Zanzibar Health Sector Public Expenditure Review

Transcription

1 Zanzibar Health Sector Public Expenditure Review 2008

2 ii Zanzibar Health Sector Public Expenditure Review Acknowledgements The 2008 edition of the health sector Public Expenditure Review for Zanzibar for the first time has been undertaken as an in-house project. It is part of the Plan of Action of the Health Sector Reform Secretariat and was carried out by Mr Kai Straehler-Pohl and Mr Abdul-Latif Haji of the Planning Unit. Ms Zainab Choum assisted on some occasions. The exercise was funded through the Zanzibar Health Sector Programme Support (HSPS) office. We would like to thank everyone who has assisted the team in making available and analysing data, providing information, or sharing their knowledge on achievements and constraints facing the MoHSW. In particular, we would like to thank the head and the staff of the Accounting Unit, of the Personnel Unit, and the Training Unit, as well as the accountant of the HSPS office. We would also like thank all those programme managers and their staff who provided information for this year s PER edition. We would also like to thank all Development Partners who have made additional information available. Useful feedback was received during a meeting of the MOHSW Senior Management, and from members of the Health Sector Reform Secretariat. Responsibility for errors in interpretation or fact remains with the authors.

3 List of Tables iii Contents List of Tables... iv List of Figures... v Acronyms... vi Selected Glossary... vii Executive Summary... ix 1. Introduction Background and Context Objectives Structure of the Document Progress on Recommendations from the PER Overview of the Health Spending in Zanzibar Financing of the Health Sector in Zanzibar Estimated Total Health Spending RGoZ Health Budget and Expenditure Health Spending of the MoHSW Sub-Sectoral Spending Detailed Analysis of Recurrent Spending Spending by Location External Financing of the Health Sector Total Volume Analysis of External Finances Cost Sharing in Public Health Facilities Total Income Current Charging Practices Protection Mechanisms The Budget for FY08/ Domestic Funding External Funding Key Findings Spending and Stated Priorities Financial Flows, Management and Monitoring Expanding the Revenue Base Recommendations References Appendices Appendix 1: Data Sources Appendix 2: Population, Inflation and Foreign Exchange Information Appendix 3: Additional Measures of MoHSW Share of RgoZ Spending Appendix 4: External Finance Reports from Programmes Appendix 5: Spending on Operation of Health Facilities... 49

4 iv Zanzibar Health Sector Public Expenditure Review List of Tables Table 1: Progress on Recommendations of PER Table 2: Total estimated health spending (nominal), FY06/07-FY07/ Table 3: Nominal MOHSW spending, FY2003/04 FY2008/09 (TSh 1m)... 4 Table 4: Summary measures of MOHSW budget, FY03/04 - FY08/09 (TSh m)... 5 Table 5: Summary measures of MOHSW expenditure, FY03/04 - FY08/09 (TSh m)... 6 Table 6: Top Receiving MDAs, budgets and expenditures FY05/06 FY07/08 (recurrent vs recurrent excl. CFS)... 9 Table 7: Health as % of Government spending in SSA... 9 Table 8: United Republic of Tanzania s GBS funds to MoHSW Zanzibar Table 9: MoHSW recurrent and development Budgets FY03/04-08/ Table 10: Budget performance: recurrent and development spending FY03/04 FY08/ Table 11: MOHSW recurrent spending on PEs and OC, FY03/04 FY08/09 (TSh m) Table 12: Budget performance: PE and OC, FY03/04 FY07/ Table 13: Crude expenditure breakdown by category, FY07/ Table 14: RGOZ spending on training activities, FY2004/05 FY08/09 (Tsh m) Table 15: RGOZ allocations to Medical supplies and services, FY03/04 FY08/ Table 16: RGOZ allocations to Medical expenses abroad, FY04/05 FY08/09 (TSh m) Table 17: Estimated allocations for running of health facilities, FY07/08 (TSh m) Table 18: Spending in Unguja and Pemba FY2003/04 - FY2007/ Table 19: DP Budgets for MoHSW Development Budget, FY03/04 FY07/08 (Tsh 1m) Table 20: DP Actual Expenditure for Development Programmes, FY06/07 FY07/08 (Tsh m) Table 21: External Finances of Programmes, FY2005/06-07/08 (TSh 1m) Table 25: External Funds: Budget, Disbursements and Expenditure FY07/ Table 23: External Finances by Source, FY2006/07-07/08 (TSh1m), DP overview data Table 24: External Finances by Donor, FY2006/07-07/08 (TSh1m), DP overview data Table 25: External Funding by Geographical Area, FY206/07-07/08 (TSh1m), detailed data on spending by location Table 27: External Funding by Programmatic Area, FY2006/07-07/08 (TSh1m), based on funding overview by DP Table 28: External Funding by Programmatic Area, FY2006/07-07/08 (TSh1m), based on detailed spending by location Table 28: External Funding by Category, FY2006/ /08 (Tsh1m) Table 30: HSF by level and zone, FY2006/07-07/08 (TSh1m) Table 30: Projected income of the MoHSW, MoFEA estimates (TSh 1m) Table 31: Actual income from Cost Sharing in FY2007/08 (TSh) Table 33: Charges at hospitals for selected services, as of December 2008 (TSh) Table 34: Draft budget for FY07/08 (TSh 1m) Table 35: MoHSW as % of RGoZ, FY08/09 (TSh1m) Table 36: External Funding by source and programmatic area FY08/09 (TSh1m) Table 37: Main recommendations Table 38: Specific recommendations Table 39: Population Data Table 40: Inflation figures FY02/03-08/ Table 41: Foreign Exchange Rates FY03/04-08/ Table 42: MOHSW share of RGOZ spending, various measures... 46

5 List of Figures v Table 43: Reporting issues at PDUs Table 44: Allocation of PEs to levels of care, FY07/ List of Figures Figure 1: Total expenditure in nominal terms by source, FY04/05-FY07/08 (TSh bn)... 2 Figure 2: Total Expenditure in real terms by source, FY04/05-FY07/08 (TSh bn)... 3 Figure 3: Shares of total expenditure by source, FY04/05-FY07/ Figure 4: Nominal MOHSW approved budget and expenditure, FY2003/04 FY2007/ Figure 5: MOHSW budget in nominal and real terms, FY2003/04 FY2007/ Figure 6: MOHSW expenditure in nominal and real terms, FY2003/04 FY2007/ Figure 7: MOHSW expenditure as % of approved budget, FY2003/04 FY2007/ Figure 8: MOHSW budget as share of RGOZ total, FY2003/04 FY2007/08 (%)... 8 Figure 9: MOHSW expenditure as a share of RGOZ total, FY2003/04 FY2007/08 (%)... 8 Figure 10: Health as % of Government spending in SSA (incl. debt service)... 9 Figure 11: Estimated allocation of PEs between activity types, FY07/ Figure 12: Staff by service area and skill level, FY07/ Figure 13: Salaries by service area and skill level, FY07/ Figure 14: Estimated RGOZ spending on running costs of health facilities (budgets), by level, FY07/ Figure 15: Budgets by Zone, FY2007/ Figure 16: Expenditure by Zone, FY07/ Figure 17: Budget performance for Other Charges, Unguja and Pemba, 2003/ / Figure 18: Estimated allocation of salary spending by geographical area, FY2007/ Figure 19: External Budget by Location, FY07/ Figure 20: External Expenditure by Location, FY07/ Figure 22: External Recurrent Expenditure FY07/ Figure 21: External Capital Expenditure, FY07/ Figure 23: Shares from CS income, FY07/ Figure 24: Underfunding of primary health care... 39

6 vi Zanzibar Health Sector Public Expenditure Review Acronyms ADB AIDS bn CFS CHS Danida DHMT EPI FHRP FY GAVI GBS GF GoT RGoZ HIV HSF HSPS HSS IMCI m MCH MDGs MDRI MEA MKUZA MMH MoFEA MoHSW MSS MTEF NBS OC OCGS PE PER PHCC PHCU PER POA RCH SMMRP TSh USD USAID WHO ZACP ZAJHSR ZHSRSP II ZMCP African Development Bank Acquired immune-deficiency syndrome billion Consolidated Fund Services College of Health Sciences Danish International Development Assistance District Health Management Team Expanded Programme on Immunisation First Health Rehabilitation Programme Financial year Global Alliance for Vaccines and Immunisation General Budget Support Global Fund for AIDS, Tuberculosis and Malaria Government of Tanzania Government of Zanzibar Human immunodeficiency virus Health Service Fund Health Sector Programme Support (Danida) Health Systems Strengthening Integrated Management of Childhood Illness million Maternal and Child Health Millennium Development Goals Multilateral Debt Relief Initiative Medical Expenses Abroad Mkakati wa Kukuza Uchumi na Kupunguza Umasikini Zanzibar (Zanzibar Strategy for Economic Growth and Poverty Reduction) Mnazi Mmoja Hospital Ministry of Finance and Economic Affairs (Zanzibar) Ministry of Health and Social Welfare Medical Supplies and Services Medium Term Expenditure Framework National Bureau of Statistics (mainland) Other charges Office of the Chief Government Statistician (Zanzibar) Personal Emoluments Public Expenditure Review Primary Health Care Centre (cottage hospital) Primary Health Care Unit Public Expenditure Review Plan of Action Reproductive and Child Health Support to Maternal Mortality Reduction Programme (ADB) Tanzanian shillings United States dollars United States Agency for International Development World Health Organisation Zanzibar AIDS Control Programme Zanzibar Annual Joint Health Sector Review Zanzibar Health Sector Reform Strategic Plan II Zanzibar Malaria Control Programme

7 Selected Glossary vii Selected Glossary Absorption capacity: Absorption capacity measures actual expenditure as a percentage of funds available, ie generally of releases or disbursements. It is an indicator of the planning and implementation capabilities of the spending organisation, although it should be borne in mind that external factors such as delays in disbursement can also affect ability to spend within a given period. Budget performance: Budget performance measures actual releases or disbursements as a percentage of the approved budget. It is a measure of budgetary execution or discipline, as it shows the extent to which the budgetary commitments are being met as a whole. It can also be seen as measuring effective priorities, with variations in budget performance between different components of the budget serving as an indicator of their effective relative priority. Consolidated Fund Services: Consolidated Fund Services are those items which have prior claim on the RGOZ budget, notably interest on debt and repayments of the principal on debt, both domestic and foreign. It also includes other items such as election expenses, pension contributions, and the special contingency fund. Nominal: Nominal budget and expenditure figures are given using the value of the currency at the time of announcing the figure (ie current prices). Real: Inflation reduces the value of currency in terms of what can be bought with it (ie the real value). If one is interested in what can be purchased with the budgets or expenditures of different years, it is necessary to account for inflation from year to year. Example: The nominal budget rises by 10% between Year 1 and Year 2 (TSh 1.0m to TSh 1.1m) and inflation is 5%. To compare the purchasing power of Y2 s budget with Y1 s budget, the Y2 nominal figure has to be deflated by the inflation rate to give the budget in Y1 prices. This is done by dividing by the index of the new price level (105) and multiplying by the old level (100). 1,100,000/105 x 100 = 1,047,619 This gives a figure of TSh 1.048m shillings, ie the 10% nominal increase is only a 4.8% real increase. When analysing the real expenditure, it is important to note the base-year. The nominal budget for a given year in real figures will take on two different values if two different base years are used. This basically means that we are more interested in trends than in values when discussing real figures. In this report, timelines use two different base years to make use of a maximum of information available. Previous PERs provide estimates for external funding back to FY04/05, so for summary figures FY04/05 is the base year. For RGoZ figures, usually five years have been taken into account, so that the base year is FY03/04. The absolute values for RGoZ real spending will therefore differ between the Summary and the RGoZ sections. This will change for next year, when information on five years of external funding will be available. Government Spending: Government spending is the term used for all budgets and actual expenditures from RGoZ s own funds, ie excluding any Development Partner s contribution. Public Spending: Public Spending is used to include all budgets and actual expenditures funded by public sources, which includes spending by development partners (even international NGOs). It does not include spending by private individuals or local NGOs (unless their funding came from a public source).

8 viii Zanzibar Health Sector Public Expenditure Review Vertical Programmes / Funding: In most developing countries health systems are set up with a central Ministry of Health (MoH) that coordinates the health sector and that has direct control over central institutions, health facilities and a number of general health services, and some priority programmes that coordinate the fight against important (public) health threats, most typically HIV/Aids and Malaria. These priority programmes often enjoy considerable independence from the central MoH, both in administration and funding, with the latter most often coming from development partners (often in large amounts). Because of the independence in its operations from the bottom, i.e. operational level, to the top, its administration, these programmes and their funding are often called vertical. There is a heated debate of the overall health systems efficiency of vertical programmes and funding. Health Systems, Health Systems Strengthening, and Horizontal Funding: The World Health Organization (WHO) defines health systems as follows: A health system consists of all the organizations, institutions, resources and people whose primary purpose is to improve health (WHO n.d.). Health system strengthening (HSS) can be defined as any array of initiatives and strategies that improves one or more of the functions of the health system and that leads to better health through improvements in access, coverage, quality, or efficiency (Islam 2007). The defining characteristic of HSS is exactly that it strengthens the system, with benefits for a variety or even all parts of the health system, instead of only one disease. In this sense, funding for HSS is sometimes called horizontal.

9 Executive Summary ix Executive Summary Introduction The Health Sector Public Expenditure Review 2008 (PER08) updates the information on the financing and spending in the public health sector provided in previous PERs. It is addressed at decision makers in the Ministry of Health and Social Welfare (MoHSW) and the Revolutionary Government of Zanzibar (RGoZ) and meant to provide the evidence needed for informed decision-making with regard to allocation of funds, efficient and equitable health care financing, and transparent and accountable management and planning. It is also meant to provide development partners (DPs) with an overview of financing and spending that will help them in coordinating their efforts. The PER08, like its predecessors, is constrained by the availability and quality of the data made available to the PER team. Some of these constraints are due to weak monitoring and accounting systems, others are due to capacity and time constraints of reporting staff. Especially the information on external financing is still incomplete, inadequate and inconsistent. Not all recipients of external funds report on their funding, the level of detail of the reports is not sufficient, and the differences of the reports to different recipients are all worrying, especially since external funding amounts to more than 50% of total funding. Overview of spending The Zanzibar Public Health Sector has three sources of funding: RGoZ funding, external funding, and income of health facilities through user-fees. The total resource envelope (expenditure) in FY07/08 was TSh20bn, or USD16.4m, which equates to TSh16,578 or USD13.7 per capita. RGoZ contributed 40% of the total, DPs 59%, and the remaining 1% has come in through user-fees. The resource envelope is not sufficient according to WHO estimates for a full package of essential interventions of USD34, but it is well above the estimates for the own Zanzibar Essential Health Care Package (EHCP) at primary level of USD3.63. It is difficult to make definitive statements about the development of the resource envelope because the level of missing information may differ over the years. The figures reported for FY07/08 have shown a recovery from a drop in FY06/07, and in nominal terms have surpassed those of FY04/05 and FY05/06. In real terms, however, the total funding in FY07/08 was lower than in FY05/06. RGoZ Health Budget and Expenditure The total budget of the MoHSW in FY07/08 was TSh9,184m, and the expenditure TSh7,967m. Per capita, the spending amounted to TSh7,696 and TSh6,676, respectively. The budget performance of the MoHSW was 86%, lower than in the previous two years, but still higher than in previous years. The budget has seen a considerable jump up from FY06/07, until when budgets and expenditures only increased moderately. The larger than usual increase in the budget also meant that for the first time in recent years, the real budget is has increased. The real expenditure has increased slightly over FY06/07, but this only means that it is basically back at the value of FY05/06. The share of RGoZ resources that were allocated to the health sector was 7.6% for the budget, and 8.7% for the expenditure, using total local spending and excluding the Consolidated Fund Services (CFS). With these shares, the MoHSW was the third biggest budget item of the RGoZ and the second largest recipient of actually expended funds (behind the Ministry of Education and Vocational Training, MoEVT). The share for the equivalent figures for recurrent spending only was 8.0% and

10 x Zanzibar Health Sector Public Expenditure Review 9.3%. This falls way short of what other SSA countries spend on health, the 15% that the Government of Tanzania (GoT) committed to under the Abuja declaration, and the 12% that the MoHSW included as a target in its Health Sector Reform Strategic Plan II. Part of the RGoZ budget is financed through the General Budget Support GBS that DPs provide to the Government of the United Republic of Tanzania (GoT). RGoZ receives 4.5% of the total GBS. This money is part of the general revenue and is spent in the same way. This means that the percentage of total RGoZ spending on health can be applied to the GBS. Like this, it is estimated that TSh2,3bn (25.3%) of the MoHSW spending was financed through GBS. Analysing the MoHSW spending in FY07/08 in detail, the following points can be observed (budget/expenditure, TSh if not otherwise indicated): The budget is heavily tilted towards recurrent expenditure; recurrent spending accounted for 91%/95% of the total. Within recurrent expenditure Personal Emoluments (PE) account for 82%/89% of the spending. This means that 75%/84% of spending is on staff. The budget performance of PE is also much higher than that of Other Charges (OC), 98% vs 57%. Most of the MoHSW spending is in the area of Curative Services (2.1bn, 46%), followed by Preventive Services (1.1bn, 24%) and Management and Administration (975m, 21%). The remaining 10% are shared between Social Welfare, Substance Abuse and the Chief Government Chemist. The staff of the MoHSW is split 75% to 25% between medical and other staff. Two thirds of medical staff is skilled, and only very few are un-skilled. Among other staff, only 20% are skilled, while nearly a third is unskilled. Skilled staff earn 72% of the total salaries, of which 67% are for medical staff (50% of total staff), and 5% for others (5%). Spending on training to increase staff capacity is fairly low, at 153m/114m. This is just 10%/13% of total OCs and only 2%/2% of total recurrent expenditure. Medical Supplies and Services received only 263m/109m, which translates into a budget performance of only 42%. It is only 17%/13% of total OC spending, or 3%/1% of total recurrent spending. Medical Expenses Abroad received 82m/43m, a budget performance of 53%. Spending by type of health facility is heavily in favour of Mnazi Mmoja hospital, which captures 2.3bn/2.3bn, or 52%/52% of the total. Other hospitals receive 31%/29%, and PHCUs 24%/23%. For each facility type, the spending is largely PEs and very little OCs. The spending is shared 5.7bn/5.3bn for Unguja and 2.7m/2.3 for Pemba, or 68%/70% against 32%/30%. Unguja s share is higher for expenditure than for budgets, because of a higher budget performance in Unguja than Pemba. The population is distributed 62% and 38% in Unguja and Pemba. At the same time, Unguja hosts most of the central level administration. External Financing The estimates of the external finances available to the public health sector are less comprehensive and precise than the figures for the government, because no comprehensive data-collection system exists that would routinely capture all funds. The MoFEA collects information through the MoHSW from the disease-specific programmes, but not from other departments and units of the MoHSW, and also not about the source and use of funds. While the MoFEA information was analyzed, the PER team also collected its own, more detailed data. The different reports do not always match and it is not certain, which numbers are correct in case of mismatches. Some information is missing due to

11 Executive Summary xi non-responses of programmes, departments and units (DPUs), and some due to non-availability of data at the central or peripheral level. The totals reported for budget/expenditure are reported below (all TSh): DP funding through the development budget: 4.8bn/2.5bn Monthly programme reports to MoFEA: 8.9bn (disbursements) PER08 reporting (totals do not coincide due to missing bits of information and reporting inconsistencies the PER team was not able to clear up): o Information by development partner: 19.9.bn/13.1bn (disbursements) o Information by location: 12.3bn/11.9bn o Information by spending category: 12.3bn/11.9bn It is difficult to assess which information is most accurate for the individual reports, but the data collected for the PER is the most detailed and most comprehensive, and has been used in deriving the total resource envelope. The PER data collection also provided a chance to analyse DP funding on a more detailed level, i.e. by funding from DP, spending by island and by spending category. The lack of ZACP information on budgets for FY06/07 and 07/08 makes the comparison between budgets and expenditures look very favourable. Looking at the funding, the largest category of donors is International Organisations (budget/disbursements: 10.1bn/5bn), followed by bilateral support from Denmark (2.5bn/2.9bn) and the US (3.5bn/2.7bn). The largest individual DPs are the GF (6bn/3.8bn), Danida through HSPS (2.5bn/2.9bn), and PEPFAR (3bn/2.3bn). Looking at the spending, it is possible to analyse it by location, i.e. the split between Headquarters (5.6bn/3.4bn), Unguja (4.0bn/5.3bn), Pemba (2.6bn/2.7bn) and non-allocated funds (116m/418m); by programmatic area, where Malaria comes out on top with 6.4bn/4.2bn, followed by Health Systems Strengthening with 6.7bn/4.2bn, and HIV/Aids (n.a./2.2bn); and finally by expenditure category, which shows that the largest part of the funding is going into recurrent expenditure (10.1bn) and only a much smaller portion for capital spending (1.7bn). Cost-Sharing Cost-sharing, i.e. user-fees, is practiced in all hospitals in Zanzibar, and also in some Primary Health Care Units. The total amount collected in user-fees is about TSh167m, or about 1% of the total resource envelope. Of the total, about TSh150m is being collected at Mnazi Mmoja Hospital, most of the rest at the District hospitals and only relatively small amounts at Primary Health Care Centres. At the same time, there is still no official MoHSW guideline operational that would standardize charges, procedures, and accountability, or would provide a common social protection to all citizens. The resulting system is therefore fragmented, at times illogical (e.g. fees at lower levels in a few cases being higher than at higher level facilities, and priority interventions like facility-based deliveries being charged for in some places and not in others), and opaque in management. There are plans to change this by passing existing draft guidelines on cost-sharing. However, on the technical level, the guidelines also still lack a clear strategy on social protection, i.e. waivers for the poor, and risk-sharing. It is recommended that the MoHSW prepare a Health Care Financing Strategy for the mid- to longterm that has strong political backing and can guide the technical staff in developing the funding methods, in order to arrive at the desired mix of funding.

12 xii Zanzibar Health Sector Public Expenditure Review The budget FY08/09 The total budget envelope for FY08/09 is TSh11.1bn, TSh9.55bn in the recurrent and TSh1.55bn in the development budget. All departments except Mnazi Mmoja increased their nominal allocations. The overall increase in nominal terms is 11.2%, which translates into a decrease of 9.1% in real terms, based on average inflation rates from July 08 to February 09. The budget had been prepared based on a projected 9.7% inflation rate, which would have given a 1.4% increase,. The FY08/09 budget includes funds made available through the Multilateral Debt Relief Initiative (MDRI). These funds are meant to free up spending on debt payments to be used in additional efforts aimed at reaching the Millennium Development Goals. It is therefore correct and necessary that the MoHSW receives parts of the MDRI spending that is made available to RGoZ. Unfortunately, a closer investigation reveals that the MDRI funds have largely replaced other RGoZ funds, instead of adding to them. On the bright side is that the MoHSW increased its share of total RGoZ spending, from 7.6% to 8.2%. This is largely due to an increased share in the development budget. Regarding external funds, a total of between TSh10.3bn and TSh10.5bn is expected, the uncertainty coming from inconsistent reporting.

13 Introduction 1 1. Introduction 1.1. Background and Context The Public Expenditure Review (PER) is viewed as an integral component of the budgeting and reporting cycle in Zanzibar according to the Zanzibar Strategy for Growth and Poverty Reduction (known as MKUZA, RGoZ 2007). The health sector PER aims to assess how recent budgetary allocations match with the stated strategic objectives of the sector, in order to provide recommendations to feed into the budget process. The overall purpose is to ensure efficient and effective use of scarce resources by strengthening the planning, budgeting and allocation functions within the sector. The health sector for the purpose of this PER includes the Ministry of Health and Social Welfare together with public facilities at central and district levels in terms of the institutions, and covers government funds, external financing, and cost-sharing at public health facilities. External finances are included as long as they were channelled through the public health system, whether they come from public funds, i.e. government contributions whether through bilateral or multilateral channels, or private, e.g. NGOs and foundations. Similarly, cost-sharing is included as it provides funds for the public health system, despite being funded privately by the Zanzibari population. The RGoZ s has made social services, of which health care is one, a priority for reducing poverty and stimulating economic growth in MKUZA. It remarked that *f+or social services to be available [RGoZ and stakeholders] require adequate resources to invest in those services (RGoZ 2007, p.34). The Health PER08 attempts to assess in how far the RGoZ and development partners (DPs) have lived up to this responsibility. Overall health sector priorities are identified in the Zanzibar Health Policy, the Zanzibar Health Sector Reform Strategic Plan II 2006/ /11. Activity planning in the MoHSW is done through a comprehensive annual Plan of Action (PoA), while RGoZ funded activities are also included in the Medium Term Expenditure Framework (MTEF). The PER also assesses how stated aims were pursued by making necessary funds available Objectives The 2008 health sector PER updates the findings of the 2007 report, and provides a summary of recent financial allocations and budgetary performance within the sector. The PER08 is the first PER conducted in-house, in the Planning Unit and as part of the Health Sector Reform Secretariat s (HSRS) PoA activities. The scope of the PER08 reflects the ToR of the previous two PERs. The main objectives are: To provide an updated assessment of the level and allocation of public financing to and within the health sector. To the extent possible, given acknowledged data constraints within the sector, this will include both RGOZ and external funding. To review the extent to which existing resources government and external are deployed in line with stated priorities as reflected in policy and strategic plan documents; To provide a preliminary analysis of the feasibility of planned sectoral activities given the available resource envelope; To provide support to the implementation and monitoring of the Zanzibar Strategy for Economic Growth and Poverty Reduction (MKUZA);

14 2 Zanzibar Health Sector Public Expenditure Review To propose further pieces of analytic work to inform fiscal policy dialogue on Zanzibar s revenue effort, tracking expenditure and service delivery, and coherence of the development budget with PRS priorities. Not all these objectives could be met, due to both time and data constraints, and recommendations are made accordingly Structure of the Document The PER08 has kept the structure provided by the previous PERs, with new information added in selected subsections. Section Two provides an overview of the financing and total spending in the public health sector in Zanzibar. It summarizes the more detailed information provided in subsequent chapters. Section Three deals with RGoZ funding. The analysis is based on the Budget Books published by MoFEA, Itemized Appropriation Accounts from the Accounting Unit, and Nominal Roll information from the Personnel Unit. The information allows for a reasonably detailed analysis of funding. Section Four looks at the external funding of the public health sector. Analysis is based on three different sources, the Capital Expenditure as per MoFEA published budgets and MoHSW appropriation accounts, reports to the MoFEA, and the PER08 reporting formats. Section Five looks at cost-sharing. Estimates for the funds collected are provided, charging mechanisms and social protection discussed to provide a limited update on the information collected for the Rapid Assessment of Cost Sharing (Lake 2007b). The budget for FY2008/09 is discussed in Section Six. Section Seven provides the conclusions and recommendations of the PER08 team.

15 Introduction Progress on Recommendations from the PER 2007 Table 1 reviews the progress on the recommendations included in the PER07. Overall, there has been progress on the recommendations related to the PER itself, while progress on broader topics has been limited. Table 1: Progress on Recommendations of PER07 Recommendation Responsible Achievement MOHSW to continue lobbying for greater share of RGOZ budget, using the PER data (and Abuja target) as ammunition RGOZ contributions for the running of health services (ie PHCUs through DHMTs, and hospitals) to be clearly identified in the official printed estimates as well as the MTEF, to be based on objective allocation criteria, and to be channelled through the ZMO accounts in the same manner as Health Service Fund monies. All recipients of external financing should be reminded of their obligation to report completely and in a timely manner as required by MOFEA TWG on Health financing (TWG HF) to review proposals for categorising expenditure, geographically and by expenditure category, in order to facilitate analysis in future PER updates The possibility of expanding routine data collection on external financing to include such categorisation should be explored TWG HF, together with TWG on Sector Performance Monitoring (TWG PM) to agree format for quarterly or semi-annual review of both technical and financial POA implementation Detailed analysis of external funding to be undertaken, either during the FY or as part of the next PER update, looking at geographical spread of spending Urgent action to be taken on the recommendations of the rapid assessment of cost-sharing in order to clarify official policy in this area, and to strengthen management (and thus transparency and accountability) of existing practices Further work to be undertaken to develop a comprehensive health financing strategy for Zanzibar Principal Secretary Planning Unit, with Chief Accountant Principal Secretary TWG HF TWG HF TWG HF and TWG PM TWG HF & Unit Principal Secretary (through TWG) TWG HF Not achieved. MoHSW allocation strongly increased (also real), but share of RGoZ budget declined Not achieved. Pemba has comprehensive budgets for all hospitals in MTEF, but Unguja does not; PHCUs are not included in MTEF; no objective allocation criteria in use; work ongoing to establish subvote for District Health Services that meets requirements Partly achieved. Planning Unit and MoFEA Dept. External Finances still report delays and gaps, but reported some improvements; Planning Unit needs stronger data security Achieved. Based on the work of the consultant for the PER07, TWG HF collected detailed information on expenditure categories and location of spending; format to be used for next PER to improve quality of reporting Not achieved. Possibility explored, but so far no attempts on integration into routine reports Achieved. Reporting format based on PoA planning format including comprehensive information to be used for PoA review in January 2009; no reporting on categories, however, as regarded too complex by TWG PM Achieved. Based on the work of the consultant for the PER07, TWG HF collected detailed information on expenditure categories and location of spending; format to be used for next PER to improve quality of reporting Not achieved. Guidelines for cost-sharing in public health facilities developed, but unresolved issues remain (especially on social protection) and guidelines not implemented. Support at political level weak; in light of PER08, MoHSW position on cost-sharing in general should be re-evaluated Not achieved. No work on this topic as to date. Plans to integrate this task into potential review of Health Policy

16

17 Overview of the Health Spending in Zanzibar 1 2. Overview of the Health Spending in Zanzibar 2.1. Financing of the Health Sector in Zanzibar The Zanzibar public health sector is funded through a mix of financing mechanisms, including general revenue of the Revolutionary Government of Zanzibar (RGoZ) disbursed to the Ministry of Health and Social Welfare (MoHSW) through the Ministry of Finance and Economic Affairs (MoFEA), external funding through development partners (DPs), and cost-sharing, i.e. user-fees from patients. Similar to many SSA countries, the government finances only account for a minority share of spending in the health sector. Nevertheless, it still provides the core funding for the sector through its funding of the health care workforce. Similarly, notwithstanding issues with the disbursement of funds for complementary inputs (Other Charges - OCs) in any given year, it provides one of the most reliable sources of funding for planning in the medium to long-term. External funding through DPs provides for the larger share of the health sector spending in Zanzibar, and without it, the system would be dysfunctional due to a lack of complementary inputs into the provision of health care. But the reliance on DP funding also poses some challenges. It is difficult to plan for the medium- and long-term, as some donors provide only annual budgets, and it increases the complexity of coordinating different sources of funds to achieve an overall efficient allocation of funds. The information on external funding presented here is still only partial; one major gap is the exclusion of in-kind contributions, another is incomplete reporting from recipients. The largest donors are international organisations and initiatives (especially the Global Fund to fight Aids, Tuberculosis and Malaria (GF)), the US Government (through its Malaria (PMI) and HIV/Aids (Pepfar) initiatives), and the Danish Government (through Danida). Compared to Tanzania mainland, there is a notable absence of direct European bi-lateral donors (except for Denmark). 1 Governments that have sent medical teams, which are not included in this PER, are Cuba, China, and Egypt. Cost-sharing makes up for only a small fraction of funds, mostly collected in the main referral hospital, Mnazi Mmoja (MMH). Important policy documents (the Zanzibar Strategy for Economic Growth and Reduction of Poverty MKUZA, the Zanzibar Health Policy ZHP, and the Zanzibar Health Sector Reform Strategic Plan II ZHSRSPII) endorse the use of cost sharing, but the potential of user-fees as a financing mechanism is not clear. The absolute revenue is small and a cost-benefit calculation might not be favourable due to the administrative costs of a well-functioning system. It will also be difficult with user-fees to guarantee the equitable access to health care to which the RGoZ is committed. Important issues, especially concerning social protection, remain to be solved Estimated Total Health Spending Table 2 shows the total estimated spending on public health services in Zanzibar from FY04/05 to FY07/08. The figures are based on the data collection for the PER08 from Programmes, Departments and Units (PDUs) on their detailed income and expenditure (reports by category, for data issues see 0). It includes details on funding by source, total spending and spending per capita, both in Tanzanian Shillings (TSh) and US Dollars (USD). 1 All DPs contributing to the General Budget Support for Tanzania, who are mainly the European bi-laterals, finance the Zanzibar health system indirectly, but with an important contribution, see Section

18 2 Zanzibar Health Sector Public Expenditure Review Table 2: Total estimated health spending (nominal), FY06/07-FY07/08 Source FY04/05 FY05/06 FY06/07 FY07/08 Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Expd GoZ (TSh m) 6,011 4,814 6,108 5,775 6,621 6,727 9,184 7,967 External Funding (TSh m) n/a 10,860 n/a 10,860 7,101 6,997 12,318 11,868 Cost-Sharing (TSh m) n/a 43 n/a 107 n/a Total TSh (m) n/a 15,717 n/a 16,742 13,722 13,886 21,668 19,999 Population 1,082,062 1,117,955 1,155,065 1,193,383 Total TSh per capita n/a 14,525 n/a 14,976 11,880 12,022 18,157 16,758 GDP (TSh m) 369, , ,250 n/a Health as % of GDP n/a 4.3% n/a 3.7% 2.5% 2.5% n/a n/a FX Rate USD1 = TSh 1,110 1,192 1,318 1,221 Total USD (m) n/a 14.2 n/a Total USD per capita n/a 13.1 n/a Table 2: Total estimated health spending (nominal), FY06/07-FY07/08 shows that average available external funding in FY06/07 and 07/08 declined compared to FY04/05 and 05/06. On the face of it, this is a worrying development, but there are two issues to be considered: The particularly low figures for external funding in FY06/07 contain two special effects: two large instalments of Global Fund (GF) grants to the Zanzibar Malaria Control Programme (ZMCP) were disbursed in FY07/08 instead of one each in FY06/07 and FY07/08. Also, data on an ADB programme in FY06/07 is missing. 2 These facts introduce a substantial downward bias in the figures for FY06/07, and an upward bias in FY07/08. It is possible that the low records are due to data availability. Data for FY06/07 was collected more than a year after the expenditures, which may have caused some problems. The data collected for external funding is also by no means accurate: sometimes different amounts for the same measures are reported for the same time period, by the same programmes, at the same time. This must be kept in mind when interpreting data, especially comparing different years. Figure 1 shows the development of total nominal expenditure in the health sector for the given time. Figure 1: Total expenditure in nominal terms by source, FY04/05-FY07/08 (TSh bn) CS External RGoZ FY04/05 FY05/06 FY06/07 FY07/08 2 The First Health Rehabilitation Programme (FHRP) ended and was replaced by the Support to Maternal Mortality Reduction Project (SMMRP). SMMRP was not able to provide information on FHRP.

19 Overview of the Health Spending in Zanzibar 3 Figure 2 presents the same analysis in real terms (at FY04/05 prices). This is of interest, as the real value determines what can be bought with the amount of money expended. After the drop in FY06/07 (which, as discussed, is partly due to special effects), the real funding has recovered. Figure 2: Total Expenditure in real terms by source, FY04/05-FY07/08 (TSh bn) CS External 6 RGoZ FY04/05 FY05/06 FY06/07 FY07/08 Figure 3 shows a similar analysis, but this time the share of the different sources is depicted. It is clearly visible that for external funding usually makes up the largest share of the spending in the sector, in FY07/08 it accounted for 59%; RGoZ funds (incl. GBS) provided another 40%, while the contribution of cost sharing has never been more than 1%. The RGoZ seems to have taken on a larger share of the financing, but the caveats on comprehensiveness of external funding still apply. Figure 3: Shares of total expenditure by source, FY04/05-FY07/08 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0% 1% 1% 1% 50% 69% 64% 59% 48% 31% 36% 40% FY04/05 FY05/06 FY06/07 FY07/08 CS External RGoZ The total funding envelope of the Zanzibar public health sector in FY07/08 was USD13.7 per capita. This is less than half of the USD34 estimated by the WHO Commission for Macroeconomics and Health (2001, p.15) as necessary to provide essential interventions. At the same time, the available funding per capita is well above the estimated financing requirements of USD3.63 for the Zanzibar Essential Health Care Package (EHCP) for the primary sector (Bijlmakers et al. 2007, p.86). The current problem is that even the EHCP needs to be financed to a large extent by RGoZ, as DPs usually provide funding tied to specific programmes and diseases. While the effective work of these programmes must be kept up, the RGoZ needs to make additional funds available from its own resources and attract budget support from DPs to provide the primary level EHCP and the coming hospital level EHCP.

20 4 Zanzibar Health Sector Public Expenditure Review 3. RGoZ Health Budget and Expenditure 3.1. Health Spending of the MoHSW Total nominal MoHSW Spending This section presents an overview of the total spending of the MoHSW. In the following sections it will be complemented by more detailed analysis. All numbers in the text are in Tanzanian Shilling if not indicated otherwise. Table 3: Nominal MOHSW spending, FY2003/04 FY2008/09 (TSh 1m) FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 FY2008/09 Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Recurrent 6,342 4,744 6,011 4,814 6,108 5,775 6,351 6,489 8,384 7,562 9,553 Development ,550 Total 6,553 4,774 6,233 4,819 6,341 6,066 6,621 6,727 9,184 7,967 11,103 %growth, year on year 10% 12% -5% 1% 2% 26% 4% 11% 39% 18% 21% The total health spending by the MoHSW has increased markedly from FY 2006/07 to 2007/08 in nominal terms. As can be seen in Table 3, the budget has increased to 9.2bn, which is an impressive increase of 39%. The actual expenditure has increased to 8bn, which translates into 18%. The different relative increase is mainly a consequence of the unusually high budget performance in FY06/07 that was not repeated in FY07/08. Reasons for this are discussed in the budget performance section (3.1.3). Figure 4: Nominal MOHSW approved budget and expenditure, FY2003/04 FY2007/08 10,000 9,184 9,000 8,000 7,967 7,000 6,553 6,233 6,341 6,066 6,621 6,727 6,000 5,000 4,774 4,819 Budget 4,000 Expenditure 3,000 2,000 1, / / / / /08 Figure 4 visualizes the information contained in Table 3. With the exception of the slight fall from FY03/04 to FY04/05, an accelerating upward trend is clearly visible. This trend is less marked for the actual expenditure than the budgets, though, and the budget-expenditure gap is widening again.

21 TSh m RGoZ Health Budget and Expenditure Total Real MoHSW Spending The central question for an assessment of health spending is: What service can be delivered to an individual patient? This question cannot be answered by just looking at the development of nominal budgets. It is important to take two things into consideration: Inflation: Prices usually increase from year to year, and TSh1,000 this year buy less than last year. To compare how much can be bought with a certain amount of money, it is important to account for the lost value of the money, i.e. to adjust for inflation. Population growth: Zanzibar s population is still growing strongly, at an annual rate of about 3.3%, which means that each year, the spending needs to be divided among more people. If it does not change, the money available per capita decreases. Table 4 shows that while the nominal figures have shown a steady increase from FY04/05 onwards (from 6.2bn to 9.2bn), the real value of the budget has declined from FY03/04 until FY06/07 (6.6bn to 4.9bn). The budget FY07/08 has recouped some of the losses, but still not all (5.8bn). Taking into account population growth, the per capita budget for FY07/08 is even below the FY04/05 value (5,1bn against 5.3bn). 3 Table 4: Summary measures of MOHSW budget, FY03/04 - FY08/09 (TSh m) FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 bdgt bdgt bdgt bdgt bdgt Nominal TSh (m) 6,553 6,233 6,341 6,621 9,184 USD (m) per capita TSh 6,257 5,760 5,672 5,732 7,696 per capita USD Real TSh (m) 6,553 5,723 5,267 4,900 5,811 real per capita TSh 6,257 5,289 4,711 4,242 4,870 CPI deflator population 1,047,373 1,082,062 1,117,955 1,155,065 1,193,383 USD exchange rate 1,066 1,110 1,192 1,318 1,221 The graphic representation of the information of Table 4 in Table 5 reinforces the need to adjust the nominal figures to conduct a meaningful analysis. Figure 5: MOHSW budget in nominal and real terms, FY2003/04 FY2007/08 10,000 9,000 8,000 7,000 6,000 5,000 4,000 3, / / / / /08 Nominal (current, TSH m) 6,553 6,233 6,341 6,621 9,184 Real (FY03/04=100, TSh m) 6,553 5,723 5,267 4,900 5,811 Real per capita (TSh) 6,257 5,289 4,711 4,242 4,870 3 The PER07 reported a real budget of TSh5,003m for 2006/07. The difference to the TSh4,900m reported here is due to updated inflation data. The PER07 used an official inflation estimate that has proven too low, due to the timing of publication. Only because of delays this has been avoided this year, see below.

22 TSh m 6 Zanzibar Health Sector Public Expenditure Review Table 5 and Figure 6 present the same analysis for the expenditure of the MoHSW. Table 5: Summary measures of MOHSW expenditure, FY03/04 - FY08/09 (TSh m) FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 expd expd expd expd expd Nominal TSh (m) 4,774 4,819 6,066 6,727 7,967 US$ (m) per capita TSh 4,558 4,454 5,426 5,824 6,676 per capita US$ Real TSh (m) 4,774 4,425 5,038 4,978 5,041 real per capita TSh 4,558 4,090 4,507 4,310 4,224 CPI deflator population 1,047,373 1,082,062 1,117,955 1,155,065 1,193,383 US$ exchange rate 1,066 1,110 1,192 1,318 1,221 The development of the expenditure has overall been more positive than that of the budget, both nominal and real. The trend is broadly upward with some small dents, but this is also due to the low starting position, i.e. the poor budget performance in FY03/04. Accounting for population growth, however, the trend has been fairly flat. The spending per capita has constantly been between TSh4,000 and TSh4,500. Figure 6: MOHSW expenditure in nominal and real terms, FY2003/04 FY2007/08 9,000 8,000 7,000 6,000 5,000 4,000 3, / / / / /08 Nominal (current, TSH m) 4,774 4,819 6,066 6,727 7,967 Real (FY03/04=100, TSh m) 4,774 4,425 5,038 4,978 5,041 Real per capita (TSh) 4,558 4,090 4,507 4,310 4,224 The inflation figures used have been kindly provided by the OCGS. Up to 2008, they are final figures. At earlier stages, however, a projection for 2008 had to be used, which turned out to be far off target. If it had not been for some serious delays, the PER08 would have reported serious overestimates of the real budget FY07/08. There is a trade-off between a report shortly after the end of a financial year, and accurate inflation data, which is reported for calendar years. At the same time, ideally, the PER should be ready before budget negotiations with MoFEA in order to provide arguments for attracting an adequate share of RGoZ resources to the Health Sector. The RGoZ needs to increase its spending on health through the MoHSW, in order to keep up with a) inflation, and b) population growth. The RGoZ reports positive economic development on a per capita basis and in USD (i.e. largely inflation adjusted) for each year since 2003 (MoFEA Zanzibar 2008b, p.1), but too little of this has been translated into a higher spending on health services for the individual patient.

23 RGoZ Health Budget and Expenditure Budget Performance Budget performance is a term used for indicating how closely budget and actual expenditure are related, it is the ratio of actual expenditure over the budget, expressed in percent. Generally, a high budget performance is positive; a low budget performance indicates problems in spending funds on the implementer s side, or problems in disbursements on the funder s side. At the same time, a high budget performance can also be a sign of unrealistically low budgets. Figure 7 shows that the budget performance of the MoHSW has improved steadily since FY03/04 until it reached a value of 102% in FY06/07, to drop back down to about 85% in FY07/08. The main reason for the exceptionally high budget performance in the financial years 05/06 and 06/07, seem to be conservative planning and an underestimation of personnel costs. In 07/08, the budget was increased more boldly than in previous years, and the budget performance immediately dropped, although not to the low levels seen in 03/04 and 04/05. Figure 7: MOHSW expenditure as % of approved budget, FY2003/04 FY2007/08 105% 100% 95% 90% 85% 80% 75% 70% 65% 60% 2003/ / / / /08 Budget Performance 100% MoHSW Spending in Comparison to RGoZ Spending The analysis of the preceding section dealt with the absolute availability of funds, this section looks at how much of the total RGoZ resources go to the MoHSW. The share a sector receives of the government spending is an indicator of the importance of the sector to the government. The analysis of both budgets and actual spending are important. Actual spending shows how much funds a sector receives for providing its services to the public, and budgets are expressions of political will and priority setting. Budgets can be set fairly freely, while actual expenditure is partly determined by the budget. 4 Other influences on actual expenditure are budgetary structures, e.g. spending on complementary inputs (OCs) can be slashed more easily than wage-related (PE) items. Figure 8 and Figure 9 show the share the MoHSW receives of the RGoZ spending. They show the MoHSW s share both if only recurrent spending is taken into account and its share if local development spending is included (total local). 4 Path dependency is often an important limit to this freedom, i.e. the relative importance of sectors is often kept fairly constant over time in order to avoid political fights over the distribution.

24 % of GOZ expenditure % of RGoZ budget 8 Zanzibar Health Sector Public Expenditure Review Figure 8: MOHSW budget as share of RGOZ total, FY2003/04 FY2007/08 (%) 9.5% 9.0% 8.5% 8.0% 7.5% 7.0% 6.5% 6.0% 5.5% 5.0% 2003/ / / / /08 % discretionary recurrent (excl CFS) 9.1% 8.6% 8.5% 8.1% 8.0% % GOZ total local (excl CFS) 6.5% 6.4% 6.4% 5.6% 5.9% As can be seen in Figure 8, the share of the MoHSW is higher if recurrent expenditure is compared with total recurrent expenditure of the RGoZ. The share drops if for both the MoHSW and RGoZ the development expenditure is included in the analysis. This indicates that the share of the MoHSW of development expenditures must be lower. Figure 9 provides the same analysis for the actual spending. The observations are similar to Figure 8, but there is more volatility in the figures. Figure 9: MOHSW expenditure as a share of RGOZ total, FY2003/04 FY2007/08 (%) 11.0% 10.0% 9.0% 8.0% 7.0% 6.0% 5.0% 2003/ / / / /08 % discretionary recurrent (excl CFS) 7.7% 7.9% 9.8% 8.3% 8.7% % GOZ total local (excl CFS) 6.7% 5.9% 6.6% 6.0% 6.8% It is important to put the figures into the context of the total budget of the RGoZ. Table 6 looks at the five votes (MDAs 5 ) that regularly receive the biggest shares of RGoZ budgets and expenditure, if CFS are excluded. 6 Overall, both the ranking and the shares change very little from year to year. 5 In FY2007/08, there were two administrative changes, the Min. of Education, Culture and Sports changed to Min. of Education and Vocational Training, and the Min. of Agriculture, Natural Resources, Energy and Cooperatives to Min. of Agriculture, Livestock and Environment. 6 In 2005/06, the Anti-Smuggling Unit had the fourth highest expenditure; in 2006/07, the third highest expenditure was recorded for the MoCT. Both votes are regularly among the next highest after the Top 5.

25 RGoZ Health Budget and Expenditure 9 Table 6: Top Receiving MDAs, budgets and expenditures FY05/06 FY07/08 (recurrent vs recurrent excl. CFS) MDA FY05/06 FY06/07 FY07/08 Bdgt Expd Bdgt Expd Bdgt Expd Rank % Rank % Rank % Rank % Rank % Rank % MoECS / MoEVT % % % % % % MoFEA % % 2 9.8% % % 3 8.9% MoHSW 3 8.5% 3 9.4% 3 8.1% 4 8.8% 3 8.0% 2 9.3% HoR 4 6.8% 6 5.6% 4 6.3% 6 5.9% 5 5.4% 4 6.5% MoANREC / MoALE 5 6.1% 5 6.0% 5 6.3% 5 5.9% 4 5.7% 5 6.3% The ranking shows a fairly positive picture for the RGoZ s commitment to the health sector. The MoHSW has been number three out of the 34 votes in the annual budget, and for expenditure, it ranked third in FY05/06, fourth in FY06/07 and second in FY07/08. The share of expenditure has always been higher than that of the budget. The positive impression made by this ranking does not hold, however, if the spending is compared to other Sub-saharan African (SSA) countries. Figure 10 shows (red, long-dashed line) that the Zanzibar health sector receives a lower share than in most other countries in the region in 2003 and 2006, (the most recent figures available, Zanzibar figures are from FY03/04 and FY06/07). Kenya is the exception. Figure 10: Health as % of Government spending in SSA (incl. debt service) 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% 6% 6% 10% 9% 6% 5% 11% n/a 14% n/a 11% 11% 12% 12% 18% n/a 9% n/a Table 7 adds information on the health share when debt service is excluded, which raises the share for all countries, including Zanzibar. Table 7: Health as % of Government spending in SSA Total local budgets incl. debt service excl. debt service share year Zanzibar (FY03/04, 05/06) 6.5% 5.6% 7.6% 2006 Tanzania (FY03/04, 05/06) 8.7% 10.2% 11.6% 2006 Kenya 6.0% 5.1% Malawi 10.8% n.a. Namibia 13.8% n.a. 15.0% 2003 South Africa 11.1% 11.0% 12.8% 2006 Uganda 11.8% 11.7% 12.5% 2006 Zambia 17.7% n.a. Zimbabwe 9.2% n.a. Source: Govender, McIntyre & Loewenson 2008, pp.13, 14; authors calculations

26 10 Zanzibar Health Sector Public Expenditure Review For a fair comparison, it should be pointed out that all other countries included here are facing a much higher burden of HIV/Aids, which greatly increases the financing needs. At the same time, most of the HIV/Aids spending is financed by development partners, and not by the government. The MoHSW has set a yardstick for the sector s share in the RGoZ budget in its ZHSRSPII of 12% (MoHSW Zanzibar 2007, p.64), to be achieved by FY10/11. Through the Abuja Declaration, RGoZ is even committed to allocating 15% to health (green, short-dashed line in Figure 10. The Government of the United Republic of Tanzania (GoT) as the representative of Zanzibar in international relations, together with other African countries, committed to set a target of allocating at least 15% of *their+ annual budget to the improvement of the health sector (OAU 2001). No matter which definition of spending is used (see also Appendix 3 for further measures), RGoZ is a long way off the targets of 12% or even 15%. More worryingly, if the comparison between local recurrent spending of MoHSW and RGoZ is used (which delivers the highest MOHSW share), the health share is falling. In order to achieve the targets set out in MKUZA, the Health Policy and the HSRSPII, it is important that RGoZ increases its funding to the level it has promised General Budget Support to URT and the MoHSW budget The United Republic of Tanzania receives part of its assistance from development partners through General Budget Support (GBS), and part of it is passed on to Zanzibar. The main feature of GBS is that once it has been transferred from the donor to the recipient, it is part of the general revenue of the recipient government and can be spent accordingly, i.e. with no strings attached regarding the way the money can be spent. 7 This means that when calculating the share of external financing of a sector, the support from GBS to the sector could be included under external funding to reflect reality. The decision to commit these funds to the health sector, however, has been that of the GoZ. Zanzibar gets a share of 4.5% of the total GBS the URT receives, and the share of GBS the MoHSW receives from the RGoZ total is equal to its share from the total RGoZ budget. Table 8 gives an overview of the GBS figures, down to the level of the MoHSW. Table 8: United Republic of Tanzania s GBS funds to MoHSW Zanzibar FY FY03/04 FY04/05 FY05/06 FY06/07 FY07/08 GBS URT TSh m 405, , , , , % to GoZ TSh m 18,227 19,551 27,727 36,205 39,659 Share MoHSW/GoZ incl. CFS % 6.5% 6.4% 6.4% 5.6% 5.9% MoHSW TSh m 1,180 1,251 1,775 2,011 2,324 Share of MoHSW total % 18.0% 20.1% 28.0% 30.4% 25.3% The total amount of GBS funding that reaches the MoHSW has been steadily increasing, reaching TSh2.3bn in FY07/08. Equally, the share the GBS funds make up as a proportion of total MoHSW funding has been increasing until FY06/07 to reach 30%. In FY07/08, the share has fallen again to slightly more than a quarter Sub-Sectoral Spending Recurrent versus Development Spending The MoHSW budget is split in two parts, the Recurrent and the Development Budget. The recurrent budget is meant to include funds for the ongoing provision of existing services (recurrent), while the 7 When a switch from programme support or sector funding is agreed, recipient and donor governments may agree that the contribution to a certain sector, e.g. social services, should not fall under the new modalities.

27 RGoZ Health Budget and Expenditure 11 development budget is meant to contain the funds for extending the services (development), either in quality or quantity. Services here include the outputs of all units, not only medical services. In fact, however, the main criterion for inclusion of a programme in the development budget seems to be political: The visibility of a development programme in the budget signals its importance. The current development programmes often coincide with disease specific programmes, and finance all sorts of non-wage expenditures of those programmes, no matter if development or recurrent. Similarly, at times construction work is included in the MoHSW MTEF for recurrent spending, which is investment or development spending. This is a smaller problem, though, since most of the spending on development is undertaken by development partners. Despite this shortfall, an analysis of the split into recurrent and development spending is presented in Table 9 and Table 10. Table 9: MoHSW recurrent and development Budgets FY03/04-08/09 FY03/04 FY04/05 FY05/06 FY06/07 FY07/08 Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Expd Recurrent 96.8% 99.4% 96.4% 99.9% 96.3% 95.2% 95.9% 96.5% 91.3% 94.9% Development 3.2% 0.6% 3.6% 0.1% 3.7% 4.8% 4.1% 3.5% 8.7% 5.1% Table 9 clearly shows how the MoHSW budget is tilted towards recurrent expenditure, although in recent years the share of the development budget as part of the total has been increasing, both for the budgets, as for the actual expenditure. In Table 10 it can be seen that the budget performance of recurrent expenditure has usually been much higher than that for development expenditure. Since FY05/06, there has been a marked improvement in the development budget performance (FY05/06 saw the special effect of a delayed release of counterpart funding to an ADB project). Table 10: Budget performance: recurrent and development spending FY03/04 FY08/09 FY03/04 FY04/05 FY05/06 FY06/07 FY07/08 Recurrent 75% 80% 95% 102% 90% Development 14% 2% 125% 88% 51% While the current reliance on donor funding for development spending is not a sustainable solution for the long-term, and the government development spending needs to increase, it should be noted that there is only limited need in Zanzibar for additions to the health infrastructure, thus limiting the need for large development spending (MoHSW Zanzibar 2007, p.51, Bijlmakers et al. 2007, p.v) Personnel versus Other Charges The recurrent budget of the MoHSW is split further into Personal Emoluments (PE), containing salaries and wages, Zanzibar Social Security Fund (ZSSF) contributions and two types of allowances (bus fares and special allowances), and Other Charges (OC), which contain all complementary inputs, from medical equipment and drugs to refreshments for meetings. Table 11 shows the split of the recurrent budget into PEs and OCs in recent years. Table 11: MOHSW recurrent spending on PEs and OC, FY03/04 FY08/09 (TSh m) FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Expd PE 4,232 4,021 4,349 4,272 4,824 5,025 4,995 5,542 6,875 6,705 OC 2, , , , , Total 6,342 4,744 6,011 4,814 6,108 5,775 6,351 6,489 8,384 7,562 PE as % of Total 67% 85% 72% 89% 79% 87% 79% 85% 82% 89%

28 12 Zanzibar Health Sector Public Expenditure Review It is immediately obvious, that the recurrent budget (which is usually about 95% of the total budget) is heavily weighted in favour of PEs, which recently have accounted for about 80% of the recurrent budget, and more than 85% of the actual expenditure. This means that around 80% of the total MoHSW budget each year is spent directly on staff. Developing countries often have a high share of staff costs in their budget (70% and more, Islam 2007, p.7.18), but Zanzibar s share is particularly high and leaves hardly any room to purchase complementary inputs. Table 12: Budget performance: PE and OC, FY03/04 FY07/08 FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 PE 95% 98% 104% 111% 98% OC 40% 33% 58% 69% 57% As pointed out above, the share of PEs is usually higher in the actual expenditure than in the budget, i.e. the budget performance for PEs is higher, as Table 12 shows. This is mainly because it is less problematic to not procure something than to not pay staff. While the budget performance for OC is a cause of concern because it is so low, the more than 100% budget performance for PEs is also likely to be rather a consequence of low-quality planning than a large unexpected influx of qualified staff (which in an environment of skill shortage would be positive) Spending by Type of Activity As in the PER07, an attempt is made to identify what kinds of activities are being financed through the government budget. The OC budgets are distributed according to their subvotes and the Nominal Role (NR) has been analysed with regard to the type of activity. Table 13 shows that most costs are incurred in the curative services, followed by Planning and Preventive to nearly equal amounts, and then Social Welfare and Substance Abuse (SW&SA) and the Chief Chemist. Table 13: Crude expenditure breakdown by category, FY07/ /08 PE OC Total Zanzibar Unguja Pemba Unguja Pemba Unguja Pemba Planning ,056.8 Preventive ,052.0 Curative 1, , ,107.5 SW&SA Chief Chemist MoHSW 2, , , , , ,939.7 The picture is very different between Unguja and Pemba. For the PE, Unguja has naturally much higher spending on Planning, since all the central administration is concentrated here. But it is remarkable that in Unguja, the PE spending on Curative Services is higher (more than three times) than for Preventive Services, while it is smaller in Pemba, where all District Hospitals are included under Curative Services. This shows the huge importance and cost of MMH in the health sector in Zanzibar. It should be noticed, however, that hospitals also offer primary health care, especially MMH, which serves as the national referral hospital, the DH/PHCC for Urban and West Districts, and as the PHCU for Stone Town, and that the classification by subvote is not equal to curative vs preventive care. Figure 11 shows a classification that avoids the problem of preventive and curative care by using the categories primary care versus hospital care. 8 In section 3.3.1, a more detailed analysis of PE spending is undertaken. Due to the time-intensive nature of the analysis of the Nominal Role, we did not analyse the NR for 2006/07. With regard to this, efforts should be made to get access to the electronic version of the NR for next year s PER.

29 RGoZ Health Budget and Expenditure 13 Figure 11: Estimated allocation of PEs between activity types, FY07/08 Primary Health Care 25% Various 5% Hospital 47% Admin & Managin g 16% Technical Support 7% This figure of PE expenditure brings out even more clearly the focus of the Zanzibar health system on hospital services. Nearly half of all salaries are paid in hospitals, with the rest nearly equally split between primary health care and administration and management and technical support. It should be noted that the numbers changed slightly against the PER Detailed Analysis of Recurrent Spending In the following, some more detailed analysis is undertaken of the recurrent spending in selected areas. First staff qualification and training are looked at, and then spending on medical supplies and services, operation of health facilities, and lastly medical spending abroad Personnel 1: Staff Qualification Human Resources are absolutely crucial for the functioning of the health sector. Sufficient staff, both in quality and quantity, is a necessary although not sufficient condition for providing health services. As pointed out before, the MoHSW directs most of its spending (around 80%) towards salaries and other staff-related costs. It is important therefore to have a more detailed look at how this substantial amount of money is spent. Figure 12: Staff by service area and skill level, FY07/08 Figure 13: Salaries by service area and skill level, FY07/08 Number of Workers 5% Salaries Health Staff Un-Skilled Health Staff Semi- Skilled Health Staff Skilled Other Staff Un-Skilled 8% 12% Total: 3,649 24% 1% Health Staff Un-Skilled Health Staff Semi- Skilled Health Staff Skilled Other Staff Un-Skilled 5% 9% 5% Total: TSh6,519m 13% 1% Other Staff Semi-Skilled Other Staff Skilled 50% Other Staff Semi-Skilled Other Staff Skilled 67% Figure 12 shows that the staff is split between health and other staff. Overall, this is a positive picture. It needs to be noted that some personnel showed as health staff may work in non-health positions, while the reverse is less likely. Within the two groups, the skill mix 10 is very different. Among health workers, one third of the current health staff is unskilled, i.e. orderlies. The plans for FY08/09 indicate an improvement of the 9 See Appendix 2 on sources of information. 10 For both groups, less than 9 months training is considered unskilled, between 9 months and 2 years is semiskilled and 2 years and more is considered skilled. These categories were developed with the Personnel Unit and cleared with the Training Unit.

30 14 Zanzibar Health Sector Public Expenditure Review situation, with a shift from unskilled to skilled staff. Among other staff, only 19% of the staff are skilled, which indicates a serious problem. Even the larger share of semi-skilled staff is not much comfort, as secretaries and drivers are included here, due to the duration of their training. Figure 123 shows the split of the salaries according to groups and categories. Health workers receive about 80% of the total amount (i.e. TSh5.3bn), which is due to the increased share of the skilled health workers. Un- or semi-skilled other staff only receives 14% of the wage bill. The overall picture shows that action is needed to increase the share of skilled personnel, as also indicated by the PS MoFEA on the Partner Coordination Meeting in April 2008, when he spoke about rightsizing the public sector, not only downsizing it. The Nominal Role for FY08/09 shows that steps in the right direction are being taken for health staff, among which the number of skilled staff has increased, while the number of unskilled staff has decreased. Unfortunately, the same does not apply for Other Staff Personnel 2: Training Increasing, or even just maintaining, the amount of qualified staff is a concern for the MoHSW. The Health Policy notes a severe shortage of medical experts and doctors and complains about a brain drain (MoHSW Zanzibar 2002, p.50). The ZHSRSPII points out that inadequacy of HRH severely constrains implementation of health activities on all levels (MoHSW Zanzibar 2007, p.11). Table 14 shows what is being done to increase the capacity of staff in the MoHSW. It includes spending on the College of Health Sciences (CHS), the institution where most of the health staff receive their pre-service training (with the notable exception of medical doctors), and the spending on in-service training. From FY08/09 onwards, medical training of doctors will be included in the development budget under the Zanzibar Medical School. Table 14: RGOZ spending on training activities, FY2004/05 FY08/09 (Tsh m) FY04/05 FY05/06 FY06/07 FY07/08 Bdgt. Expd. Bdgt. Expd. Bdgt. Expd. Bdgt. Expd. CHS Other training Total As % of OC 10% 13% 12% 9% 11% 17% 10% 13% As % of Recurrent 3% 1% 3% 1% 2% 2% 2% 2% Expenditure as % of Budget 43% 45% 106% 75% The overall picture regarding capacity building is bleak. Only around 10% of OC spending is budgeted each year for training, and less is spent than budgeted (training s share of total expenditure can still be higher than of total budget because both numerator and denominator change). One worrying development is that the share of the budget has been decreasing constantly over the last four FYs. While the present figures are worrying, they do not represent the total amount of capacity building that is going on, as DPs often fund trainings (see also section 4.2.4). In fact, some DHMTs and hospitals pointed out that at times health facilities are understaffed because of the many trainings (and workshops) Programmes are organising. 11 Trainings need to be better coordinated, either by the Training Unit or the Department of Preventive Services, which coordinates the Programmes in other matters. At the same time, all DHMTs and hospitals appreciated trainings as a means to improve capacity and also as a (usually) non-financial incentive for their staff. Similarly, the RGoZ 11 At Micheweni PHCC, the PER team was told, only part-jokingly, that they were lucky to come during that week, as in the previous week hardly anyone had been at the facility due to training and other activities.

31 RGoZ Health Budget and Expenditure 15 funding for CHS is dwarfed by the spending of DPs. According to the 2008/09 MTEF, DPs plan to spend nearly 2bn on CHS (of which 1.3bn are for construction of buildings, however). While this reliance on external finances is clearly not sustainable in the long term, this funding helps to address the most pressing capacity issues at present Other Charges: Medical Supplies and Services Medical Supplies and Services (MSS) include running costs for hospitals, PHCCs and PHCUs, and the procurement of drugs, i.e. the complementary inputs to qualified staff for provision of health services. Table 15 shows that the absolute amount allocated for MSS is fairly small. This is not surprising, considering the small share of the total budget that goes to OCs in general. Table 15: RGOZ allocations to Medical supplies and services, FY03/04 FY08/09 FY03/04 FY04/05 FY05/06 FY06/07 FY07/08 Bdgt. Expd. Bdgt. Expd. Bdgt. Expd. Bdgt. Expd. Bdgt. Expd. Total MSS (TSh m) As % OC 31% 16% 23% 14% 17% 20% 21% 23% 17% 13% As % of Recurrent 10% 3% 6% 2% 4% 3% 5% 3% 3% 1% Expd as % of Budget 21% 20% 68% 76% 42% The numbers are somewhat worrying. Even as a percentage of OC budgets, MMS have accounted for only around 20% (with the exception of FY03/04), and there is a downward trend. The expenditures have always been smaller than the budgets in absolute figures, although relative to the total OC spending, in FY05/06 and 06/07 the expenditure share was larger than the budget share. This means that the budget performance of MMS was better than that of other OCs. The figures for FY07/08 are clearly worrying. Budgets and expenditures are smaller in absolute terms than in the previous year, and they are the smallest shares on record, both when compared to OC and the total recurrent budget. A health system that spends only 1% of its total resource envelope is clearly not sustainable on its own. The gap is currently filled by development partners. For the mid- and long-run this is clearly a problem, as donor funding has traditionally been volatile and depends on national and international developments outside of the health sector. In the shortrun, the assessment depends on how the rest of the RGoZ funds are spent. There clearly are problems in attracting and retaining qualified staff, so that sorting out these issues by focusing on PE spending while there is partner-funding for MSS available, can be a useful strategy if this would be agreed with donors. At the same time, the previous sections have shown that the MoHSW also has not made sufficient progress in addressing staff-related issues. There is still a large overhang of un- or semi-qualified staff, especially among other (non-medical) staff, and funding for training expenditures has also been limited. Some work has been done to address the problem of deployment in rural areas by building more staff housing (with ADB funds), but so far no other mechanisms are in place that could address this problem, either in ways that would affect the budget or not. So it does not seem as if the MoHSW s low spending on MSS would be a strategic decision Medical Expenses Abroad Medical Expenses Abroad (MEA) cover the costs of specialist treatment that has been received outside of Zanzibar, usually because adequate treatment is not available domestically. Table 16 shows the development of budgets and expenditure for this category. MEA often involve high profile cases, dealing with life threatening conditions. At the same time, services abroad are often subject to the same access constraints for poor and rural parts of the

32 16 Zanzibar Health Sector Public Expenditure Review population like hospital services (compare Schieber 2006, p.8). MEA spending is therefore controversial from an equity-of-access point of view. This is even more so the case, if the budget performance of MEA is regularly higher than that for domestic MSS, as in Zanzibar. Table 16: RGOZ allocations to Medical expenses abroad, FY04/05 FY08/09 (TSh m) FY04/05 FY05/06 FY06/07 FY07/08 Bdgt. Expd. Bdgt. Expd. Bdgt. Expd. Bdgt. Expd. Unguja Pemba Total As % of OC 6% 13% 5% 14% 5% 7% 5% 5% As % of Recurrent 2% 1% 1% 2% 1% 1% 1% 1% Budget Performance 75% 158% 100% 53% Health Service Delivery: Operation of Health Facilities Operation of health facilities is a subcategory of MSS. The spending on different types of health facilities can be separated out using the MoHSW budget and the Nominal Roll (NR) (see Appendix 5 for the methodology). Table 17: Estimated allocations for running of health facilities, FY07/08 (TSh m) FY2007/08 Personal Emoluments Other Charges Total Bdgt Actual Bdgt Actual Bdgt Actual PHCU 1, , , ,109.0 PHCC DH MMH 2, , , ,277.4 Total 4, , , ,792.8 Table again shows that the MoHSW has a much greater role in financing PEs compared to OCs, and that the lion s share of budgets and expenditures goes towards hospital care. Figure 14 illustrates this very clearly: Nearly half of MoHSW expenditures go to the MMH and another 20% to the three district hospitals. This seems to be an imbalance. PHCCs, or Cottage Hospitals, have always been treated as primary care in the PER. Adding the shares of PHCUs (23%) and PHCCs (9%) gives primary care a 32% share of budgets. In FY05/06 the share was still at 34%. Figure 14: Estimated RGOZ spending on running costs of health facilities (budgets), by level, FY07/08 MMH 48% PHCU 23% PHCC 9% DH 20% 12 The methodology underlying this table is explained in Appendix 5. It was noted that the total of the NR is TSh1.2bn lower than the Salaries and Wages budgeted by MoFEA. It needs to be cleared why this difference exists.

33 RGoZ Health Budget and Expenditure 17 The ZHSRSP II calls primary health care the cornerstone of the Zanzibar health care system (MoHSW Zanzibar 2007, p.21). It goes on to argue that it is the most cost-effective, extensive and equitable form of health care delivery to the Zanzibari people. With only 32% of the available funding allocated to primary care, it is not sure how strong this cornerstone is. While access to primary care is theoretically very good, with 95% of the population living less than 5km away from their nearest PHCU, the figures here already indicate service quality issues. When visited by the PER team, all visited DHMTs and hospitals complained about the problematic release of RGoZ funds and indicated that the basis of their activity planning was the HSF (see Section 4.2.5). There is an urgent need to increase the funding and release budgeted funds for district and primary health care, both on the grounds of efficiency as well as equity Spending by Location The analysis of the spending by location is important for assessing the equity of the spending, i.e. if all Zanzibar citizens benefit equally from the RGoZ health spending. The population in FY07/08 is estimated to be about 1.2m, with 62% in Unguja and 38% in Pemba. These values can be taken as a benchmark for assessing the spending. Table 18 shows the distribution of spending. Table 18: Spending in Unguja and Pemba FY2003/04 - FY2007/08 FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Expd Bdgt Expd Unguja 4, , , , , , , , , ,304.2 Pemba 1, , , , , , , , , ,257.8 The allocation of funds is facilitated by the fact that for most departments, Unguja and Pemba have their own subvotes. In line with the PER06 and 07, the Public Health Department and the Chief Government Chemist are split between the islands for the analysis, and MMH 95-5 for Unguja. The allocation is not ideal, especially for Public Health, but to maintain consistency with previous PERs, the share is maintained. 13 Numbers for PE and OC spending by island have changed slightly against the PER07, as PE expenditure in subvotes other than Planning (Unguja and Pemba) and MMH had wrongly been allocated to OCs before; this has been corrected for this PER. Figure 15 and Figure 155 show what each island received as a share of budget and expenditure in FY07/08. Unguja received slightly more than two thirds of the budgets and 70% of the MoHSW expenditure. The 2006 Census (based on population figures from 2002) estimates that in 07/08, about 62% of the population resides in Unguja and 38% in Pemba. Poverty and Under-5 mortality rates also point towards a greater need for health care in Pemba. In this light, Pemba s share seems too low. But it also has to be kept in mind that the overall administration and decision making is taking place in Unguja. Therefore, a share somewhat higher than by population is warranted. 13 For MMH, the split is based on a bed census carried out for the EHCP. For the CGC, a split by population or by the UNG/PBA average of all other departments would seem more useful. The same applies to Public Health, with a greater share for Unguja, as Public Health in Pemba is partly under the Dept. of Preventive Services.

34 18 Zanzibar Health Sector Public Expenditure Review Figure 15: Budgets by Zone, FY2007/08 Figure 16: Expenditure by Zone, FY07/08 Pemba 32% Pemba 30% Unguja 68% Unguja 70% Figure 17 shows the budget performance of the OC allocations to Unguja and Pemba islands. While Unguja has (and has always had) a better budget performance than Pemba, the gap has shrunk in the last years. This is very positive, as it makes the access to health care on both islands more equal. Unfortunately, the main reason for the closing of the gap is a decrease in the budget performance in Unguja, not an increase in the performance in Pemba. The need to increase the OC budget performance in general, and in Pemba in particular, remains. Figure 17: Budget performance for Other Charges, Unguja and Pemba, 2003/ / % 80.0% 76.8% 78.3% 70.0% 59.5% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 45.8% 40.3% 23.9% 16.7% 28.1% 53.9% 51.4% Unguja Pemba 0.0% 2003/ / / / /08 The PE spending can be broken down to the district level, at least for the wage part, which is included in the Nominal Role (NR) (and the ZSSF contributions, which are proportional, i.e. 10% for all civil servants). Figure 18shows the distribution of salary spending according to the districts.

35 RGoZ Health Budget and Expenditure 19 Figure 18: Estimated allocation of salary spending by geographical area, FY2007/08 National Pemba Unguja Micheweni Chake Chake Wete Mkoani North A North B Urban West Central South Unallocated 1% 3% 3% 2% 3% 5% 6% 6% 8% 8% 12% 12% 14% 17% 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% The figure shows the large share of salaries that are paid for institutions with a national remit, 17%. It also shows the importance of hospitals for salary costs, West and Unguja who share most of the MMH salaries (see below), and the three Pemba districts with DHs have the highest salary costs. The sparsely populated rural districts in Unguja, together with Micheweni, have the lowest salary costs. The classification of salary cost-centres is taken from the PER07. National are all central institutions that exist only once on the islands (e.g. CGC) or the headquarters of disease programmes (e.g. ZMCP). Unallocated are those salaries that are transferred to bank accounts. 14 Unguja includes the ZHMT, as well as Kidongo Chekundu Hospital, Mwembeladu Maternity Hospital and 20% of MMH salary costs. Pemba includes the ZHMT, all head offices of departments, and 5% of MMH salary costs. Districts are allocated PHCUs, PHCCs, and DHs located in their district. West and Urban are allocated an additional 37.5% of MMH salary costs. North A and B have swapped their allocations compared to the PER07, after PHCC Kivunge has been correctly allocated to North A (from North B). 14 With a better coverage of the islands by banks, the proportion of bank transfers is likely to go up, making the current analysis less useful. The personnel unit should register bank transfer recipients by their work place.

36 20 Zanzibar Health Sector Public Expenditure Review 4. External Financing of the Health Sector 4.1. Total Volume The total volume of external financing is estimated at around TSh13.1bn for the budgets and around TSh11.9bn for the expenditures. The estimate is based on the detailed reporting for the PER08 and does not include the estimate for GBS funding that reaches the MoHSW. There are other sources of information that provide other estimates, but these are not as comprehensive. In the following, however, all sources of information are analysed. The accuracy of the external financing is more limited than that of domestic funds. There are three types of external financing: On-budget funding through the development budget, funding that is channelled through the programmes and that is planned for and executed by MoHSW departments, units or programmes, i.e. quasi on-budget, and lastly funds that are administered and spent by development partners themselves. While the development budget should be comprehensive for the included programmes, there is some uncertainty about the comprehensiveness on reporting on funds administered through programmes, and more about the funds spent by DPs themselves; substantial amounts are possibly not captured. There are two issues, one is that reporting of contributions from outside sources is not sufficiently organised within the MOHSW (only the programmes report regularly, because of MoFEA requirements), and the other is that partners at times disburse funding to recipients outside of the MoHSW. Therefore, the MoHSW needs to (a) establish a system in order to ensure that all units, departments, programmes and health facilities report all funds they receive, and (b) that the MoHSW comes to an agreement with the Development Partners to report all those disbursements to non-mohsw institutions, or alternatively make notification of the MoHSW by the recipient a requirement for their funding. The MoHSW does not routinely record and quantify in-kind contributions to the MoHSW, and the PER team has not made any attempts to do so either. A register for all in-kind contributions should be developed, and a methodology for converting these into monetary terms. Apart from the problems of completeness of records, there are also sometimes problems in reporting (to MoHSW and MoFEA) and in record-keeping. The estimate for the total external funding should therefore be seen as a lower boundary for the funds available On-Budget Funding through Development Budget The on-budget funding for the MoHSW is shown in Table 19 (as included in the MoFEA Budget Estimate Books - BEBs). The development budget in FY2007/08 stood at TSh4.8bn, but it has been very volatile over the past years. This has been due to changes in the contributions of individual partners but also to changes of partners over the years. Programmes are sometimes listed by incorrect or outdated names in the BEBs, which can be the source of some confusion. At times it is also not clear how the programmes in the Development Budget match the existing institutions, and how partners contributions are allocated in the Appropriations Accounts. The Danida-sponsored Health Sector Reform Programme (HSRP) is an example: no expenditure is recorded, in the Appropriations Account, but the Health Sector Reform Secretariat and other reform-related units received funds from the Health Sector Programme Support (i.e. Danida) and also expended them in FY2007/08. Why this does not show in the Appropriations Account is not clear. This issue needs to be taken up with the MoHSW Chief Accountant.

37 External Financing of the Health Sector 21 Table 19: DP Budgets for MoHSW Development Budget, FY03/04 FY07/08 (Tsh 1m) Name Donors Type FY03/04 FY04/05 FY05/06 FY06/07 FY07/08 Immunisation and MCH support (1) GAVI,Unicef,WHO Grant Reproductive and Child Health UNFPA Grant Programme (2) Support to Maternal Mortality ADB/ADF Loan 1, , , ,820.0 Reduction Programme (3) AIDS Control Programme (4) UNDP,CDC,WB,GF Grant , ,286.4 Uni of Columbia Health Sector Reform Programme Danida Grant Total 2,342 7,573 3,515 3,632 4,796 Notes (1) Unicef and WHO joined GAVI support in 2005/06 (2) In the Budget Estimate Books RCHP is incorrectly still listed as Family Planning (3) SMMRP in FY07/08 took over from the Programme for Primary Health Care (name in BEBs) (4)UNDP only in 2004/05, in 2007/08 WB, CDC, Univ. of Columbia, in 2008/09 GF only (BEBs) Table 20 shows the expenditure as recorded by the MoHSW. Accounts on externally financed programme expenditure were started only in FY06/07. Budget performance is very low and unstable, according to the itemized appropriations account. For the two years available, there is not one Programme that has recorded a budget performance of more than 30% both years. Table 20: DP Actual Expenditure for Development Programmes, FY06/07 FY07/08 (Tsh m) Name FY2006/07 FY2007/08 Total Budget Perf. Total Budget Perf. Immunisation and MCH support % % Reproductive and Child Health Programme (2) % % Support to Maternal Mortality Reduction Programme (3) % % Mental hospital rehabilitation - n/a n/a AIDS Control Programme % 2, % Health Sector Reform Programme - 0% - 0% Total ,462.1 Source: Itemized Appropriations Accounts FY2007/08, MoHSW s Accountants Office, as of 2 September 2008/09 Notes: see Table MoFEA Reports on MoHSW External Finances Since FY2005/06, the MoFEA requires the MoHSW to submit information for all programmes (not only development budget ones) to its External Finance Department. Table 21 gives an overview. The reports to MoFEA give a more comprehensive picture of external financing than the reports on the Development Budget. Programmes sponsored by DPs are included, whether they are part of the development budget or not. Some doubts have been raised on the accuracy of the figures during the research. The Kataa Malaria and the MCH programme were reported by the MoHSW Planning Unit to not have existed. Since the records of the MoHSW have been destroyed in a computer failure, it was not possible to determine how they were included in the MoFEA report. To avoid future ambiguities, it is absolutely crucial that this kind of records are stored and backed up in safe places. 15 This also holds for instances when the amounts reported are suspiciously high or out of line with experience. During the research a data entry mistake (in the MoHSW or the MoFEA) in ZMCP figures was discovered that led to over-reporting of about TSh2.5bn (now cleared). Reporting mistakes of this order can influence DPs funding decisions and need to be avoided. Programmes, Planning Unit and MoFEA External Finances must coordinate to control the quality of the records. 15 In some cases, important information and records only exist in laptop computers or even USB sticks of individuals. There is no data security.

38 22 Zanzibar Health Sector Public Expenditure Review Table 21: External Finances of Programmes, FY2005/06-07/08 (TSh 1m) Programme DPs FY05/06 FY06/07 FY07/08 EPI UNICEF, GAVI, WHO Eye Care Services Programme SSI-UK Family Planning Programme (1) USAID Filariasis Programme WB/WHO/GER SMMRP (2) ADB Health Sector Programme Support III (3) Danida , ,994.0 Health Sector Reform Programme Danida Helminths Programme Health Foundation Human Resources for Health USAID Integrated Management of Childhood Illnesses UNICEF Kataa Malaria (4) USG / PMI / Gl fund - 7, Maternal & Child Health (5) UNICEF Nutrition (6) UNICEF TB and Leprosy GF Reproductive and Child Health Programme UNFPA USAID Zanzibar Aids Control Programme GF,WB,OPEC,GR , Zanzibar Malaria Control Programme GR,GF,WB 3, ,795.3 Total MoHSW 6, , ,261.9 Total MoHSW w/o Kataa Malaria 6, , ,261.9 Child Survival Protection and Dev. (MoFEA) WB ZAC (CMO) UNICEF - 2, Total 6, , ,872.7 Source: Department of External Finance, MoFEA Notes (1) Family Planning: Unclear why FP has been listed separately, should probably be part of RCHP (2) SMMRP: in FY07/08 took over from the Programme for Primary Health Care (name in BEBs) (3) HSPS: In MoFEA records incorrectly named Health Care Programme Support III (3) Kataa Malaria: Planning Unit reported that no programme of this name existed, but that it should be ZMCP. (4) Maternal & Child Health: Planning Unit reported that programme did not exist, possibly mix-up with RCHP. (5) Normally included under Child Survival and Protection (MoFEA) The biggest MoHSW recipients of external funds in FY2007/08 according to the MoFEA records are Zanzibar Malaria Control Programme (ZMCP, TSh3.8bn, 46%), Zanzibar Aids Control Programme (ZACP, TSh870m, 11%), Health Sector Programme Support (HCPS, TSh2bn, 24%), and ADB s First Health Rehabilitation Programme (FHRP, TSh715m, 9%). Fewer funds were received by EPI, IMCI and RCH (total TSh454m, 5%). Eye care seems to have done comparably well, capturing another 3%. There is no clear picture as to the trend of the contributions. Some recipients have seen a steady increase in funds (Eye Care, Filariasis, HCPS, IMCI), EPI a steady decrease, and most an up and down, including more than doubling of resources or drops to less than half. Table 21 also includes two programmes that are relevant to the MoHSW, but which are not administered by the MoHSW itself. These are the Zanzibar Aids Commission (ZAC) under the Chief Minister s Office (CMO) and the Child Survival Programme (CSP) under MoFEA. These programmes have a wider reach than the health sector and have been allocated to other ministries as a consequence, but expenditures benefit the health sector substantially.

39 External Financing of the Health Sector PER08 Data Collection: Finances by Recipient PDU Table 22 provides an overview of the external financing data collected for the PER08. The purpose of the own data collection is a more detailed analysis than is possible with the information provided by MoFEA, and taking into account MoHSW departments and units. There are only a few data gaps. 16 Two different sets of information were asked for: One on Funding overview by DP, including budgets and disbursements, and one on Funding details by category/location, including budgets and expenditures. All information collected is included in Table 22. Table 22: External Funds: Budget, Disbursements and Expenditure FY07/08 FY2007/08 Budget - Overview Budget - Details Disbursements - Overview Expenditure - Details ADB 3,607 3, ZBTS CHS CMS ,187 1,187 DSW EPI Eye Care HCEU Helminths HMIS HSF HSPS HSRS IMCI Mental Nutrition RCH TB/L Training ZACP (1) 5,802-3,804 2,219 ZMCP 6,407 4,216 4,216 4,216 Total 19,900 12,318 13,130 11,868 (1) ZACP did not provide a budget in the reports on detailed spending Comparing this information to the MoFEA estimates, the overall estimate of development partner contributions (disbursed) has increased, to TSh13.1bn, if. This is to be expected if more sources and funds are covered. The largest recipients in FY2007/08 are again ZMCP (TSh4.2bn, 32%) and ZACP (TSh3.8bn, 29%). The next largest recipients are CMS (TSh1.2bn, 9%), which captures the drug purchases through HSPS, and the HSF (TSh721m, 6%), also funded by HSPS. But there are some worrying differences to the MoFEA data. Eye Care and IMCI reported smaller amounts to the PER team than to MoFEA. For FY2006/07, this is the case for EPI, IMCI, Nutrition and TB&Leprosy. It is not clear why these differences exist and what information can be trusted. The second thing that can be noticed is that PDUs sometimes were not consistent in their reporting of their donors budgets. The two reports usually will have been filled out by the same person (many PDUs did not indicate on who compiled the information) around the same time, and both were included in the same excel file. Still, some programmes reported different budgets in the two sections (DSW, EPI, Helminths, IMCI, and ZMCP), and ZACP did not submit a detailed budget at all. It has not been possible to find out why these differences are present, as follow up was difficult due to time constraints at the reporting units. The divergence shows that DPUs have at times serious problems getting the necessary data together when confronted with a request for information like 16 Follow-up for Dept. Substance Abuse, Environmental Health Unit and Mental Health was unsuccessful.

40 24 Zanzibar Health Sector Public Expenditure Review the PER. The MoHSW should agree on a set of indicators for categories and reporting that it wants to include in the PER08, and integrate these into the standard planning and reporting system. This way, all DPUs would start taking this information into account from the start and also be able to provide the required information in a consistent manner. Comparing budgets from the overview data with disbursements, then it can be seen that the overall budget shortfall has been nearly TSh6.8bn. This is despite the fact that more than 50% of PDUs reported to have received more money than budgeted. Using the budget from the detailed analysis, the picture hardly changes with regard to the percentage of PDUs reporting budget shortfalls or surpluses, but since ZMCPs budget now coincides with the disbursements, the overall budget shortfall is reduced by TSh2.2bn (ZACP has to be excluded in the comparison due to missing data). Comparing disbursements and expenditures it can be seen that two thirds of PDUs have reported that about 100% of disbursements have been spent (ADB, CHS, CMS, DSW, Eye Care, HCEU, Helminths, HMIS, HSF, HSPS, HSRS, Nutrition, TB/Leprosy, Training, ZMCP). But because of funding modalities, budget performance is not a very informative measure in many cases. In the case of ADB, the disbursement and implementing agent are the same. In other cases, DPs do not provide prospective budgets for the year for PDUs, but they work based on a (bi-)annual workplan and release money according to budgets for specific activities. This puts budget and disbursement on par. Remaining funds then have to be retired to the DP, equalizing disbursements and expenditure. The UN agencies work like this. For the PDUs that are by majority funded by a donor with (multi-)annual budgets fixed in advance, e.g. HSPS and GF, however, a budget performance around 100% is very positive and shows that the PDUs have no problem to absorb the funds they were given. CHS (excluding ADB funded activities), CMS, HCEU, HMIS, HSF, HSPS, HSRS and ZMCP are such PDUs. It has not been possible to date to find out what the problems of ZACP had been in expending 40% of their disbursements. Selected programmes, the largest by funding, were visited to enquire where the major difficulties in implementation lay, and if they could effectively implement additional funds if made available to them. Most programmes reported that delays in fund releases, were problematic for keeping up with established workplans. This is a special concern if activities are sequenced. If one activity builds on another, delays accumulate. In a few extreme cases, programmes had to reduce their activities to an absolute minimum due to the lack of funds, e.g. IMCI (did not have any funds for the first half of the year) and TB/Leprosy (last GF disbursement in April 2006, i.e. FY05/06, still waiting for initiation of Phase 2 of grant). The reasons for the more common delays seemed to be on both the funding and recipient side, often related to delays in reporting to funders or evaluation of reports by the funders. 17 Concerning the availability of additional funds, most programmes stated that they could absorb more funds, but also admitted that staff numbers and capacity may become a constraint for absorbing large additional funds. Qualification of new and existing staff is a major concern for them. MMH and other hospitals have not been asked about contributions directly to them, but they often receive some direct contributions. This is an oversight that should be addressed in the next PER. The PER team also did not make efforts to record or quantify in-kind contributions to the MoHSW, due to 17 Particular problems here occurred with GF funding. Reporting to GF Headquarters goes through a Local Fund Agent (PwC Tanzania), which has, according to the Zanzibar GFCCM resulted in increased delays of reports reaching their destination and ultimately disbursements. The GFCCM now copies reports directly to GF HQ.

41 External Financing of the Health Sector 25 the difficulties in estimating values for these contributions. A routine register for in-kind contributions should be established and a methodology for converting these into monetary terms developed; as a minimum, the issue should be looked at in the PER09. This would also capture the contributions of the Governments of China, Cuba and Egypt through medical teams, which are important, but have so far remained outside of the scope of the PER Analysis of External Finances The data collected for the PER08 allow taking the analysis a step further than in previous years. In the following, finances will be analysed by donor, geographic areas, programmatic areas, and spending categories. The data presented here is as reported, which means that some data issues that the PER team was not able to clear up remain in the presented figures Finances by Source of Funding Table 23 provides an overview of the source of the funding to the MoHSW. It is split into the categories Multilaterals, International NGOs and other institutions (INGO), and bilateral government support (USA, Denmark and GoT). The biggest source of external finance has been Multilaterals, followed by bilateral funding from the US and the Danish Government. The contribution by the GoT in FY2006/07 is a reported contribution of the National TB and Leprosy Programme (NTLP) to the Zanzibar TB&Leprosy programme. Table 23: External Finances by Source, FY2006/07-07/08 (TSh1m), DP overview data FY06/07 FY07/08 Source Budget Disbursed Budget Disbursed Multilaterals 3, , , ,753.2 INGO DK 2, , , ,944.0 US 4, , , ,703.0 GOT Total 10, , , ,129.7

42 26 Zanzibar Health Sector Public Expenditure Review Table 24 splits the support by individual development partner. The analysis is incomplete, due to some remaining data problems, but it is clear that most of the external funding comes from a very limited number of sources. The GF has become the most important donor to the Zanzibar Health Sector, having a committed budget of nearly TSh8bn for FY07/08 and disbursed funding of TSh5.3bn. ZMCP, at the time of reporting, was also still waiting for a disbursement scheduled for that year. The second largest donor according to disbursements is Danida, through HSPS, with nearly TSh3bn. These two sources of money account for about 63% of the total funding. PEPRFAR has been the third largest donor, with around TSh2.3bn. ADB has been the fourth large donor with a budget of TSh3.6bn and disbursements of TSh688m. ADB did fund the MoHSW in FY2006/07 through FHRP, but with the end of the programme in that year, the records were repatriated to the ADB headquarters. According to the MoFEA reports, TSh276m were spent in FY2006/07. Adding ADB and PEPFAR to GF and HSPS, 86% of the external funding are accounted for. Other big donors with more than TSh100m support are UNFPA, World Bank, Unicef, CDC, USAID, SSI, and Engender Health.

43 External Financing of the Health Sector 27 Table 24: External Finances by Donor, FY2006/07-07/08 (TSh1m), DP overview data Development Partner FY06/07 FY07/08 (by Disbursement in FY07/08) Budget Disbursed Budget Disbursed GF 1, , , ,314.3 HSPS 2, , , ,944.0 PEPFAR 4, , , ,307.5 ADB , UNFPA CHAI WB 1, , Unicef CDC USAID SSI Engender Health WHO GLRA Contrast TRI GAVI Columbia U Italian Cooperation NTLP Puget Sound Total 10, , , ,129.7 The reliance on only two external funding sources for 61% and on four for 86% is somewhat worrying. A positive consequence of having the GF and Danida as the main funding sources is that this makes longer-term planning possible, as both commit their funding for a five year time-frame. 18 This allows PDUs to plan better, which then benefits the service provision. Most other partners have only annual or bi-annual plans. While RGoZ and the MoHSW have very limited say in this matter, they should seek to convince all donors, that reliable and predictable funding need to be committed over a time-frame that extends the short-term Finances by Location The PER08 data also includes a split of the external finances by geographic location of spending. Budgeted and expended allocations for Unguja, Pemba, and Headquarters (HQ) were asked for, taking into account that many funds spent at the Unguja based benefit both islands. 18 GF funding is subject to a project review after two years, Phase 1, and an adjustment of budgets for Phase 2, but usually Phase 2 is approved and budgets are somewhat in line with initial plans.

44 28 Zanzibar Health Sector Public Expenditure Review Table 25 provides a summary of this analysis. Values differ from the analysis by donor because here expenditures, rather than disbursements, are provided. The figures reported may be smaller or bigger than disbursements, because funds can be carried over sometimes. HQ contains all funding for central level institutions of which most are based in Unguja, but deliver services to both islands, such as HSRS, HMIS, CMS, HCEU and others. Activities of programmes and departments that span both islands are included here, too, if they benefit both islands but benefits cannot readily be specified for each island, e.g. the formulation of policies, strategies, treatment guidelines. Reporting agents were advised to separate out Unguja and Pemba, whenever possible, e.g. training activities with participants from both islands.

45 External Financing of the Health Sector 29 Table 25: External Funding by Geographical Area, FY206/07-07/08 (TSh1m), detailed data on spending by location Location FY06/07 FY07/08 Budget Expenditure Budget Expenditure HQ 2, , , ,412.6 Unguja 1, , , ,344.5 Pemba 1, , , ,692.9 Unallocated 2, , Total 7, , , ,867.9 Most funds in FY2007/08 are recorded for Unguja (TSh5.4bn), followed by HQ (TSh3.3bn) and Pemba (TSh2.7bn). The percentage distribution is shown in Figure 19 and Figure Figure 19: External Budget by Location, FY07/08 Figure 20: External Expenditure by Location, FY07/08 Pemba 21% HQ 45% Unallocate d 1% Unallocate d 3% Pemba 23% HQ 29% Unguja 33% Unguja 45% HQ receives 45%/28% of the budget/expenditure, respectively, Unguja receives 33%/46%, and Pemba 21%/23%. Exlcluding HQ and Unallocated funds than the budget compares 61% to 39% and the expenditure 66% to 34% for the two islands. This is a slightly favourable distribution for Pemba compared to the RGoZ figures (Pemba received 32%/30% of budget/expenditure) Finances by Programmatic Areas As in the PER07, the allocation of external funding by programmatic area is only a rough estimate of the contributions to each sector. All allocations to a programme working in a specific area have been taken to benefit this area, despite there being spill-over effects into other areas. In particular, the amounts stated for the HIV/Aids and Malaria response will be overestimated, with an according downward bias for other areas. Part of the large funding that Malaria and Aids programmes receive is used to strengthen health facilities, health systems, and other sectors in order to effectively provide their services, e.g. through equipment or facility upgrades. Also, some interventions in the priority diseases benefit RCH, e.g. PMTCT in HIV/Aids, IPT for pregnant women in Malaria, or even social welfare, e.g. food distribution to OMVC through HIV/Aids funds. This issue is at the heart of the current discussion about benefits and disadvantages of vertical funding (e.g. Ooms et al. 2008). Table 26 provides an overview of rough estimates for funding by programmatic areas. Due to the nature of the records collected for the PER08, the basis for allocating funds to the programmatic areas is only the recipient programme, and not a mix between recipient programme and donor organisation. ZMCP is Malaria, ZACP is HIV/Aids, and HSS consists of a large number of PDUs: HSPS, ADB (SMMRP), NTBC, CHS, CMS, HCEU, HMIS, HSF, HSRS, PHL and Training. While ADB and ZBTS funding is primarily aimed at RCH and HIV/Aids, respectively, building staff houses and similar installations (ADB) and establishing safe blood transfusion services have positive effects on the whole health system. RCH, IMCI and EPI are captured under RCH. All other PDUs (DSW, Eye Care, Helminths, Mental, and Nutrition) are included under other.

46 30 Zanzibar Health Sector Public Expenditure Review Table 26: External Funding by Programmatic Area, FY2006/07-07/08 (TSh1m), based on funding overview by DP Programmatic Area FY06/07 FY07/08 (by expenditure in FY07/08) Budget Disbursed Budget Disbursed Malaria 1, , , ,215.5 HSS 4, , , ,171.9 HIV/Aids 3, , , ,804.0 RCH Other TB/Leprosy Total 10, , , ,095.8 Malaria is the programmatic area receiving most funding in FY07/08 (TSh6.4bn/4.22bn). This is mostly due to the large sums of GF money going to the ZMCP. The large difference between budgets and expenditures FY06/07 and FY07/08 are due to the fact, that there was no disbursement of GF funds in FY06/07. The operational funding for 06/07 came from the comparably small disbursements under Round 1 and 4 at the end of FY05/06 that were carried over into FY06/07. In FY07/08, however, there were two large disbursements under Round 4, of which one came in at the very start of the year. The money included funds for ITNs and insecticides that are usually ordered several years at the same time. Malaria reduction is one of Zanzibar s success stories, up to the point that a Malaria Elimination Study is currently under way. The large funding thus has had a visible impact. With the Round 8 proposal approved by the GF board, substantial funding will also be available in the future. HSS receives the second highest share of funding, at nearly the same level as Malaria (TSh4.17bn). This is mostly due to HSPS and ADB funding, but also because of the very large contribution by the US Government through PEPFAR to the ZBTS, especially in FY06/07. While the figures seem very large, it needs to be kept in mind that HSS includes a vast area of support, from support of in- and pre-service training, over HMIS and administrative support, to purchasing drugs. It is therefore essential to keep up and if possibly extend the support for health systems. Efforts should be stepped up to get GF funds for HSS as soon as possible. In March 2009, the GF has accepted the HSS component of the Round 8 Malaria proposal, which will mean that over the course of 5 years, up to EUR2.6m could be made available by the GF. 19 HIV/Aids is the third area that received a very large amount of funding (TSh5.8bn/3.8bn). Most of this money has come from PEPFAR, with the GF as the second largest donor. The World Bank (WB) has had a budget of over TSh1bn, but only TSh264m of this was then actually disbursed. The funding from CHAI and Columbia University to ZACP makes up most of the funding from INGOs overall. TB/Leprosy has had only limited funding available in the recorded time. This is, at least partly, due to the postponed start of the Phase 2 support from the Round 3 GF grant. In the absence of GF funds, the available funding has come from the GLRA, Clinton Foundation and NTLP. 19 The approval, pending a number of clarifications, came after the GF Board had turned the application down after the Technical Review Panel had endorsed the HSS component due to funding constraints at the GF.

47 External Financing of the Health Sector 31 Table 27 provides similar information like Table 26 on budgets and expenditure (instead of budgets and disbursements). It is based on the detailed spending reports by location. As budgets have been discussed when looking at Table 26, we will focus here on an analysis of the expenditure.

48 32 Zanzibar Health Sector Public Expenditure Review Table 27: External Funding by Programmatic Area, FY2006/07-07/08 (TSh1m), based on detailed spending by location Programmatic Area FY06/07 FY07/08 (by expenditure in FY07/08) Budget Expenditure Budget Expenditure Malaria 1, , , ,215.5 HSS 4, , , ,215.1 HIV/Aids 0.0 1, ,219.0 RCH , Other TB/Leprosy Total 7, , , ,867.9 In FY07/08, Malaria was the programmatic area, in which most money was spent, with expenditures of TSh4.2bn. Activities funded included strengthening the recognition and care of Malaria in health care facilities, purchasing of ITNs and Monitoring and Evaluation. About an equal amount of money was spent on HSS. Activities are very diverse and include everything from spending on drugs, over information systems, to physical infrastructure. The third biggest area was HIV/Aids with TSh2.2bn. Major expenditures were Care and Treatment, Counselling and Testing, and PMTCT. In the area of RCH, TSh834m were spent on various activities for improving the access to RCH services, children s vaccinations, and awareness-raising Finances by Categories Table 28 provides a summary as-reported analysis of the information obtained on spending by category from external funding. ADB data for FY06/07 is not included (see section 4.2.1). It should also be noted that there are many problems remaining with the data: from categorisation problems (activities need to be allocated and it may have been impossible to separate out different inputs), to missing records and the need to make estimate based on information for other years. Appendix 4 provides an overview of the data issues, and an analysis by programme. Table 28: External Funding by Category, FY2006/ /08 (Tsh1m) FY2006/07 FY2007/08 Category Budget Expenditure Budget Expenditure Total % Total % Total % Total % Drugs and medical supplies 2, , , , In-service training 1, , , , Other running costs (stationery, minor repair) , Staff costs (salary top-ups, allowances etc) 1, , , Transport and fuel , , Total Recurrent Costs 5, , , , Long-term training Other capital costs Physical Infrastructure 1, , , Purchase of equipment (including vehicles) Total Capital Costs 1, , , , Total 7, , , , It can be seen clearly that the large majority of external funds go into recurrent expenditure, even more so in FY07/08 than in FY06/07. This holds true both for budgets and expenditures. Within recurrent expenditure, the biggest item by a long way is drugs and medical supplies (30%/37% in FY07/08). The other categories share the rest fairly equally. Staff costs in FY07/08 are reported here with TSh1bn (11%), but it is difficult to say how accurate the figure is. When classifying expenditure for some of the programmes, the PER team allocated e.g. planning workshops to staff costs, as by experience the major costs are incurred through allowances.

49 External Financing of the Health Sector 33 But this misallocates some other expenditure, e.g. rent costs, and inflates the staff cost figure. But it is also suspected that some programmes will have allocated staff costs to other categories. This ambiguity could be resolved by training informants for the classification, but it would be feasible only if the reporting by categories would become part of the routine data collection. In any case, it shows that staff can improve salaries substantially by taking part in extra activities, and that DPs have taken on a strong role in training. Figure 22 and Figure 21 present the shares of the single categories within recurrent and capital spending (different from Table 28). Figure 22: External Recurrent Expenditure FY07/08 Figure 21: External Capital Expenditure, FY07/08 Transport and fuel 15% Staff costs 14% Other running costs 13% Drugs and medical supplies 43% In-service training 15% Purchase of equipment 41% Long-term training 8% Other capital costs 2% Physical Infrastructure 49% Reported capital spending only makes up a minor fraction of external funding, less than 30% and 20% for FY06/07 and FY07/08, respectively. Most of this is then spent on equipment or physical infrastructure. Physical infrastructure is usually big individual projects, so this figure can be subject to large fluctuations over the years. With plans by ADB to fund substantial building work, this figure will go up again, after recording very low in FY07/08. Regarding the need to improve the capacity of MoHSW staff, a higher allocation for long-term training would be positive. All in all, the capital spending seems too low, even considering the MoHSW s position that there is no need to construct significantly more health facilities (MoHSW Zanzibar 2007, p.51) Districts: Health Service Fund The Health Service Fund (HSF) was introduced in FY2004/05 under the current Danida HSPS. In the absence of any substantial government funding for operational costs for health facilities, the HSF provides the major part of funding for day to day operations of the DHMTs in running district health services. It is meant to be a precursor to a health basket for recurrent budget support to the ZHMTs and DHMTs in Zanzibar, which would include contributions from RGoZ, Danida HSPS and other DPs. So far, no other DPs have joined. Very recently, however, the GF approved a grant for HSS as part of the Round8 Malaria proposal. This HSS grant also includes a contribution to the HSF, so that pending the grant agreement and technical questions, the HSF will turn into a proper basket. There has also been interest from UN Agencies, and proposals have also been submitted to GAVI. The HSF allocation between districts is determined according to a weighted capitation formula based on population, land area, the U5MR as a proxy for burden of disease, and poverty. Relative weightings are currently as follows: population 50%; land area 10%; U5MR 15%; and poverty 25%. Each district contributes 7% of its HSF allocation to the HCEU and 3% to the ZHMTs. In addition, contributions are made as follows to hospitals within the district in recognition of their contribution to primary health care delivery to the catchment populations:

50 34 Zanzibar Health Sector Public Expenditure Review 20% from DHMT to a PHCC in the district (South for Makunduchi, North A and North B each for Kivunge, Micheweni, Chake Chake for Vitongoji); 30% from DHMT to a District hospital (Wete and Mkoani), reduced to 20% for Chake Chake; 30% from Urban and West district to MMH. The formula provided here takes into account the change of the formula in 2006 and is therefore different from previous PERs. The outcome of this distribution is shown in Table 29 shows the HSF by level and by zone, for both budgets against expenditures. In FY07/08 ZHMTs received 3% and 5% of total budgets and expenditures, respectively. DHMTs received 69%/70%, and hospitals 28%/26%. Unguja received 60%/62% and Pemba the remaining 40%/38%. Since all stakeholders agreed to the use of the allocation formula, this share can be used as a fair yardstick against which the allocations from the RGoZ an external funding can be measured (see section 6.1.1). Table 29: HSF by level and zone, FY2006/07-07/08 (TSh1m) FY2006/07 FY2007/08 Bdgt. Expd. Bdgt. Perf. Bdgt Expd Bdgt. Perf. ZHMTs % % DHMTs % % Hospitals % % Total % % Unguja % % Pemba % % Total % % The figures show that the HSF has fully matured. The PER07 reported that for FY06/07, there had been a substantial improvement in budget performance compared to previous years. FY07/08 has seen a further increase, to above 100% in all categories. A 100%+ budget performance is possible because unspent funds can be carried forward into following years. The only DHMT that did not exhaust its budget was Urban (92%). Hospitals more often did not use up their full allocation, these were Chake Chake DH (99%), Mkoani DH (90%), Vitongoji PHCC (79%), and Kivunge PHCC (92%). The increased budget performance is remarkable in view of the increased overall funding. Regarding the financial management capacity, the external audit for FY2006/07 carried out by PWC showed that there are no concerns. The few issues initially criticised by PWC were resolved subsequently and there is full accountability. Similarly, the 2007/08 PKF audit only contained some minor complaints about the HSF, e.g. missing or incomplete documentation for a small number of activities. This is a positive development that will hopefully strengthen the argument for other partners to join the HSF. When DHMTs and hospitals in the districts were visited on both islands, all commended the HSF for disbursements in full and on time. Since RGoZ budgets are comparatively small and disbursements very unreliable, both in amount and time, and other donor s contributions are mostly ad-hoc, the HSF funds are seen as the absolute basis for planning and effective service provision. As in previous years, it is recommended that the MoHSW establish a subvote for District Health Services to integrate the HSF fully into its own administration and accounting system, and put in a contribution of its own (preferably ring-fenced). This would show the commitment of the MoHSW to the districts (and especially primary care) and further encourage other development partners to join a health basket for district services. Importantly, Danida supports and encourages this integration of HSPS funds into the MoHSW budget.

51 35 Zanzibar Health Sector Public Expenditure Review 5. Cost-Sharing in Public Health Facilities 5.1. Total Income The situation concerning cost-sharing still has not changed substantially since the Rapid Assessment of Cost-Sharing (Lake 2007b). Cost-sharing is widely practiced, all hospitals and PHCCs use it, and even a few PHCUs have started collecting some money from the patients or the general public to improve service quality and availability, but it still has not been formally introduced in Zanzibar. It has developed spontaneously as a result of fund shortages at the service delivery point, like in many other countries (Schieber 2006, p.16). There has been some work on the conceptual side, a new set of Cost-Sharing Guidelines has been produced at the end of the year 2007, and there is work ongoing for piloting a viable waiver mechanism and a risk-sharing scheme, but none of these initiatives have been implemented yet. The Cost-Sharing Guidelines which were meant to take up a number of the recommendations from the Rapid Assessment, have been approved by the Leadership Committee of the MoHSW, but have been awaiting submission to the Revolutionary Council for some time now. It is not clear when this step will be taken. Despite not having been formally introduced, the MoFEA Budget Estimate Books contain an estimate for income from cost sharing of the MoHSW. Interestingly, the information provided there indicates that the MoHSW is allowed to retain 100% of its income. Table 30 shows the estimates, for the total (which possibly includes non-cost-sharing charges like port health charges), and for curative services. Table 30: Projected income of the MoHSW, MoFEA estimates (TSh 1m) FY07/08 FY08/09: Total Curative (X-rays etc): These estimates however, do not in any way reflect the real numbers. Table 31 provides an overview of the income and expenditure from cost sharing in all the hospitals in Zanzibar. Table 31: Actual income from Cost Sharing in FY2007/08 (TSh) Zone Name Income Expenditure Pemba Chake Chake DH 4,420,700 4,696,500 Mkoani DH 5,231,100 4,724,773 Wete DH 3,426,288 4,521,000 Micheweni PHCC 1,565, ,000 Vitongojii PHCC 305,100 - Pemba Total 14,948,488 14,224,273 Unguja Kivunge PHCC 321,000 - Makunduchi PHCC 1,302, ,000 Mnazi Mmoja Hospital 151,069, ,684,754 Unguja Total 152,692, ,354,754 Grand Total 167,641, ,579,027 The total income from cost-sharing is TSh167m. Table 31, but even more so Figure 23 show clearly that the vast majority of this income is concentrated in only one hospital, Mnazi Mmoja. 90% of the total funds from cost-sharing were contributed there. The other seven hospitals collected a total of about TSh16.5m, of which TSh13m came from the three district hospitals and TSh3.5 from the four PHCCs. This means that, on average, DHs only had a declared income of TSh5.5m in a whole year, and PHCCs even less than TSh1m.

52 36 Zanzibar Health Sector Public Expenditure Review The Unguja hospitals keep their income and spend it directly, while the Pemba hospitals are required to submit their income through the Zonal Office to the MoFEA, from where they have to request it if they plan expenditures. Why there is this difference is not clear. Figure 23: Shares from CS income, FY07/08 The income estimates are the values stated by the hospitals when visited, the expenditure for Unguja also from the hospitals, while the estimates for Pemba come from the Zonal Office (with the exception of Mkoani DH). The income figures collected from the Zonal Office Pemba often differed from the figures stated by the hospitals, which indicates either that hospitals do not submit all income to the Zonal Office (and then probably also have higher expenditures than captured here), or that there are problems in book-keeping. 90% Chake Chake Mkoani Wete Makunduchi Micheweni Vitongojii Kivunge Mnazi Mmoja 5.2. Current Charging Practices The delay in putting the guidelines on cost-sharing in place has left the fee structures in most places untouched since the PER07 and the Rapid Assessment of Cost Sharing. Table 32 provides an overview over charges for selected services at the hospitals. In some cases, it is not clear if blanks are omissions in the submitted information, if the service is not charged for, or if it is not provided at all. If a service is provided but not charged for, then a - is placed in the cell. Table 32: Charges at hospitals for selected services, as of December 2008 (TSh) Dept. Service MMH Chake DH Wete DH Mkoani DH M weni PHCC V goji PHCC Kivunge PHCC M duchi PHCC Lab Blood ESR Blood Grouping 1, ,000 Blood Picture 1,000 1,000 1,000 1,000 - Blood Sugar 1, ,000 1,000 1,000 1,000 Haemoglobin 300 1, Stool routine Urine Pregnancy 1, ,000 2,000 1,000-1,000 Urine Routine X-Ray Uniform Price - - 2,000 2,000 - n/a - - Standard Bone 2,000 2, ,500 n/a 1,500 1,000 Head 4,000 3, ,000 n/a 3,000 2,000 Hip Joint 4,000 3, ,000 n/a 3,000 2,000 Spine 4,000 3, n/a 3,000 2,000 Stomach (Barium) 12,000 12, ,000 n/a 10,000 10,000 Ultrasound Ultrasound 6,000 4,000 4,500 n/a Dental Extraction 1,000 1,000 1,000 1, Maternity Charge - 1,000 1,000 1,000 1,000 1, bring own items yes yes yes yes yes yes yes yes Usually, MMH is the most expensive hospital for the services provided and charged for, with the exception of some laboratory tests that are more expensive at Chake Chake DH. Overall, the prices at different levels seem to be in line with the referral chain, although there is only limited information available for Vitongoji PHCC. Not included in this table are the fees for operations at MMH, which range from TSh3,000 for a circumcision to TSh100,000 for a prostatectomy, and most being in the range TSh20,000-50,000.