health MTEFs in neighboring countries. The goal i s to develop a tool that will allow both planning and monitoring o f all health sector resources (budget, partners, cost recovery), and will strengthen the links between resource allocations and PRSP and PNDS objectives. The Ministry i s also considering applying a recently development epidemiological model ("Marginal Budgeting for Bottlenecks") that assists in identifying and estimating the costs o f key interventions neededto achieve health MDGs. 201. The MTEF has the potential to become the basis for results-based financing system for the sector, encompassing donor and cost recovery resources as well as the national budget. Further effort is necessary to complete this first MTEF, however, (completing data gaps, finalizing nomenclature, estimating medium-term financing needs to reach key indicators), and to ensure full "buy-in" from Ministry officials and partners, so that it becomes the basis for resource planning and monitoring - not simply another parallel exercise. The process must then be institutionalized within the Ministry,along with accompanying reforms to strengthen systems for health infonnationand budget data, aligning with the district planning process, and so on. 202. PIanningand execution of personnelexpenditure. InJuly 2001, the Ministry ofHealth had a staff o f 9,052 employees, o fwhich 657 work inadministration. The public service ministry controls nearly all matters related to personnel, however. There currently is no consolidated statement o f the wages paid to the personnel working on behalf o f the Ministry o f Health, nor a personnel system that monitors where employees are located. Wages are managed by a number of different structures, none o f which centralizes the information in order to produce a complete and comprehensive statement o f wage costs, or the number and categories o f staff by district or health facility. 203. There are two groups o f staff: those paidthrough the government payroll process, and the personnelo f the administrative public entities (EPAs), who are paid directly by those structures. - 72 - Wage projections for the latter category are prepared by the EPAs, which receive subsidies from the government budget. The forecasts for the staff covered bythe payroll process are prepared by Directorate of Human Resources (DRH), which did not assume this task until 2003 (the DRHis a recently created directorate, previously located under the Budget directorate). To do so, it had to gather information from the information technology services directorate of the Ministry of Finance and Budget inorder to access the payroll file. Inother words, it does not yet have an up- to-date personnel roster. DRH has obtained personnel management software adapted to the ministry's needs, but the requisite hardware has not yet been purchased and installed. A personnel census is planned this year for purposes o f buildingthe basic personnel roster -but the personnel software must be fully function ifthese data are to be updated. 204. Recruitment policy. Burkina Faso currently trains health staff in special schools. These schools train licensed nurses, registered nurses, laboratory technicians, and other paramedical personnel. Health personnel are recruited into the Ministry of Health throughtwo methods. First, a limited number of students are recruited directly into the civil service based on competitive entry examinations for the health training schools (ENSPs), based on positions authorized by the civil service ministry. Tuition and board for these students i s fully funded by the government, and these student-employees are automatically incorporated into the civil service upon completion of their training. Other students can still enroll, but must pay. For non-"bursary" students, the civil service ministry sponsors an annual civil service entrance exam, from which the health ministry recruits a limited number o f additional personnel, depending on available budget. Many of these are recruited as contractual staff (using HIPC funds), rather than as civil servants. Physicians attend the Medical School at the University of Ouagadougou, and are recruited in accordance with the ceiling available for new measures, which is indicated by the MTEF. 205. There are several weaknesses inthe current training and recruiting system. First, there is a structural gap between the ministry's staffing requirements and the authorized recruitment ceilings. Despite the hiring o f 1,180 contractual employees, known as HIPC personnel, unmet personnel needs remain sizable. In2003, 670 positions were provided out o f a total expressed need for 1,578. Second, there is no real coordination between humanresource planningby the in human resource directorate and enrollment decisions by the health training schools, Thus, although a key objective of the national health strategy i s to ensure that 90 percent o f health centers fulfill personnel norms, there is no system to ensure training and recruitment are adequate to meet this target, Third, the schools are currently overcrowded and understaffed, particularly the regional training schools. Thus many students graduate, for example, from the nurse midwife program without havingperformed enough defivery. 206. ~ ~ ~ t r i ~of~ healthnpersonnel. As noted above, inequitable distribution of health t i o personnel is a long-standing problem in Burkina Faso. The distribution of student-employees i s carried out based on the needs expressed by the regions and the personnel available, with positions to be filled ("new measures") designated in advance. New staff is allocated among health regions, and the regional health directorates are responsible for distributing staff within the region. The HIPC personnel-who are recruited as contractual staff-are assigned automatically to the regions they designated when registering for the recruitment competition. - 73 - 207. There are several problems with the current system. The first is the willingness of health personnel to relocate to remote areas. Indeed, physicians often turn down job offers ifthe position to be filled is located in a rural area. In 2002, for example, out of about fifty positions opened, only 30 were staffed even though an adequate number of trained physicians entered the market, Second, although staff recruited to a given region and are expected to remain there for a least three years, in principle staff frequently succeed in being transferred back to urban areas despite not having fulfilled the minimum. Even those who remain the minimum requirement frequently spend the rest of their careers inurban facilities. Third, and related, the "budget posts" are associated with the civil servants themselves rather than the health facility or district (with the exception of HIPC-recruited staff), so that staff who transfer out o f a rural facility take their salaries with them. The centralization of recruitment results in many other problems with transporting newly assigned staff to their duty posts. As the DRSs have no budget line for this . purpose, it may take 6 to 7 months before staffreports to their posts. 208. Planning and execution of non-wage recurrent expenditures. A persistent challenge for the health sector - at central and particularly decentralized levels - is assuring the timely availability of hnds for functioning of curative and preventive health services. Services cannot be delivered if funds arrive late; cumbersome, centralized procedures contribute to delays in implementing activities and to reduced budget execution. While the Ministry of Health has taken some steps to decentralize budget management.and execution, more remains to be done if the budget is to become an efficient source o f fimdingfor health service delivery. 209, For centra1 structures, non-wage recurrent spending is executed by DAF, which uses two complementary expenditure procedures to do so. The first i s through the "normal" procurement and expenditure procedures common to most francophone budget systems, namely, passage throughthe stages of commitment, validation, payment authorization, payment order issue, and payment itself. At the centra1 level, this permits grouped orders, with potentials for cost savings. The DAF has introduced transparent management of these grouped orders by adopting the principle of the advance allocation, as the contract is being drawn up, of the products ordered by beneficiaries for delivery on site. The new public procurement procedures, which require a purchase order for any purchase smaller than CFAF 1 million, has complicated management of Title 3 expenditure under the responsibility of DAF, however. The second method is through a cash advance account (regie dhvance) that the DAF manages with 116 of the Title 3 appropriations. These resources make it possible to cover small outlays by the services pending the completion ofgrouped orders. The account is replenished once the expenses arejustified. 210. Planning and execution of current transfers. Title 4 expenditure mostly consists of operating subsidies transferred to the EPAs (including national and regional hospitals and the national public health training school) and national centers (CNRFP, CNAOB, CNLCC, CNRSN). This budget line also includes contributions to international organizations such as the WHO (the amounts o f which are set in the relevant conventions), student allowances, and the resources transferred to certain funds (FONALEP). Their preliminary draft budgets are centralized by DAF. Only the budgets of the project management units have not yet been centralized. The subsidy i s disbursed by releasing h n d s quarterly, and an expenditure breakdown i s provided (wage portion, investment portion). - 74 - 211. Difficulties with the current budgeting and execution process include the following. First, EPAs prepare annual budgets, and many have initiated preparation o f three-year program budget.26 But the process for presenting and defending these budgets is relatively pro-forma, with relatively little influence on their budget allocation. As noted above, there is no explicit performance contract between the government and EPAs, or criteria for determining allocations among hospitals. Second, while the EPA subsidies are transferred by releasing funds quarterly, the first tranche often made available inMarch, with frequent problems related to preparationof the documentationrequiredto unfreeze the subsequenttranches." Third, budget, information and accounting systems are insufficiently developed. Implementation o f the EPA accounting system is experiencing problems relating to completion o f the inventories o f available equipment. While these inventories are beginning, there is currently no system for amortizing equipment, or for forecasting recurrent costs. Finally, DAF does not prepare summary statements that would be tantamount to a consolidated statement o f the operating expenditure o f all structures in the ministry, including the operating expenditure o f the EPAs and the project management units reporting to the ministry. In addition, some major expenses, including water and electricity, are funded centrally and not charged to hospital budgets. This reduces transparency and incentives for conservation. 212. Planningand executionof investments.Health sector investments are included intitle 6 o f the budget (title 5 of the newly introduced WAEMU nomenclature). This category includes three major spending categories, namely, expenditure financed against own resources (known as financing by the govemment alone), expenditure financed in the form o f subsidies (with or without counterpart funding), and expenditure financed by borrowing (also with or without counterpart funding). Expenditure financed against own resources: There are two categories o f expenditure financed against own resources: (a) spending on construction or the purchase o f materials for the administrations or for the structures operating inthe field (new measures); and (b) national counterpart spending intended to cover the operating budgets o f the coordinating bodies for projects financed by loans or grants. The first type o f expenditure tends to be implemented in accordance with public procurement procedures. For construction, DAF enters into a delegated management contract with Faso Baara, the national executing agency. For purchases o f materials, DAF has introduced a new procedure consisting inpreparing contracts in which the beneficiary services that will receive deliveries on site are identified. EPA appropriations devoted to construction or to the purchase o f materials are released to them, and they execute the expenditure themselves in accordance with the applicable procedures. For the national counterpart funding to finance the operating budgets o f project coordination units, DAF initiates releases of funds intheir favor. Expenditure financed by subsidies or byborrowing: This spending is carried out directly by the project units in accordance with the procedures required by the various donors. A disbursement request processed through DGCOOP is generally l6By way of illustration, when the Banfora CHR, which has EPA status, receivesthe budget circular (at the start of the year), it preparesthe preliminary draft budget (estimating personnelexpenditure, supply expenditure, and investments). This preliminary draft draws a distinction between the wages o f contractual staff and of govemment civil servants, and is filed with the ministry by April. The process of preparing the CHR's preliminary draft budget is carried out through a budget committee established to consolidatethe needs expressed by the various services as well as the revenueforecasts. *'In fact, the resourcesmade available to the EPAs are also spent in accordance with the normal expenditure procedures, and quite often certain spendingis finalized late. There is actually nojustification for this unwieldiness becausethe EPAs are directly accountableto the Audit Offce for their own management. - 75 - addressed to the donor. There are often significant delays in such expenditure because the donors' non-objection and other procedures. 213. Planning and execution of HIPC-related expenditure. HPC resources are used to finance all categories of health-related expenditure (wages, operating budget, investments). At the outset, when HIPC resources were first available in 2000, the proposed allocations provided by the health sector were approximate and not sufficiently targeted on the priority needs of the beneficiary services. The special account mechanism established for managing funds provided in principle for some degree of flexibility with resources, in that it was possible to carry balances forward from one fiscal year to the next. The Ministry of Finance and Budget curtailed this flexibility because of the difficulties experienced with mobilizing debt relief. For example, for the health sector, of the CFAF 10 billion to be camed forward to 2005, only CFAF 4 billion actually was. Moreover, the special account mechanism has created to an over-centralized planningand execution process, andcontributed to shortfalls inHPCexecution (Table 24). Table24: HIPC HealthAllocationsandExpenditures (millions ofCFAF) I I 2001 I 2002 I 2003 214. District staff acknowledges that they have received more resources, primarily equipment and new staff, since the HIPC Initiative came into force. But owing to the absence of consultations, the equipment received more often than note didnot match the areas of need. The districts were informed o f the quantities delivered, but no prices are specified. They consequently have no idea o f the financial cost corresponding to these items. Since 2003, beneficiary services have been asked to express priorities, but the request anived too late in the district planning process. Plans to integrated HIPC into the government budget would be welcome, as would set aside a portion of HIPC funds to be delegated directly to healthdistricts. 215. Participation of deconcentrated structures in the budget process. The operational structures at the deconcentrated level include: (a) the regional directorates, which fulfill a coordination function; supervision and technical support to the districts (b) the national or regional hospital centers (CHNs and CHRs), which have the status of health public entities (EPHs) and have managerial autonomy; and (c) the districts, which supervise the medical centers (some with andsome without surgical facilities) and the health and promotion centers. 216. The regional health directorates (DRSs), the districts, andthe health and social promotion centers (CSPSs) have limited involvement in the government budget process beyond estimating wage costs at the start of each year. They do not participate on any budget commissions, and have relatively little influenceonthe budget allocations they receive from the central level. - 76 - 217. Since the late 1990s, each health district and regional health directorate has prepared an annual action planon the basis o fguidelines established byDEP, which are drafted inNovember and presented at regional workshops in January of each year. These plans initially focused primarily on planning for donor resources, but have been gradually broadened to include planning for delegated budget credits and cost recovery funds, although not consistently so. The plans do not include costs o f personnel or govemment-~nancedinvestment expenditures. The action plans are discussed with regional and central ministry officials as well as partners, but are prepared after the govemment budget preparation process has been completed. For the most part, the district plans are a consolidation o f plans by the district hospital and health clinics (CSPS) micro-plans, together with activities to be financed by external partners. However, the DRSs, the districts, and the CSPSs are unaware of how the tradeoffs are made at the central level, and do not know the total share o fthe government budget they receive. 218. Deconcentrated appropriations, The use o f deconcentrated appropriations, through which health regions and districts are given their budgets for non-wage recurrent expenses (Title 3), is well advanced in the health area. Health districts and regions are responsible for committing these appropriations, which are made available to them in the form of notices o f appropriation awards (AOCs). Expenditure execution then follows the normal procedure, o f commitment, validation, payment authorization, payment order issue, and payment. With the gradual deconcentration o f payment authorization to the regional level, all these expenditures will ultimately be devolvedto the local levelinthe very near future. 219. The deconcentration process currently has a number o f drawbacks associated with the excessively centralized nature o f the funding and the lack o f information on the appropriations delegated to the districts by the DRSs. First, notices o f appropriation awards (AOCs) for the delegated appropriations have consistently arrived late (generally between April and June). District staff refers to the period between January and arrival o f the funds as the "dead" period, when they are unable to carry out many activities on their work programsdue to unavailability o f funds. While the DAF has made an effort to reduce the time lag (most funds for 2004 arrivedby March), the procedures remain unduly cumbersome-for example, both the health budget director and the budget director for the Ministry o f Finance must each individuallysign on every delegated budget line (19 line items per district) for all 55 health districts and 13 health regions: or over 1000 signatures each, Second, because the DRS sometimes receive supplies for the districts, the distribution may not take account o f purchases made by the district using the delegated appropriations. Third, the "no~-mal" procedures for using these appropriations remain difficult to implement for purchases valued at more than CFAF 1,000,000. In particular, utilization o f a "purchase letter" proves difficult owing to the extended unavailability o f some members o f the provincial procurement commission (Secretary General of the province in his capacity as commission chair, or sometimes the Financial Controller, etc.). Proposals have been made that would limit the procedures to requesting prices to be submitted to the Financial Controller. 220. Use of district resources. For items purchased through the delegated credits, the Chief Physician o f the District (MCD) draws up an allocation statement for the equipment purchased and then distributes it. The equipment or products are handed over at the monthly meetings: (a) patient food i s distributed to the CHR and the CM; (b) protective equipment (smocks, - 77 - disinfectants, rubber gloves) is distributed in accordance with an allocation statement between the CSPS andthe district offices; (c) ofice supplies are retained at the district; (d) spare parts for 2-wheeled vehicles (motorcycles) for which the CSPSs' needs are centralized with the MCD, who places the orders and determines the table for allocation, maintenance, and repair; (e) fuel up to the amount o f CFAF 200,000 per quarter for all districts, despiie the differences in the travel requirements o f individual districts; and (f) medicines. 221. Cash advance accounts (regies d'avance): In order to impart greater flexibility to executing some delegated appropriation lines, cash advance accounts were established for all health regions and districts inNovember 2002. This reform was partly inresponse to the closing of the World Bank-financed PDSN project, which had financed district action plans through local private bank accounts. With the transition to budget support, the Ministry of Finance did not want to put public money into private bank, so the "regies d'avance" were proposed to function similar to a private bank checking account" for districts, but usingpublic funds. 222. While the health ministry welcomed this reform, it experienced a number o f problems and delays in the first year. First, while the account managers ("regisseurs") were named in March 2003, they did not assume office until November 20, 2003. Moreover, training did not extend to other members of the district health teams and regional treasury officials, who did not fully understand this new mechanism. Second, while the advances funds were created to facilitate the execution o f fungible expenditure-equipment maintenance, petty expenditure, personnel support other than health care, maintenance o f computers and peripherals-many districts assumed that they had to use these accounts to cover all delegated appropriations, even though procedures were better adaptedfor large orders ofmedicines or supplies.28 223. Third, the rules governing the accounts were undulyrigid: (a) only one-sixth of the total delegated credit was advanced at a time; (b) districts were not allowed to expend more than one- sixth of any given budget line, making it impossible to purchase supplies inbulk; (c) the list of eligible expenditures was seen as too narrow; (d) to qualify for account replenishment, receipts had to be sent to Ouagadougou (except in the one region that had decentralized the treasury function). Finally due to delays and confusion about procedures for applying for funds, the central Treasury did not transfer the first (one-sixth) tranche of funds into most accounts until after September 2003. Faced with a risk of significant shortfalls in district execution, the Treasury agreed to release the final 516 tranche once the first tranche was justified (thus amving inDecember o fJanuary inmost districts). 224. Despite these problems, most districts were able to execute over 90 percent o f their planned delegated budget, but with significant delays. Moreover, most district and regional officials interviewed remained positive about the reform's potential, and anticipated that performance would improve in2004. 225. Execution of non-deconcentrated portions of the government budget. The deconcentrated health structures do not participate in executing the government budget for 28 . The Banfora district officials suggested that the cash advances hnd approachbe used only for the following lines: (i)waste removal (with a more precise specification of the utilization methodology); (ii)petty cash funds; and (iii)personnel support (other than sanitary facilities), corresponding to travel expenses. - 78 - allocations that are not deconcentrated. In fact, they receive the bulk of their budget from the center in kind, in the form of personnel and supplies, for which the financial counterpart is not specified. For example, for repair works, the specifications are sent directly to the ministry, which handles this area by locating contractors from Ouagado~gou.~~Consequently, control is not always properly exercised and the quality of the infrastructures delivered i s unsatisfactory. It has therefore been suggested that the reform be sped up as regards deconcentrating the management of the health infrastructures line. In 2002 the DAF provided for the first time limitedhnds for regions anddistricts to supervise civil works. The regional directorates and the districts also receive fuel and office fumishings from the central level. The furniture is first received by the DRS, which then distributes it in accordance with the number of medical administration commissions (CMAs), medical centers (CMs), and CSPSs per district. However, often this equipment is poorly matched to pressing needs or is o f very poor quality For example, for sociocultural reasons, patients refused to be transported on the motorcycle ambulances purchased against HIPC resources. 226. Budget planning and execution by CSPSs and COGES.Since 2000, the CSPSs have been preparing what they call micro plans, which set forth the activities programmed for the entire year. COGES members participate in the preparation of these plans. Normally, the plans are prepared in September and October and then forwarded to the districts. InNovember, the nurses heading health posts (ICPs) meet to finalize and adopt these plans. The quality of plans varies considerably, however, and the plans often do not prioritize community health activities. The costs of the activities listed in these plans are quantified, and the ICP meeting amounts to a tradeoff committee that determines the final budget for each CSPS. To carry out these plans, each CSPS prepares, between the 1st and 5th o f each month, a provisional monthly cash flow planknown as the Monthly Budgeted Activities Packet (PMPLB), which is amended and adopted by the district. This plan is financed by the own resources generated by the COGESs from sales of essential generic drugs (MEGs) and in the form o f endowments in kind received from the district. 227. The contribution from the government budget arrives inkindas well. The CSPSs receive: (a) fuel under the advanced vaccination strategy for Expanded Program on Immunization-EPI (syringes, medicine for meningitis); (b) a financial contribution to fimd monitoring activities (food allowance for employees present); (c) products for the emergency treatment of pregnant women; (d) prenatal consultation (CPN) cards for weighing pregnant women. However, the permanence o f these supplies has been raised as a concern; (e) technical medical equipment (blood pressure kits, small surgical kits, stethoscopes); each CSPS has an equipment inventory file that is centralized by the district that supplies them in accordance with their needs and with what is available at their level. The CSPSs are not told the price o f these goods, and do not know what impact they have on the budget. In addition, each CSPS transfers between CFAF 10,000 andCFAF30,000 to the district supervisory team (ECD) eachquarter. 228. Six members elected by the members of the neighborhood health units (the sectors) head the COGESs. The six members of the COGES office are: (1) a chair and his or her deputy; (2) a 29 The Cascades DRS is currently experiencing difficulties in preparing such specifications, because the local contractors preparingthe specifications do not participateinthe centralizedcompetitivebidding in Ouagadougou.The DRSalso emphasized the problemsassociatedwith monitoringworksbecausethey are centralized. - 79 - treasurer and his or her deputy; (3) an auditor and his or her deputy; and (4) the IPC, which participates inthe committee meetings as a technical advisor. The election o f COGES members is not always transparent, however. Politically influential individuals from the village often hold the chair. Disbursements of funds to cover spending initially required three signatures (the IPC, the chair, and the treasurer). Embezzlement by some IPCs who had illiterate COGES chairs sign for improper amounts (for example, CFAF 2,000,000 instead of CFAF 200,000), led to the elimination o f IPC signature authority in the COGES. Unfortunately, as the chair and the treasurer are also not subject to controls, they often divert funds and are difficult to sanction owing to their political stature. Strengthened community oversight and electing women to the COGES offices could help reduce embezzlement. It is said about the COGES and it is true that they are more interestedinfinancial managementthan inpromotinghealthcare activities. 229. Auditing and monitoring of government, donor, and cost recovery funds. With support from the PADS project, the Ministry undertook the first global audit o f health districts, includingall government, donor, and cost recovery funds. While the auditors were able to overall certify the accounts of the PADS fimds, they were not able to do so for manydonors, nor for cost recovery funds. The audit resulted in a number o f recommendations, and will be repeated on a sample basis each year. Hospitals are not yet subject to independent, global audits, although they may be for the first time in 2005. In addition, the Ministry of Health also to establish a consolidated system for tracking income and expenditure of all COGES, cost recovery funds from district hospitals and district drug depots, and regional and national hospitals. This system is essentialboth for planningpurposes, as well as increasing accountability for these funds. E. CONCLUSIONSAND RECOMMENDATIONS 230. Establish results-based budgeting and monitoring system for the health sector. The MTEFhasthe potential to becomethe basis for results-basedfinancing system, andto strengthen the Ministry o f Health's ability to negotiate for additional resources from Finance and international partner. However, other reforms will be necessary to make the system fulfill its potential: e Establish a system for regularly monitoring program execution, including from partners, cost recovery, andbudget. e Strengthenthe health information system, to allow key national indicators to be available bylate-Januaryanddiscussedduringannualjoint sector reviews. e revise and strengtheningplanningdirectives for districts, to take into account priorities in MTEF; e Consider revisions in the planning and budgeting cycle, to allow priorities identified by healthdistricts andregions to be better integratedinto the final budget. 231. Improve the levei, efficiency and equity of resource allocations within the sector, particularly to ensure adequate financing for MDGpriorities. While resources for the health sector are limited, existing resources could be more efficiently allocated to accelerating progress toward the MDGs. Important steps are: -80- 0 increase percentage of national budget allocated for health (Abuja Declaration calls for 15 percent for health); 0 increase allocations for programs addressing major causes o f mortality and morbidity, including: malaria, reproductive health (contraceptive security), vaccination security, nutrition, integrated management o fchild illness, 0 increase the percentage o f the budget allocated to front-line services at the district and healthcenter level, and increase financing for community andhousehold interventions 0 develop a transparent criteria to improve equity o f budget allocations among regions and districts (e.g., based on population, poverty, etc.), and seek to improve geographic equity distribution of funds from international partners. 232. Improve equity and financial accessibility for preventive and curative services, Recent efforts to reduce margins on medicines and provide key preventative services for free appear to have improve utilization o f health services by the poor, but financial accessibility remains a bamer to use o f health services by the poor, particularly for emergency and curative services: 0 Strengthen monitoring and improve application o f existing prices reductions for preventive and primary care services; 0 Further reduce prices and develop transparent subsidy mechanism for critical services, including: emergency obstetric services (including cesarean sections); insecticide impregnatedbednets. Study costs and benefits o freducing costs ofnormal deliveries. 0 Clarify regulations and definition of "indigents", and pilot mechanisms for subsidized or free care for the very poor. 0 Continue to study potential role o f community insurance. 0 Since drug purchases represent the vast majority o f out o f pocket health expenditures, priorities include: improve prescription practices in public and private sector, continue campaigns to promote use of generic drugs, continue efforts to reduce profit margins o f essential medicines. 233. Improve the timeliness, flexibility, and reliability of sector funding. While trends are positive, funds from both government and partners are often unreliable, arriving late, or at less than budgeted amounts, and making it difficult to manage and execute programs. Priorities include: 0 Further streamline procedures for release o f delegated credits to health districts and regions (by Ministry o f Finance and MOH), to allow funds to be available by late January; 0 Improve functioning of cash accounts (vegies d 'avunce), including through continued training for account managers, and streamlining procedures (increasing the tranche release from 116 to ?4or 113 o f the total credit; removing line-item restrictions for expenditures; streamlinin~decentralizingprocess forjustifgng expenses); 0 Further decentralize management o f HIPC hnds, to allow these funds to be fully integrated into districtlregional planningandbudgeting process. - 81 - 0 Harmonize donor procedures, including through pooled financing and common reporting and auditing procedures, and strengthen systems to monitor execution of donor projects compared to commitments. ' 0 Reducepractice by Ministry o fFinance of blockingexpenditures for certain budget lines. 234. Strengthen accountability for funds at district and community level. The Ministry needs to put inplace a system to annually record receipts, expenses, and account balances at the district and community level (including district drug depots). The Ministryhas also initiated for the first time global audits o f accounts at health districts, which" should be continued and expanded to other levels of the health system. The texts goveming COGES also need to be revised in order to strengthen accountability to and participation of communities, expand membership (particularly women and youth), and strengthen their role in financing preventive andpublic health services. 235. Improve accountability and performancemanagementof hospitals.The government currently transfers resources directly to national and regional hospitals without a clearly defined performance agreement, but it also imposes various unfounded mandateson hospitals. 0 Develop performance contracts between government and hospitals, including indicators on service delivery, financial performance, and equity; 0 Establish system of annual reviews of hospital performance, and develop more transparent criteria for allocating budgetary resources amonghospitals; 0 Revise the system of user fees, and clarify financial and organizational framework for providingexemptions to the poor, and for mandated subsidizedcare for civil servants and students; 0 Initiate annual global audits for all national and regionalhospitals (planned for 2004); 0 Grant hospitals greater autonomy inhumanresourcemanagement; 0 strengthen systems for management and monitoring of finances, medicine supplies, and equipment, and humanresources; 0 strengthenquality monitoringand oversight role for the Ministry of Health. 236. Improvedistributionand motivation of health personnel.The inequitable distribution o f health personnel i s a long-standing problem in the health sector, which will require a multi- pronged approach, including: (a) strengthening incentives for practice in rural and poor areas, including possible financial incentives, and priority for training and professional advancement; (b) decentralization of recruitment and budget posts to the regional and district level; (c) strengthen systems for human resource planning, training, and management, including making the humanresourcessoftware systemoperational. - 82 - V. THEEDUCATIONSECTORINBURKLNAFASO: PERFORMANCEAND PROGRESS 237. This section analyzes the performance of the education sector inBurkina Faso duringthe past five years in terms o f outcomes, financing and expenditures and budget allocation and management. 238. The main message is that the education sector's performance has been improving steadily, but the challenges in terms of coverage, quality and effxiency of resource use remain high. Better coordination and linkages need to be made between sector priorities and programs and budget allocation and execution to improve the efficiency and effectiveness o f expenditure inmeetingthe educationdevelopmentgoalsofthe country. A. EDUCATION TRENDS OUTCOMES AND 239. Access and Coverage of the Education System. The number of students in the education system inBurkina Faso grew steadily over the 1999-2002period at an average annual growth rate o f 5.3 percent, reaching 1.2 million students in 2002. Eighty one percent o f the students in the system are at the primary level, which grew at an annual rate of 4.7 during this period. Student numbersat the secondary level grew much faster at an average annual rateo f 7.6 percent duringthe same period, accounting for 18 percent of the total number of students inthe system. Higher education accounts for the remaining 1percent. The overall gross enrollment rate (GER) (the proportion of the school age population who are enrolled at all levels) increased modestly by 3 percentagepoints from 19 percent in 1999 to 21 percent in 2002. At he primary level, the GER rose from 41.1 percent in 1999 to 45.1 percent in 2002 and reached47.5 percent in2003. Table 25: Gross Enrollment Rates by Instructional Level, 1998-2002 1998-99 1999-2000 2000-01 2001-02 2002-03 Pupillstudentenrollment 999 476 1041849 1110029 1169362 Primary education 816 393 852 160 901 291 938 238 1012 150 Secondary education 173 205 189 689 199 278 217 176 Higher education 9 878 9 460 13 948 Eligiblepopulation 5 227 340 5 352 842 5 429 408 5 560 366 2 129 125 Pop. 7-12 years (WSD adjusted) 1984 571 2 032 247 2 029 118 2 078 469 2 129 125 Pop, 13-19 years 1703 611 1744497 1786 365 1 829 238 Pop. 20-30 years 1539 158 1576 098 1613 925 1 652 659 Grossenrollmentratio(%) 19.1 19.5 20.4 21.0 Primary education 41.1 41.9 44.4 45.1 47.5 Secondary education 10.2 10.9 11.2 11.9 Highereducation 0.6 0.0 0.6 0.8 DEPI MEBA,DEPI MESSRSand MSD - 83 - x It in Table 26: Distributionof Enrollmentby Type of Establishment (publiclprivate), 1998-2002 1998-99 1999-2000 2000-01 2001-02 Primary 816 393 852 160 901291 938 238 Publicschoolpupils 727990 755 090 792 880 819 338 Private schoolpupils 88 403 97 070 108411 118900 % ofprivateschool pupils 10.80 11.40 12.00 12.70 Secondary 173 205 189689 199278 217 176 Publicschool pupils 115 878 124790 130711 139781 Privateschoolpupils 57 327 64 899 68 567 77 395 %privateschoolpupils 33.10 34.20 34.40 35.60 Source: DEPI MEBA and DEP i MESSRS 248. The private sector is also investing progressively inhigher education. In2002, the private sector's share o f higher education enrollments had risen to 10 percent. Most private institutions are concentrated inthe major cities (Ouagadougou and Bobo-Dioulasso) which raises access and equity concerns. 249. Equity within the educational system. Disaggregated data on enrollment rates indicate large disparities between urbanlrural populations, genders and income groups. Figures 31and 32 show trends in primary and secondary school GERs, respectively, by urbanlrural location and sex from 1994 to 2003. Urbanlrural disparities in enrollment rates are very large at the primary level. Between 1994 and 1998, the gross enrollment rate grew much faster inurbanthan in rural areas such that by 1998, the urban enrollment rate was more than triple the rural rate. Between 1998 and 2003, urban enrollment stabilized and rural enrollment rates continued its steady upward trend, narrowing the gap between the two. Nonetheless, in 2003 rural children are 3 times less likely to be enrolled inschool than their urbancounterparts. 250. There are also gender disparities inprimary GERs. The GER rate for girls i s 0.75 percent o f the enrollment rate for boys. Gender gaps are higher inrural than in urban areas. The gap in GERbetween boys and girls inurban areasnarrowedsignificantly inthe past five years to a very low level in 2003 (104 percent for boys versus 100 percent for girls). On the contrary, the primary GER for rural boys grew at a faster rate than for rural girls increasing the gender gap. Girls from rural areas are the most disadvantaged in terms o f access to education in Burkina Faso, both incomparison to boys inrural areas and to girls inurbanareas. Only one in four rural girls i s enrolled inschool, compared to one intwo rural boys. - 86 - En by la 03 : .;.I .! I :1 : L 1 ,.. . . ... . . ..... . .... . . . ... . . . . . - . iment rates, by locntronand 3 The more affluent a household is, the more likely i t is to enroll its children in school. Table 27 shows the trend ingross enrollment rates accordingto standard-of-living quintiles. Table 27: Primary and SecondaryEnrollment Ratesby Location, Sex and Standard of LivingQuintiles, 2003. Poorest Q2 Q3 Q4 Richest Primary education GER Rural 21 27 28 28 32 Male 28 31 33 31 38 Female 14 23 22 26 26 Urban 60 65 69 79 84 Male 61 62 76 81 89 Female 59 68 62 76 80 Total 23 30 34 37 53 Male 29 34 39 39 57 Female 16 26 28 35 48 Secondary education GER Rural 4 4 6 5 8 Male 5 5 8 6 12 Female 3 4 4 3 4 Urban 13 15 30 34 50 Male 13 19 27 37 55 Female 12 11 32 32 46 Total 5 5 10 12 27 Male 6 6 . '11 14 33 Female 3 5 9 10 23 Source: INSDBurkinaFaso Priority Survey, 2003 254. At the primary school level, poor households are half as likely to enroll their children as affluent households. Observing patterns by urbadrural and gender reveal that the disparities between urbadrural households are much more pronounced than between income groups. That is, children from rural households whether rich or poor, are more than two times less likely to enroll in school than children from urban households. Another observation is that gender gaps exist in both rural and urban households, regardless o f living standard. The most disadvantaged are clearly rural girls from the poorest households. 255. The divergence deepens at the secondary level, but the same pattern holds. Poor households inthe first quintile are only one-sixth as likely to enroll their children as those inthe last quintile. Moreover, rural households are at least 3 times less likely to send their children to secondary school, across all wealth quintiles. As with the primary level, it is rural girls from the poorest households who are the most disadvantaged. 256. Quality of the educationa1,system.Repetition and dropout rates are proxy measures o f the quality ofthe education system. At the primarylevel, repetition and dropout rates, which had been steadily worsening over the 1999-2002 period, began to improve in 2003 (Table 28). At least 3 out of 50 students drop out of school inmid-year between the CP1 and CM1, and at least - 88 - one out of 10pupils repeats a grade at the CP1and CP2 levels. From the CE1and CM1onward, about 2 out o f 10 repeat a grade. The repetition rate is high in the CM2, and the figure has scarcely improved since 1998, oscillating between 34 and 42 percent. This means that over 113 o f CM2 pupils repeat their grade. The highrate o f repetition in the CM2 explains the very low rate o f transition from the primary to the secondary level. Indeed, over the 1998 - 2002 period, only about one CM2 pupilinthree was able to go on to secondary school. Table 28: Repetition and Dropout Rates at the PrimaryLevel, 1998-2003 1998-99 1999-2000 2000-01 2001-02 2002-03 1998-2003 Repetition rate in primary Repetition rate inCP1 11% 12% 12% 12% 11% 12% Repetition rate inCP2 11% 13% 13% 13% 11% 12% Repetitionrate inCE1 16% 17% 17% 18% 16% 17% Repetition rate inCE2 13% 16% 16% 16% 15% 15% Repetition rate in CMl 16% 18% 19% 19% 17% 18% Repetition rate inCM2 34% 37% 42% 36% 34% 36% Dropout rate in primary Dropout rate inCP1 6% 7% 8% 5% 6% Dropout rate in CP2 7% 6% 6% 3% 5% Dropout rate inCE1 9% 10% 10% 6% 9% Dropout rate inCE2 10% 10% 10% 6% 9% Dropout rate inCM1 10% 12% 9% 6% 9% Source: DEPJMEBA 257. Performance o f the secondary level i s even lower (Table 29). The repetition rate is high from the 6" class to the final, or terminale, class. Over the 1998-2001period, at least one pupilin four inthe 6" class repeated the grade, and the figures for mid-year dropouts at the same level are similar. This means that, as early as the 6", halfo f those enrolled do not go on to the next grade. After the 6" class, the rate o f repetition increases while the dropout rate decreases, to settle at a level under 20 percent for the rest o f the first cycle o f secondary education. This suggests that those who continue to 6e class are less likely to drop out before reaching lere, but are likely to repeat. 258. The real bottleneck in the system i s at the final class o f secondary terminale where both repetition and dropout rates are particularly high. Of five terminale pupils, at least two repeat the terminale grade, and nearly three drop out o f school inthe course o fthe year. 259. The poor quality o f the education system may be partially explained by the fact that the pupillteacher ratio is high in Burkina Faso. At the primary level, the pupillteacher ratio was 54: 1 in 1998 compared to 40:l for Sub-Saharan Africa as a whole, excluding Nigeria and the Republic o f South Africa. While this ratio improved gradually over the 1998-2002 period, reaching 51:1duringthe 2002103 school year, it is nonetheless still high. - 89 - Tabie 29: Repetition and Dropout Rates at the Secondary Level, 1998-2003 1998-1999 1999-2000 2000-2001 1998-2001 Repetition rate insecondary Repetitionrate inde 24% 23% 24% 24% Repetitionrate in5e 31% 23% 31% 28% Repetitionrate in4e 33% 30% 30% 31% Repetitionrate in3e 38% 39% 44% 40% Repetitionrate in2e 21% 19% 22% 21% Repetition rate inIere 26% 23% 23% 24% Repetition rate interminale 38% 46% 44% 43% Dropout rate insecondary Dropout rate in6e 20% 22% 24% 22% Dropout rate in5e 14% 21% 14% 16% Dropout rate in#e 12% 12% 12% 12% Dropout rate in3e 15% 16% 19% 17% Dropout rate in2e 23% 24% 20% 22% Dropout rate inIere 27% 24% 19% 23% Dropoutrate inlermina/e 62% 54% 56% 57% Source: DEPMESSRS 260. At the secondary level, the pupillteacher ratio is also relatively high compared to WAEMU and other Afkican countries. The pupillteacher ratio nearly doubled from 1999to 2001, rising from 28: 1 to 53:1. This means that a secondary school teacher in2001 had about twice the number o f pupils as in 1999. Staff shortages are one o f the main reasons for the poor performance o f the secondary education system. Available data on staff shortages indicates an overall deficit o f 1,099 secondary teachers, all subject areas combined, for the 200312004 school year.30 This deficit represents nearly 113 o f all secondary teachers in 2002. The Regional Directorate for Secondary Education (Direction regionale de 1'enseignementsecondaire, DRES) o f the Cascades province estimates its staff shortage in most years at about 30-40 teachers, a shortage that i s still only partially covered. Staff is concentrated in major cities such as Ouagadougou andBobo-Dioulasso. Although rural teachers receive a larger hardship allowance, the advantages of the larger cities are significant: lifestyle (electricity, activity, recreation), opportunities for adjunct teaching in private schools, etc. The result is that rural areas have a teacher shortage. 261. The teacher shortage is more pronounced in the scientific disciplines (math, physics, chemistry, life and earth sciences) than in other subject areas, representing 60 percent of the overall deficit (Figure 33). Parentlteacher associations (APES) sometimes recruit and pay teachers. There is also a shortage o f superintendents for highschools andmiddle schools. 30Source: Humanresources directorate, MESSRS, 2004 - 90 - rates, the education system would be able to cater for 3.5 times more students through the primary cycle, 4.4 times more students to the BEPC and 3 times more to the BEPC without additional resources. Table30: AverageCost perStudentCompletinga Cycle Instructional Level Average yearly Duration of Average cost of Actual cost of cost of instruction cycle student per cycle student per cycle Primary (CEPE) 37 714 6 226 283 791 992 Secondary (le cycle 1BEPC) 70 742 4 282 968 1245 059 Secondary (2nd cycle BAC) 70 742 3 212 226 615 455 University ( degree) 854 924 4 3 419 696 26 331 662 Overall cost of educating one 495 194 1860514 secondary school graduate (BEPC + BAC) Overall cost of a student 4 141 174 28 984 167 completing a university degree 266. The highest wastage is at the university level, where the actual cost o f educating a university student is 8 times more than what it should be. There are also other inefficiency at the higher level in the national grant and loan scheme. About 415 o f students receive government support in the form o f grants or loans from the National Education and Research Fund (Fonds national pour 1'&ducationet la recherche, FONER). Annual per-student aid amounts were CFAF 130,000 for grants and CFAF 165,000 for loans. Recipients o f non-reimbursable grants and aid made up over 213 o f the student body inhigher education (Table 31). 267. FONER aid and loans cost the government about CFAF 1.5 billion per year, without counting study grants. FONER, which was created in order to replace direct social assistance with reimbursable loans, accounts for only 13 percent o f students in higher education. This finding implies that the mechanism o f replacing direct aid with reimbursable loans failed to work, since 213 o f students continue to receive direct subsidies (aid or grants). Moreover, this fund, which uses government-provided money, has not recovered any loans since it was created (0 percent recovery rate.) That being so, the fund is apparently ineffective in terms o f fulfilling the objectives that were behind its creation, and the loans are being used as a form o f grant. 268. The inefficiency of higher education weighs heavily on the educational system as a whole. It limits the Government's ability to develop the entire system rapidly. Economies of scale could be achieved ifthe performance o f higher education were improved. Furthermore, in order to improve the efficacy andbalance o f the entire educational system, it will be necessary to redirect a portion of the resources o f higher education to the lower levels (Le., primary and secondary), while developing mechanisms for participation inthe costs o f education. - 92 - Table 31: Status of Student Grants, Aid and Loans for Higher Education,200112002 Grants FONER FONER Total (grants, aid& TotalperBurkinabl Aid Loan loans) student Studentsreceivingaid 2 192 8 144 1975 12 311 15 398 Proportionof total students 14% 53% 13% 80% 100% 269. Both the quality and type o f education impact on the extemal efficiency o f the system in terms of providing the necessary skills to meet the demand o f the labor market and o f the development needs of the country in general. The focus on formal theoretical training at the secondary and higher levels o f the education system does not meet the reality o f the country's development goals. 270. The Government has focused its policy on general education, which accounts for about 97 percent of total secondary enrollment, while the technical category represents only 3 percent. Within technical and vocational training, the focus is on providing tertiary skills (accounting, secretarial, sales and marketing, information technology, etc.) more than industrial skills (agriculture, masonry, electronics, mechanics, industrialproduction, engineering, etc.) 271. Similarly inhigher education three quarters o f students are enrolled inthe tertiary branch studying science, economics, law, literature, health, etc. while only a quarter are inthe'industrial branch studying life sciences, physical sciences, mathematics, statistics, computer science, engineering and allied technologies, processing industries, etc. Within the tertiary branch, training options are concentrated mainly inthe area o f accounting andmarketing, which account for 60 percent o f the establishments in the sector, whereas fields such as hotel management, banking and information technology, taken together, account for only 5 percent of establishments in the sector. Inthe industrial sector, training focuses mainly on auto mechanics, masonry and civil engineering. The options for training in the agro-pastoral area are limited, despite the fact that the economy o f the country is primarily based on agriculture. Establishments providing training in agro-pastoral subjects represent 7 percent o f industrial sector schools and about 3 percent of all technical and vocational schools. B. EDUCATION FINANCINGAND EXPENDITURE MANAGEMENT 272. Financing of Education. Education inBurkina Faso is essentially financed through the government, and by families, NGOs and the private sector. Data on financing by communities, NGOs and the private sector are not available. Public financing of educatiod is effected either through the government's own resources (including non-targeted budgetary support), or through externally funded projects. The HIPC Initiative offers additional funding. Other technical and financial partners intervene, either directly or through NGOs, without going through the government budget. 273. The Government's own resources are the main source o f public funding for education. In 1998, government funding represented 77.9 percent o f such financing, whereas externally - 93 - financed projects accounted for only 22.1 percent. A change in the structure of financing was notedbetween 1998 and 2002; the share of capital financing fell from 34 to */3, This difference is essentially attributable to the start-up o f the PDDEB and to the availability of HIPC resources, HIPC resourcesmade a significant contribution to the financing of education in2002, amounting to 10.2 percent of total education expenditures. This mode of financing of education became effective as of2001. 274. Levels and trends in education expenditures. Public expenditure devoted to education (at the primary, secondary and tertiary levels) increased steadily over the 1998-2002 period (Table 32). Its volume increased from CFAF 47.2 billion @e., 2.6 percent of GDP) in 1998 to CFAF 87.6 billion in 2002 (Le., 3.9 percent of GDP). This increase translates into an annual growth rate of about 15 percent. Accounting for inflation so as to gauge expenditures in real terms, the increaseworks out to 12.7 percent. Table 32: Sourcesof Financingfor Education (in billions of CFAF), 1998-2002 Fiscal year 1998 1999 2000 2001 2002 Total public expenditurefor education(inCFAFbillion) 47.2 51.5 68.0 74.9 87.6 Of which Govemment funds (includingbudgetary support (inCFAFbillion) 36.7 47.4 49.3 56.2 58.3 HIPC (in CFAF billion) 2.7 8.9 Extemal financing(excludingbudgetarysupport (in CFAFbillion) 10.4 4.1 18.6 16.0 20.3 Share of Govemment funds (includingbudgetary support) in educationbudget ("h) 77.9 91.9 72.6 75.0 66.6 Share of HIPCresources in educationbudget(%) 3.6 10.2 Share of extemal resources (excludingbudgetarysupport in%) 22.1 8.1 27.4 21.3 23.2 Share of govemmentfunds devotedto educationrelativeto total govemment 15.4 18.6% 19.9% 20.9% 19.3% resources(includingbudgetarysupport in%) Share of extemal financingdevotedto educationrelativeto external govemment 6.4% 2.6% 10.1% 6.6% 9.9% resources (excludingbudgetarysupport inYO) Total Govemmentexpenditure(in CFAFbillion) 400.7 416.9 432.8 510.6 509.1 Of which Govemment resources(includingbudgetarysupport in CFAFbillion) 238.9 254.6 248.5 268.8 302.6 Extemalfinancing(in CFAFbillion) 161.8 162.4 184.3 241.7 206.5 Source CID, DCEI, DGCOOP, DCCF 275. Education expenditures, as a percentage o f the total govemment budget, oscillated between 11.8 percent and 17.2 percent over the .1998-2002 period (Table 33). On average over the five years, education expenditures accounted for 14.2 percent of the entire government budget. Comparing the share of education expenditures for the 1998-99 period and that for the 2000-02 period, which correspondsto the approval and implementation of the PRSP, shows an increase of 3 percentagepoints in the share o f education in total expenditures. This expenditure trend over the second period is attributable particularly to the substantial resources channeled into the sector, e.g., the HIPC Initiative, the start-up o f the PDDEB in2002, and implementation of the World Bank-financed Post-Primary Education Project (PPEP) and o f the African Development Bank's Education Tv Project. - 94 - Table33: EducationSector ExpenditureIndicators,1998-2002 Fiscal year 1998 1999 2000 2001 2002 Totalpublic expenditure (inCFAF billion) 400.7 416.9 432.8 510.6 509.1 Expenditures allocated to education (inCFAF billion) 47.2 51.5 68.0 74.9 87.6 Annual rate of increase ineducation expenditures (%) 9.2 31.9 10.2 16.9 Share of education budget relative to total public 11.8 12.4 15.7 14.7 17.2 expenditure (%) Share of education budget relative to nominal GDP (%) 2.6 2.8 3.6 3.6 3.9 Nominal GDP (inCFAFbillion) 1826 1834 1890 2062 2233 Inflationrate ("A) -1.1 -0.3 4.9 2.3 Source: CID, DGCOOP, INSDand IAP 276. Despite this steady increase in Burkina Faso's education expenditures, the share o f GDP that they represent i s still smaller than it is inother countries, andparticularly inother WAEMU countries. Thus, from 1993 to 1998, Burkina Faso spent an average of significantly less than 3 percent per annum of its nominal GDP on education, whereas the figure was 4.5 percent for Togo, 3.7 percent for C6te d'Ivoire, and 3.5 percent for Senegal. Burkina Faso is, however, making a similar effort as some other WAEMU members such as, for example, Niger (2.7 percent) and Benin(2.6 percent). 277. Distribution of education expenditures by economic classification, Recurrent expenditures (salaries, goods and services, current accounts transfers) constitute the largest share o f education expenditures. In 1998, they account for about 213 o f education expenditures, whereas investment represented 113. This trend did not continue over the 1998-2002 period, however. The share devoted to recurrent expenditure decreased significantly. With an average annual growth rate of 5 percent, current expenditure represented only 60.3 percent in 2002, whereas the more rapidly growing investment expenditures (38 percent on average, between 2000 and 2002) accounted for 39.7 percent of total public expenditure for education. This i s mainly due to the large investment projects that were implementedinthis sector over the period (e.g., the Post-Primary Education Project) and by the start-up o f the Ten-Year Basic Education Program (Programme dkcennal de I '&ducation de base, PDDEB) which relies upon a different form of financing, Le., sectoral support. Inaddition, H P C resources were devoted to the primary education sub-sector, which, in 2002, represented, when taken alone, 10.2 percent o f total sector expenditures. 278. Recurrent expenditures in the education sector consist mainly of salaries (66 percent o f recurrent spending). The relatively small proportion o f expenditures devoted to goods and services or to operations and maintenance (O&M) (5 percent) is partially attributable to the fact that parents also contributeto the operation o f schools andpublic institutions. However, with the introduction o f the delegated credit mechanism at the primary level, as well as the implementation o f specific poverty-reduction measures, particularly inthe twenty provinces with the lowest enrollment rates (e-g., provision of free school supplies, waiver o f school fees for girls, contributions by parents), the share represented by O&M expenditures has increased significantly, reaching 8.4 percent in2002. - 95 - 279. Recurrent transfers in the education sector are made primarily for the benefit of training institutions and other autonomous establishments, such as public establishments (EPA) and public cultural and scientific establishments (~~ablisse~entspublics Ir caract2re culture1 et sczentzjique, EPCS). Some non-autonomous entities (e.g., FONER, National Commission of UNESCO, etc) also receive current transfers for their operations. Transfers are an important component of the education sector's current expenditures, but one that i s declining steadily. Between 1998 and 2002, transfers went from about 113 to ?4of current expehditures. Consisting primarily of the salaries o f autonomous establishments, study grants and subsidies for pupils and students, and operating supplies, this decline is partially due to the quota policy instituted inthe award of study grants. 280. Distribution of education expenditures by administrative and operational level?' Assuming that all expenditures on the part of the central units are of an administrative nature and that the administrative expenditures of decentralized units are marginal, it becomes clear that most education expenditures are devoted to operational structures (Table 34). In 1998, expenditures by operational structures represented 94 percent of total public expenditure for education, whereas the share attributable to administrative structures was 6 percent. 281, Average annual operational and administrative expenditures over the 1998-2002 period increased at a rate o f 13 percent. However, with the implementation period o f the PRSP, Le., the period from 2000 onward, operational expenditures increased more rapidly, Le., at an average annual rate of 17 percent, whereas administrative expenditures are increasing at an average rate of 8 percent per year. This reflects the government's objective to devolve responsibility to the structures directly involved inproducing educational services. Table 34: Distributionof EducationExpenditureby Administrativeand OperationalLevel, 1998-2002 Fiscal year 1998 1999 2000 2001 2002 Operatingexpenditures (inCFAF billion) 44.2 46.9 63.2 68.4 80.8 Administrative expenditures (inCFAF billion) 2.9 4.6 4.8 6.6 6.8 Total public expenditure for education (in CFAF 47.1 51.5 68.0 75.0 87.6 billion) Share of operating expenditures in education 94% 91% 93% 91% 92% expenditures (%) Share of administrative expenditures in education 6% 9% 7% 9% 8% expenditures (%) Growth rate of operating expenditures (%) 6% 35% 8% 18% Growth rate o f administrative expenditures (%) 59% 4% 38% 3% Source: CID, DCEI, DGCOOP,DCCF 282. Distribution of education expenditure by education level. Educational expenditures are concentrated on the primary system, which absorbed over 63.4 percent o f public funds 31 Since the expenditures devoted to regional education departments cannot be broken down in such a way as to identify those that are directed at schools and institutions, any classification o fexpenditures according to operational and administrative structures can only be approximate. - 96 - devoted to education over the 1998-2000 period. The secondary system accounted for 21.4 percent. The remaining 14.2 percent was for higher education. This trend continued over the 2000-2002 period with the approval o f the PRSP as a frame o f reference for development policy and the launching o f the PDDEB. Over this period, annual expenditures for the primary level averaged 62.0 percent, for the secondary level 24.7 percent and for higher education 13.4 percent. The period also saw a slight increase in the share o f expenditures for secondary education. 283. Innominal terms, expenditure at all levels increased significantly (Table 35). Expenditure at the primary level almost doubled reaching 58 billion CFAF'. At the secondary level, expenditure tripled between 1998 and 2001 and slightly dropped to 19 billion CFAF in 2002. Higher level expenditure increased less dramatically from 7.4 billion CFAF in 1998 to 10 billion CFAF in2002. As a share o f total public education expenditure, expenditure at all three levels o f education fluctuated duringthe 1998-2002 period. The share ofprimary and higher education in 2002 droppedfrom their level in 1998 while that o f secondaryincreased. Table35: DistributionofEducationExpendituresby InstructionalLevel, 1998-2002 Fiscalyear 1998 1999 2000 2001 2002 Total public expenditures for education (inCFAF billion) 47.2 51.5 68.0 74.9 87.6 Expenditures allocated to primary education (in CFAF billion) 32.2 30.4 43.5 40.8 58.5 Expenditures allocated to secondary education (inCFAF billion) 7.5 13.6 15.0 22.8 19.1 Expenditures allocated to higher education {in CFAF billion) 7.4 7.6 9.4 11.3 10.1 Share o fprimary ed. inpublic expenditures for education {%) 68.3 58.0 64.0 54.5 66.7 Shareo f secondary ed. inpublic expenditures for education (%) 15.9 26.4 22.1 30.5 21.8 Share of higher ed. inpublic expenditures for education (%) 15.7 14.7 13.9 15.1 11.5 Source: CID#DCEI,DGCOOP,DCCF 284. Overall unit costs are on an upward trend in current prices, having risen from CFAF 49.663 in 1999 to CFAF 54.261 in 2002 (Table 36). As a percent o f nominal GDP, however, they tended to drop from 1999 to 2002, from 35.3 percent to 33 percent o f per capita GDP. At the primary and secondary levels, unit costs increased more or less steadily, but at the level of higher education, unit costs per student oscillated. The per-student unit cost trend is downward, however. The exceptional case of 2001,when unit costs spiked at CFAF 1.193.45 1per student, or 751.5 percent of per capita GDP, i s explained by the social upheaval experienced by the University in 2000 with the cancellation o f the academic year. The government increased operating expenditure sharply in order to make up for the impact of a missed year, as well as to implementits new re-establishment policy for the University. 285. The difference between unit costs at the various instructional levels is particularly striking, andmeans that higher education i s still strongly supportedby the Government budget at the expense o f the lower levels. A public secondary school pupil is the beneficiary, on average, o f twice the amount of annual Government subsidy as a public primary school pupil. This disparity is even more profound if one compares the Government subsidy received by a pupillstudent.A student in higher education receives 23 times more subsidies than one in the primary level, and 12 times more than a pupil at the secondary level. Inother words, it costs the - 97 - Government about as much to educate two university students as it does to educate a class of 50 pupils at the primary level. The relatively highunit costs at the higher education level are due to the high salary levels o f university professors, the high cost o f certain laboratory materials used inuniversities and research centers, and the existence of student aid inthe form of grants and loans. Other types o f subsidization also contribute the costs o f higher education: e,g., student meal plans, transportation and housing. At the Polytechnical University o f Bobo-Dioulasso, the Government covers transportation for students, teachers and administrative staff for the entire academic year, andthis is reflected inper-student unitcosts. Table 36: PublicExpendituresfor Education, by Student and by InstructionalLevel, 1998-2002 Fiscalyear 1998 1999 2000 2001 2002 GDPicapita(inCFAF) 137 578 140874 146 580 158815 164 241 Current expenditures(in CFAF billion) 33.6 42.4 45.7 50.7 52.8 Primary(inCFAF billion) 20.4 26.4 27.9 29.7 33.1 Secondary(in CFAFbillion) 5.8 8.5 8.3 9.7 9.6 Higher (inCFAF billion) 7.4 7.6 9.4 11.3 10.1 Public enrollment 853 746 879 880 933 051 973 067 Primary 727 990 755 090 792 880 819 338 Secondary 108493 115 878 124790 130711 139781 Higher 9 160 9 878 9 460 13 948 Unitcost perpupilistudent 49 663 51 939 54 338 54 261 Primary(inCFAF) 36 197 36 952 37 453 40 384 Secondary(inCFAF) 53 214 73 253 66 689 74 320 68 980 Highered. (CFAF) 809 092 766 319 1193451 721 935 Unit cost as a % of GDPicapita 35.3 35.4 34.2 33.0 Primary(%) 25.7 25.2 23.6 24.6 Secondary (%) 38.7 52.0 45.5 46.8 42.0 Higher ed. (%) 588.1 544.0 751.5 439.6 Source: CID, DCEI, DGCOOP, DCCF, INSD, IAP, DEPIMEBA, DEPNESSRS 286. Efforts will thus need to be made to diminish per-student costs at the university level, and to transfer more resources to the lower levels &e., the primary and secondary schools). In addition, the rapid growth in enrollments means that a mechanism for contribution and cost recovery inhigher education mustbe established to supplement direct public expenditure. c. PROGRAMBUDGET ALLOCATIONAND EXECUTION 287. Program budgets. The program budget reiterates inprinciple all the public expenditure forecasts at the level o f the sector in question. In practice, program budgets are not always exhaustive, however. This is true, for example, in primary education where personnel expenditures are not reflected inthe program budget. The programs inthe following table can be examined in the light o f information contained in the program budgets of the two education ministries. - 98 - Table 37: Distributionof Program Budget by Level of Education,2000-2003 Promambudgetof theeducationalsystem Allocation 2001 Forecast 2002 Forecast 2003 Primary 22.2 14.3 20.1 30.1 1. Increaseof supplyof primary education 6.9 5.9 8.3 12.4 2. Improvedquality andrelevanceof basiceducation 2.4 3 4.3 6.4 3. Increasedsupply ofliteracyservices 0 0.2 0.2 0.3 4. Improvementof MEBA's capacityto plan, manageand 12.8 5.2 7.3 10.9 Secondary 4.7 19.6 18.4 21.3 1 , Increasedaccommodation capacityin secondaryschools 1.9 2.9 2.3 1.6 2. Increasedaccommodationcapacityin secondary 0.2 9.5 9.7 14.3 3, Improvedquality ofsecondary education 1.6 3.4 4 3.4 4. Enhancedadministrativeandmanagerialcapacitiesat 1.1 3.8 2.4 2 Tertiary 14.6 31.8 27.5 23.6 1. Increasedacco~odationcapacityin highereducation 0.9 5.6 3.8 1.6 3. improvedqualityofhigher education 6.6 17.5 16.2 15 4. Strengtheningofresearchentities I.5 1 0.4 0.3 5. Disseminationof resultsof scientific research 1.5 1.1 0.5 0.3 of which Shareofprimary (as a YO) 53.4 21.8 30.5 40.1 Shareof secondary(as a%) 11.3 29.8 27.9 28.4 Shareof higher ed. (as a%) 35.1 48.4 41.7 31.4 Source: MEBA et MESSRS 288. Programs are concentrated at the primary and higher education levels, althoughpersonnel expenditures at the primary level are not accounted for (Table 37). Between 2001 and 2003, the forecast share of the primary program represented 34.5 percent as an annual average, that of the secondary program 22.7 percent, and that o f the higher education program 38.6 percent. This indicates very clearly a certain bias in favor of the development of the primary and higher education levels. 289. Within each sub-system, the program for increasing educational services at the primary level is receiving nearly the same level o f resources as the program concerned with improving managerial and administrative capacities. This is explained in part by the fact that teacher salaries are not taken into account in the service expansion program. Moreover, the program intended to improve managerial and administrative capacities treats the project's operating costs, as well as the equipment, services and supplies o f the central and decentralized departments, as national counterpart funding, thereby making this part o fthe program quite large. 290. At the secondary level, the distribution among the various programs developed and among instructional categories within the sub-system seems to be better. The program aimed at increasing the absorptive capacity o f secondary technical and vocational education is receiving 213 o f the resources earmarked for the sub-system. The program to improve the quality of higher education is absorbing 213 of the resources earmarked for the tertiary system. This is due to the fact that this program takes into account the salaries o f university teachers as well as student grants. 291. Forecasts under program budgets can also be grouped into three broad programs at the level of the educational system as a whole: namely, the program to increase educational system - 99 - offerings, the program to improve educational quality and relevance, and the program to strengthenthe system's administrative and managerialcapacities. 292. In 2001 and 2003, forecasts contained in program budgets earmarked for service expansion programs in the various sub-systems (Le., primary, secondary and higher), an annual average o f 38 percent o f the resources allocated to the sector. The program for improvingoverall system quality receives an annual average of 37 percent. Nearly 114 is allocated annually, on average; to the program aimed at enhancing the managerial and administrative capacities of the ministries. 293. Within the budgetary allocation for 2000, the structure of expenditures according to the three main programs was: 23.8 percent for service expansion; 32.7 percent for improvingsystem quality; and 43.5 percent for strengtheningmanagerial and administrative capacities (Table 38). Ifthedefinitivebudgets,asdevelopedandexecuted,hadreflectedtheprogrambudgets,this would have been a revolutionary change in the financial structuring of the programs in the educational system infavor of increased supplyandimprovedquality and relevance. Table 38: Distributionof ProgramBudgetby ProgramObjective, 2000-2003 Fiscalyear Allocation 2000 2001 Forecast2002 Forecast2003 Forecast Increasedsupply 9.9 24.1 24.3 30.2 of which Primary 6.9 6.1 8.5 12.7 Secondary 2.1 12.4 12 15.9 Improvedquality 13.6 26 25.4 25.4 o fwhich Primary 2.4 3 4.3 6.4 Secondary 1.6 3.4 4 3.4 Enhancement of administrativeandmanagerialcapacities 18.1 15.6 16.1 19.3 o f which Primary 12.8 5.2 7.3 10.9 Secondary 1.1 3.8 2.4 2 Total 41.6 65.7 65.8 74.9 Shareof programsdevotedto increasingsupply 23.80% 36.70% 36.90% 40.30% of which share of primary 69.70% 25.30% 35.00% 42.10% o fwhich share of secondary 21.20% 51.50% 49.40% 52.60% Shareof programsaimedat improving quality 32.70% 39.60% 38.60% 33.90% of which share of primary 17.60% I1.SO% 16.90% 25.20% o fwhich share of secondary 11.80% 13.10% 15.70% 13.40% Share of programsdevoted to improving administrativeand 43.50% 23.70% 24.50% 25.80% of which shareofprimary 70.70% 33.30% 45.30% 56.50% of which shareofsecondary 6.10% 24.40% 14.90% 10.40% Source: MEBA et MESSRS 294. Absorptive capacity of the educational system. In general, the system's absorptive capacity is somewhat better than the overall rate of execution of the government budget (Figure 34). Budget execution at the primary level is higher than that o f the secondary level, due to the - 100- thit feci in t 297. The execution rate for budgets inthe basic education sector improved over the 2000-2002 period, reaching an annual average of 94.7 percent (Table 39). This improvement is mainly due to the highrate o f execution of the portion financed nationally, because ofreforms implemented in the sector to improve budgeting procedures, e.g., the delegation to decentralized units of a portion o f operating credits, given the poor quality of annual forecasting of initial allocations for personnel expenditures. 298. The execution rate for expenditures under externally-financed and HIPC projects also improved perceptibly over the period. Between 2001 and 2002, however, it dipped to low levels (51 percent and 62 percent, respectively), mainly because o f the weak absorptive capacity with regard to HIPC funds. 299. Absorptive capacity in secondary and higher education improved overall over the 1998- 2002 period. It is stillweaker thanthat ofMEBA, for three primaryreasons: 0 The ministry's strong centralization o f credit management. All credits for central and decentralized departments, with the exception o f autonomous establishments such as the EPAs and EPICS,are executed bythe DAF. 0 National budgeting procedures are still cumbersome. Procurement at the level of the MESSRS can take 4 to 5 months. Budget execution begins in February, and commitments close inNovember. Constraints associated with the procurement planmean that budget execution cannot be concentrated within that same timeframe. Inaddition, it i s impossible in certain specific areas to group purchases. In addition to these procedural constraints, many credit-related regulations have been instituted over the past few years. 0 Procedures for the execution o f expenditures under donor-financed projects are another source o fthe weak absorptive capacity o f MESSRS. 300. Budget execution and efficacy of resource management. The delegated credit mechanism introduced to MEBA is a positive development that makes local actors accountable and enhances the transparency o f the budgetingprocess, as well as the efficiency o fthe supply o f educational services. 301. The regional departments o f basic education and literacy (DREBAs) receive the following for purposes o f implementingeducational policy at the regional level: 0 Delegated credits for the regional directorates, which they then manage directly; 0 &el and equipment (pupils' desks, teachers' desks, chairs) which are sent by the ministry. The monetary value of the equipment receivedis not specified; 0 construction o f school complexes (3 classrooms, latrines, teacher's quarters, pupils' desks, chairs, teacher's desk) is carried out in the regions, but regional departments are not involved inworks oversight; 0 teaching personnel, some of whom are paid out o f HIPC resources, with their salaries (CFAF 50,0001month) transiting directly through the regional department to the beneficiaries. 302. Under the HIPC specifically, the regional departments that deal with the 20 provinces with the lowest school enrollments do not receive money directly, but instead get supplies for - 102- these provinces. These usually consist of notebooks, binders, T-squares, compasses, pens, etc. The supplier delivers this materiel to the localities, and it is delivered from there to the schools. Orders are placed directly with Ouagadougou, and delivery occurs on site. The school cafeteria system does not exist inall regions. Regions deemed self-sufficient, such as Cascades and Hauts- Bassins, do not have them. 303. Delegated credits are earmarked for certain types o f expenditures, such as: e consumables (e.g., chalk, blackboard surfacing, etc) distributed on the basis of the number o f classes per school. These supplies are usually distributed directly to the various localities without going through the provincial departments. Heads of localities handle the distribution to the schools. The supplies received by the localities are then distributed among the schools: 2 boxes of white chalk and 1 box of colored chalk per quarter and per school. Fwnishings are distributed o n the basis of needs expressedby the schools; e maintenance and repair: each provincial department is asked to specify its needs. A general order is placed with a licensed mechanic who then performs the repairs; the petty cash fund (CFAF 50,000 per circumscription per quarter): the process bars a circumscription from replenishing its fund more than once a year. These funds are managed by individuals appointed by the Ministry o f Finance. 304. The execution rate o f delegated credits i s generally satisfactory (Table 40). For the DREBA in Cascades province, for example, it stands at 98.3 percent. This mechanism has enhanced the efficiency of the decentralized structures, and especially o f the DREBAs and schools. 305. Until 2004, the provincial directorates (Directions provinciales de I 'enseignement de base et de I ' a ~ h a ~ ~ t i s a t iDPBA) and localities were not direct recipients of the delegated o n , credits distributed. They received materials, however, based on their statement of needs in terms o f consumable school supplies, furniture and maintenance products. 306. Under the current mode o f supply, from localities to the schools, some inspection departments keep a portion of the material, because it was noticed inpractice that some teachers make o f f with leftover, unused supplies. The DREBA also holds back some supplies as a margin of security. 307. The process o f delegating credits should be eventually expanded to the localities. There will also be a need to decentralize the management of school equipment (pupils' and teachers' desks, etc.) in order to reduce transaction costs and improve the quality of the equipment received. With the assignment of buildingtechniques to the provincial directorates, repair work may also be delegated to the provincial level. 308. Access of decentralized units to resources and their utilization at the secondary level. The delegated credit mechanism developed byMEBA does not exist within MESSRS. The decentralized units (DRES, lower and upper secondary schools) do not receive delegated credits. Their credits, which come in the form of a budget allocation for equipment expenditures within - 103 - the Appropriations Law, are managed at the central level, i.e., by the DAF. They do, however, receive materialinkind(albeit ininsufficient quantity) consisting of desks, chalk, paper, stencils, etc., without there being any corresponding financing. The centralized ordering system also poses many problems with delivery to the site. The equipment often anives late because the DREShas no budget for transportation. The desks are built inOuagadougou, where chalk is also purchased, and these materials are dispatchedwith significant loss or damage. Distributionto the highschoolsandmiddleschoolsoccurs onthe basisofinitial reportspertainingto the start of the school year. The criterion is the size of the establishment in terms o f the number of classes and the shortage interms of slots. The Government is almost totally uninvolved inthe construction o f highschools andmiddleschools. Apart from the PEPP, localpopulations andcommunities build infrastructures. Table 40: Execution of Delegated Appropriations of the DREBA, Cascades Province, 2003 Designation o f credit lines Initial credit Executed Execution rate amount Various nat'l. mtgsiconflfairs 960 000 959 300 99.9% Office furniture and equipment 3 743 000 3 742 167 100.0% Newspaper and magazine subscriptions 356 000 356 000 100.0% Various activities for organization oftesting and competitive 26 000 000 25 997 400 100.0% examinations, 2003 Maint. Fees Pers. to mtgsitravel costs to 2003 examinations 13 939 000 13 926 900 99.9% Materiel .Transport. Vehicles, 4-wheel. 3 847 000 3 843 175 99.9% Maintenance products - misc. equipment 1288 000 1279 533 99.3% Printed material, data sheets 1500 000 1497 420 99.8% Other specific materials 3 073 000 3 071 776 100.0% Office supplies 6 500 000 6 498 614 100.0% Materiel . transport. Vehicles, 2-wheeled 2 019 000 2 017 446 99.9% Misc. expenditures 1498 000 400 000 26.7% Definitive transportadon costs 1440 000 1437 184 99.8% Total 66 163 000 65 026 915 98.3% Source: DRE3A, Cascades,2003 309. Government intervention in the operations o f high schools and middle schools generally amounts to the provision of teaching personnel and some supplies of poor quality (paper, chalk, ink, stencils, etc.), as well as books (in the first cycle of primary school) which are rented to pupils for CFAF 500 per year. These books are of good quality but too few in number. The Lompolo Kon6 high school is totally without English textbooks for the Sth and 9'classes. Emphasis also needs to be placed on the publication o f books for the second cycle. The high school also receives foodstuffs of poor quality (rice, yam, beans, noodles, oil, tomatoes, etc.). Meals cost the pupilsCFAF 100. - 104- 310. Regarding the managemento f school cafeterias inhighschools and middle schools, food orders are issued inOuagadougou and are then distributed to the various highschools andmiddle schools. The main problem is the low quality o f the products provided, as well as the costs associatedwith transporting these products. There is the case, for example, o f yams produced in Gaoua, transported to Ouagadougou, orderedby the ministry, dispatched to Bobo-Dioulasso and then re-transported to Gaoua for the use of the region's high schools and middle schools. It is therefore crucial that budget allocations (for materiel, equipment and foodstuffs) be delegatedto the DRESSso that the quality o f the material supplied can be improved, transaction costs eliminated, and delivery times reduced. 311. Other sources of financing of decentralized structures. At the level of MEBA's decentralized operational structures, i.e., the schools, there are other contributions to their functioning. These include the contributions o fparents, which are managedby the parenuteacher associations (Associations des parents d 'klhes, APE). They range from CFAF 1,000 to CFAF 1,500 dependingonwhether the school is ina rural or urbanarea. 312. The management of APE contributions is problematic at the regional level. The illiteracy o f APE members is a significant obstacle. There is no nomenclature for APE budgets and they lack managerial expertise. The Treasury should also see that APEs do not use their members' accountsto save APE funds. 313. Some new initiatives have also emerged in the circumscription o f Banfora I,in connection with greater accountability and involvement for women within the APEs. These include the Association of Volunteer Mothers (Association des d r e s kducatives). However, the coexistence ofbothentities within a single school is difficult. 314. At the secondarylevel, other sources of financing ofthe regional directorates include the funds of the high schools and middle schools themselves. These consist mainly of school fees, which range from CFAF 20,000 for first-year student's textbooks to CFAF 5,000 for other pupils. These school fees are distributed in accordance with a 1994 Bylaw stipulating that 75 percent must go to the operation of the institution, 25 percent to the ministry, o f which 5 percent goes to the regional directorate, 10 percent to FONER and 10 percent for outreachisupport activities. 315. The budgets of high schools and middle schools essentially consist o f revenues derived from school fees and parent contributions. The budget covers operating costs, including in particular the payment of contractual employees' salaries (e.g,, watchman, mat~physics/chemistryteacher), substitute teaching costs and materiel. For the 2002-2003 school year, for example, substitute teaching expenses amounted to CFAF 2.851.800, at CFAF 2,OOO/hour at the Lompolo KonC school inBanfora. 316. The APE also contributes to the operation o f high schools and middle schools. Their budget is mainly derived from contributions ranging from CFAF 3,000 - 5,000, depending on the school. The example o f the role of the APE at the provincial school at Lompolo KonC (Banfora) is one that should be extended to the country as a whole, Inthat school, there are two general meetingsof the APE, one at the beginningo f the year and the other at the end, at which - 105 - time it reviews the past year's activity. The school administration attends the APE'S budget session inorder to provide advice. Since most members o f the APE are illiterate, the DRES has also begunto provide them with training inbudget management. Taken together, these initiatives have helped to enhance the transparency of the management of their funds. It was not easy to set up anoperatingmechanismofthis kind, since transparency isdifficult for most APESto achieve. 317. The technical high schools and middle schools also engage in extracurricular and extra- scholastic activities (e.g., gardening, woodworking, cloth dying, etc.). This type of activity is extremely important, especially for rural schools that have very few options for self-financing and receive very little in the way o f funds from the Government. Tn this connection, the provincial high school at Lompolo Kone has set up a botanical garden project that has received in-kindfinancing inthe form o fequipment. 318. Resource transfers at the level of the central administration. Mechanisms exist for transferring resources fiom the operational structures to the administrative or central structures. At the primary school level, the mechanism for transfemng resources is still quite underdeveloped, but the system does exist. For example, within the provincial directorate of Banfora, each school must transfer, out o f parents' APE dues, CFAF 1,500 per class andper year to its circumscription to help cover operating costs. This is necessary because Government resources reachingthe localities are not adequate to ensure their functioning. 319. A t the secondary level, despite the fact that most highschools andmiddle schools receive very little Government support for their operation and maintenance, these schools transfer a substantial portion of their resources to the central administrations, i.e., 25 percent o f revenues fiom school fees, o f which 5 percent is earmarked for the operating expenses o f the DES. In addition, every high school and middle school must transfer between CFAF 35,000 and CFAF 60,000 (depending on the size o f the establishment) to DRES each year to help cover its operating costs. The explanation offered for this was that the remaining 20 percent is distributed -- 10percent to FONERto finance student loans and 10percent for teacher training andsupport. 320. In general, the reforms introduced within the high schools to ensure their autonomous operation have yielded positive results. They receive very little operational support from the Government. On the contrary, they transfer a significant portion o f their resources to the centralized structures, thereby limiting their capacity for autonomous functioning. Since that is the case, reservations are again in order with regard to the Government's capacity to ensure cost- free education at the primarylevel, and to expand it to the first cycle of secondary school. 321. Participation in budget formulation and execution. The level o f involvement of the various units inthe budgeting process is low, in terms of both budget preparation and during the execution o f expenditure, in both ministries o f education and at the central and regional levels. Except for the units that have delegated credit mechanisms, other decentralized units receive very few resources, andthe resources allocated to them are not specified. 322, Partic~ationof MEBA structures in budget f o r ~ u ~ a t i ~MEBA's central and n . decentralized departments are not sufficiently involved in the ministry's budget formulation. However, since the launching of the PDDEB, the process has been moving inthe direction o f far - 106- greater involvement, on the part o f sectoral agencies, in the budget process. The DAF and DEP usually prepare the budget. In 2003, for example, a budget commission met in connection with the preparation of the 2004 budget and convened all o f the ministry's central and decentralized structures (regional directorate, central directorate, etc.). 323. The new instruments usedto prepare the budget are the PDDEB document, the provincial action plans for implementing the PDDEB and the medium-term framework for basic education expenditures, which is still being developed. These instruments are still not well coordinated in terms o ftheir linkages to the budgetingprocess at the sectoral level. 324. At the decentralized level, the regional department of basic education, which does not draw up an action plan at the regional level, helps drawing up the ministry's budget. It also has an allocation under the budget law, whereas the provincial directorates, which develop and implement operational action plans within the framework o f PDDEB, did not have budget allocations under the budget law until2004. 325. Action plans developed at the provincial level are invaluable tools in the preparation o f the sector's program budget. Iftheir quality were enhanced interms o f the participation of local actors (e.g., schools and localities) and their thoroughness (e.g., the extent to which they take into account all activities at the provincial level), MEBA's budget development process could be improved. For example, the content o f provincial action plans must include human resource requirements and costs. In order to strengthen the feeling o f ownership and participation of all actors at the decentralized level, these plans could be reviewed by a regional budget commission within the sector, and could yield a regional action plan for basic education. These action plans also deal with a one-year time horizon at present. They could be broadened to cover three-year periods, thereby facilitating the integration o f local action plans into the ministry's program budget. 326. Participation of MESSRSstructures in budgetfor~ulation.For lack o f a coherent policy for the entire sector, there is not transparency inthe distribution o f resources within the ministry. Most MESSRS units, whether central or decentralized, reported little involvement in the ministry's budget formulation. Some central structures are consulted only inconnection with the distribution o f public investment allocations. They include the Department o f Studies and Planning (DEP) and the EPAs and EPICS(e.g., the University o f Ouagadougou andthe CNRST.) 327. There is no commission charged with arbitrating resource allocations within the ministry mainly due to the short timeframe within which the ministry must develop its budget. The regional directorates for secondary education (DRES) acknowledged that they were only consulted on issues o f personnel expenditure forecasts at the regional level, in the context of preparing each year's budget. 328. Operational entities at the decentralized level, such as high schools and middle schools, are also excluded from the budget preparation process for the sector. They draw up their own operating budgets on the basis o f their own locally-generated resources. The DRESSalso draw uptheir operating budgets on the basis oflocally-generated funds. - 107 - D. CONCLUSIONSANDRECOMMENDATIONS 329. The overall performance of the education system has been improving steadily during the past five years, but the challenge still remains interms of coverage, quality and efficiency. Over half the school age children are out of school. Those who are enrolled are receiving low quality education, resultinginhighrepetition rates and large waste o f the limitedresources available for the sector. There are also serious inequities incoveragebetween urbanand rural areas, between boys and between the poor and rich. While the current sectoral strategies and action plans highlight some of these issues, it is important to ensure budget allocation, execution and managementneed to be better geared towards better education outcomes. The following are the key recommendations: a Consider the whole education sector as apriority sector andnotjust primaryeducation. 0 Optimize investments in schools infrastructures by adopting and implementing the educational map. 330. Improve budget planning and execution. Further improvements in budget planning and execution may include the following: 0 Greater involvement of the regional departments in the budget planning process to ensure their needs are reflected in the budget. This would help in identifying the particular regionalneeds. e Greater transparency in budget discussions through the disclosure o f information and larger participation o fthe relevant departments inthe center and the regions. a ~impli~cationprocurement procedures and credit-related regulations in MESSRS to of avoid unnecessarydelays inbudget execution. a Constraints in donor-financed projects need to be addressedsystematically with donors. Options to simplify and unify disbursement procedures amongst donors need to be explored. 331. Improvebudgetdecentralizationand monitoring. Given the success of the delegated credit system under MEBA, budget decentralization could be further improved through the following: a Extending the delegated credit system to MESSRS. Although decentralized units receive material inkind, the delays inthe bureaucratic procedures at the central level may be avoided if credits are delegatedto the regional units to purchase their material needs. This would eliminate the transaction costs and would avoid delays due to the transportation o f the goods from the center to the regions. In the case of school cafeterias, this will also improve the quality o f the food which could be purchased locally. - 108- Reinforcing the capacity of both MEBA and MESSRS at the regional and provincial levels to manage the implementation o f provincial action plans for improved planning, execution and monitoring o f budgets, through greater skill development and a simplified system ofbudget management andmonitoring o fresults. e Strengthening the accountability of the schools and EPAs through a system o f information recording, and audits to ensure appropriate use o f funds. 332. Improvethe efficiency and equity o f resource allocations Within the sector. It is clear fiom the analysis there are large efficiency and equity gains in the sector through better management o f resources and improvements inthe quality o f inputs. These would include: e Better management of recruitment and deployment of teachers and inspectors. This has to be addressed within a regional and provincial action plan which is comprehensive and complete in order to incorporate all prioritized needs related to the development o f the educational system at the level considered. It would also require putting in place a mapping of vacancies for teachers and inspectors by province and region. A policy o f regionalized recruitment will also need to be instituted at the secondary level in order to limit the strong tendency o f teachers in rural zones to migrate to larger cities such as Ouagadougou andBobo-Dioulasso, including the offer o f more attractive pay. e Teacher training, especially in the sciences (mathematics, physics and chemistry, life sciences) to meet the shortages inthis area and to improve the quality o f education. 333. Cost recovery at higher levels of education. Efforts will need to be made to diminish per-student costs at the university level, and to transfer more resources to the lower levels (ie., the primary and secondary schools). In addition, the rapid growth in enrollments means that a mechanism for contribution and cost recovery in higher education must be established to supplement direct public expenditure. The Government would need to define an appropriate and viable long-term strategy to promote higher education in Burkina Faso on a sustainable basis. Such a strategy must incorporate cost recovery, particularly through the enhancement o f FONER's capacity to recover loans, which can then be re-allocated to students. 334. Improve the qualiq of education. It is necessary to strictly apply the measure regarding sub-cycles and extend it if possible to the first grade o f secondary education. Moreover, the government should develop a program for the provision and distribution o f books for the institutions involved in secondary education. Sufficient financing should be made available for such a program. 335. Increase the period o f time for teaching t the primary and secondary schools by enforcing the rule o f 24 hours per teacher andper week. - 109- ANNEX 1: BURKINA FASOAT A GLANCE Sub- POVERTYandSOCIAL Burkina Saharan LOW- Faso Africa income Developmenidiamond' 2002 Population, midyear (millionst 11.8 688 2,495 Lifeexpectancy GNI percapita (Atlas method, US$) 250 450 430 GNI(Atlas mthWr, US$bi//hsJ 3.0 306 1,072 - Averageannualgrowth, 199642 Population(%J 2.4 2.4 1.9 Laborforce ( X ) 2.0 2.5 2.3 GNI Gross Mostrecentestimate (latestyear available,199642) per +-- PnmW capita nrollment Poverty (% ofmulation belownationalpoverly line) 45 Urban populationI%of total wulatim) 17 33 30 Lifeexpectancyat birth(Yeam) 44 46 59 1 Infantmttality(per 1,000live births) 104 105 81 Childmalnutrition dchildrenunder5) 34 Accessto improvedwater source Accessto an improvedwater source(% of~ulationj 78 58 7% Illiteracy1% ofpopulationage f5+1 74 37 37 Gross primaryenrollment (% dschd-age population) 43 86 95 Burkina Faso Male 51 92 103 Low4ncomegroup Female 35 80 87 KEYECONOMICRATIOSand LONG-TERMTRENDS 1982 1992 2001 2002 Economicratios' GDP (US%billions) 1.6 3.4 2.8 3.2 Grossdomestic investmenWGDP 21.8 18.8 18.4 Exportsofgoodsand secviceslGDP 9.2 9.0 9.2 8.5 Trade Grossdomestic savingvrGDP -3.9 9.3 4.9 5.3 GrossnationalsavingsiGDP T 16.3 7.7 7.8 i Currentacmnt balancdGDP -5.5 -11.7 -11.0 lnteresfpaymenWGDP 0.5 0.4 0.2 0.2 TotaldebWGDP 22.6 30.8 47.3 52.2 Totaldebt servicelexports 6.7 7.4 23.1 19.1 I Presentvalueof debWGDP Presentvaiueof debvexports Indebtedness 1982-92 199242 2001 2002 200248 (average annualgrowth) GDP 5.0 5.8 5.9 4.4 BurkinaFaso ~~ 5.6 GDP percapita 2.5 3.3 3.4 1.9 3.2 Low-Incomegroup Exportsof goodsandservices 2.2 1.8 3.0 11.7 14.2 STRUCTUREof the ECONOMY 1982 1992 2001 2002 Growthof investment andGDP (Y.) (% of GDP) Agriculture 32.1 29.6 31.4 30.8 Industry 20.9 20.8 19.3 18.7 Manufacturing Services 47.0 49.5 49.3 50.3 Private consumption 88.3 68.5 82.9 81.6 1 97 98 39 M) 01 02 Generalgovemmentconsumption 15.6 22.2 12.2 13.1 i Importsof goodsandservices 33.4 21.4 23.1 21.6 6----GDI -0'GDP 1 1982-92 199242 2001 2002 1 (average annualgrolvlht Growthof exportsand imports(XI 7 Agriculture 3.8 4.9 4.1 1.7 60 Industry 4.5 4.7 7.6 8.5 40 Manufacturing 2.2 Services 5.4 6.6 6.0 6.9 20 Privateconsumption 3.2 3.9 6.2 1.7 0 Generalgovernmentconsumption 2.2 -1.3 2.5 2.5 d o Gross domestic investment 2.1 9.7 10.5 10.6 Importsof goodsandservices 3.5 2.4 3.1 -3.1 - ~~ Note 2002 data are preliminaryestimates 'Thediamondsshowfourkeyindicatorsinthecountry(inbold)comparedwithitsincome-groupaverageIfdataaremissing,thediamondwill beincomplete - 110- 3 TRADE I4 , - 113- REFERENCES Fonds monetaire international, Revue du respects des Standards et des Codes (Module Transparence buggktaire), Rapport du FMIsur les pays 021142,2002 Diamond, J., Performance Budgeting, Managing the Reform Process, Document de travail du FMIWP103133. Lallement, D., McMillan,D., O'Sullivan, K., Plane P., et Savadogo, K., Evaluation du cadre de dkveloppement intkgre (CDI). Etude de cas :le Burkina Faso, septembre 2002 Ministbe des Finances et du Budget, Plan d'Action pour le renforcement de la gestion budgktaire, Ouagadougou, 2002. Banquemondiale, Burkina Faso :ktude de cas sur les aspects budgktaires de la mise en oeuvre du CSLP, Institut de laBanquemondiale 2004. Banquemondiale, Evaluation de la responsabilitk~nanci~redupays, 2002. Banque Mondiale. 2003. Sante et Pauvreteau BurkinaFaso. Washington, DC. Institut National de la Statistique et de la Demographie (INSD). Enquete Prioritaire 2003, Ouagadougou :2003. Institut National de la Statistique et de la Demographie. - Enquete Prioritaire 1998, Ouagadougou : 1998. Institut National de la Statistique et de la Demographie (INSD). Enquete Demographique et de Sante : 2003, Ministere de 1'Economie et du Developpement et Macro Intemational, Ouagadougouet Calverton, MD.2004 InstitutNationalde la Statistique et de laDemographie(INSD).. EnqueteDkmographique et de Sante : 1998-99, Ministere de 1'Economie et des Finances et Macro Intemational, Ouagadougouet Calverton, MD. 1999 InstitutNationalde la Statistique et de laDemographie (INSD).. EnqueteDemographiqueet de Sante : 1993, Ministere de 1'Economie et des Finances et Macro Intemational, Ouagadougouet Calverton, MD. 1993 INSD. 2003. Enquete sur le Suivi du CSLP : Services Dkconcentres duMinistere de la Sante et duMinisterede 1'Enseignement de Base et de l'AlphabCtisation, ))Ouagadougou. INSD. 2003. <(Enqutte sw: les Ecarts Entre Les Prix Unitaires d'un Certain Nombre des Produits Obtenus par les Marches Publics et Ceux Obtenus par le Secteur Prive, )) Ouagadougou. - 112- Burkina Faso, Ministbre de l'enseignement de base et de l'alphabktisation, annuaires statistiques de 1998 a 2003. Burkina faso, Ministbre des enseignements secondaire, supkneur et de la recherche scientifique, annuaires statistiques de 1998 A 2000 et donnkes globales de 2002. Burkina Faso, Ministbre de 1'Economie et des Finances. -Loi de Finances du budget de 1'Etat - Gestion 1998 a 2003. Burkina Faso, Ministere de 1'Cconomie et des finances, Direction de la solde et de l'ordonnancement, situations des agents de 1'Etat par catkgone et par section, janvier 2003. Burkina Faso, Ministkre de l'kconomie et des finances, Bilan du programme d'investissement public(PIP), de 1998 a 2002 . World Bank, African developpment indicators, 2003. - 113 -