Public expenditure review for Tanzania prepared by the World Bank

PUBLIC EXPENDITURE REVIEW FY04

Report on Fiscal Developments and Budget Management IssuesFY03-FY04

June 14, 2004

TABLE OF CONTENTS

Introduction..........................................................................................................................i1. Overall Fiscal Performance.....................................................................................1

Aggregate Fiscal Discipline.............................................................................1Domestic Revenue...........................................................................................3Findings from the Investment Climate Survey on the Tax System.................4Donor Assistance.............................................................................................6Government Expenditures...............................................................................6

1.2 Strategic Resource Allocation.............................................................................9Use of the Contingency under Vote 50..........................................................14

1.3 Priority Sector Expenditures..............................................................................15Priority sector trends 1999/2000 – 2003/04...................................................15Actual expenditures on the priority sectors during the first half of FY04.....17

2. Consistency in budget planning and execution.....................................................193. Strengthening budget process................................................................................21

Strengthening the Budget Guidelines............................................................22Implications for other planning instruments..................................................24Implications for donor consultations on the budget......................................24

4. Strengthening a result-based approach in policy making and budgeting..............25Introduction....................................................................................................25Monitoring systems.......................................................................................25Developing a Result-Based Tanzania PRS....................................................27Performance-compatible incentive structures................................................29

5. Public Expenditure Management Issues at Local Government Level...................30Recurrent vis-vis Development Expenditure.................................................34Expenditure on Personal Emoluments and Other Charges at LG level.........36Emerging Issues from the preliminary analysis of actual LG expenditure. . .38

ANNEX Consistency between Policy, Planning, and Budgeting..................................40A.1 Introduction........................................................................................................40A.2 The Evolution of Planning, Programming, and Execution................................40A.3 Defining the Budget Data..................................................................................42A.4 Overall Budget Trends.......................................................................................43

Data Sources and Definitions........................................................................43Within-Year Consistency of Aggregate Expenditure Estimates (Budget Guidelines – MTEF – Budget – Actual Expenditures)..................................44Aggregate Resource and Expenditure Estimates...........................................44Assessment of Outer Year Estimates.............................................................47Overall Trends...............................................................................................47Recurrent Expenditure Sub-Trends...............................................................49Analysis of budget deviation at sub vote level..............................................52

A.5 Trends in the Priority Sectors............................................................................53Data Sources and Definitions........................................................................53Overall Trends...............................................................................................53Sector Trends.................................................................................................55

A.6 Conclusion.........................................................................................................57

BOXESBox 1: Supplementary Budget and Additional Expenditure Needs in FY04................................13

FIGURESFigure 1: Ability to Finance Recurrent Expenditure from Domestic Revenue...............................2Figure 2: Share of Development Expenditure Financed from Non-Project Support Resources.....3Figure 3: Percent of enterprises rating tax rates and administration as major or very severe

obstacles...................................................................................................................................5Figure 4: LGAs – Sectoral Shares of Recurrent Expenditure 2001...............................................35Figure 5: LGAs – Sectoral Shares of Recurrent Expenditure 2003...............................................36Figure 6: LGAs – Sectoral Shares of Exchequer Issues to OC 2003............................................37Figure 7: LGAs – Sectoral Development Expenditure 2001.........................................................37Figure 8: LGAs – Sectoral Development Expenditure 2003.........................................................38Figure 9: Differences in Total Resource Estimates.......................................................................44Figure 10: Domestic Revenue, Program Grants and Loans, and Project Grants and Loans,

Difference between Budget and Budget Guidelines, FY00 – FY04......................................45Figure 11: Total, Recurrent and Development Expenditure, Difference Between Budget and

Budget Guidelines Estimates, FY00 – FY04.........................................................................45Figure 12: Wages and Salaries, Operations and Maintenance, and Debt Service, Difference

between Budget and Budget Guidelines, FY00 – FY04........................................................46Figure 13: Total Expenditure, Budget Guidelines, Annual Budget, and Actual Expenditure, FY00

– FY07....................................................................................................................................47Figure 14: Recurrent Expenditure, Budget Guidelines, Annual Budget, and Actual Outturn,

FY00 – FY07.........................................................................................................................48Figure 15: Development Expenditure, Budget Guidelines, Annual Budget, and Actual Outturn,

FY00 – FY07.........................................................................................................................49Figure 16: OC Expenditure, Budget Guidelines (BG), Annual Budget, and Actual Expenditure,

FY00 – FY07.........................................................................................................................50Figure 17: PE Expenditure, Budget Guidelines (BG), Annual Budget, and Actual Expenditure,

FY00 – FY07.........................................................................................................................51Figure 18: CFS Expenditures, Budget Guidelines (BG), Annual Budget, and Actual Expenditure,

FY00 – FY07.........................................................................................................................51Figure 19: Recurrent Expenditure of Selected Priority Sectors (Education, Health, Water, and

Roads), Budet Guidelines, MTEF, Annual Budget, and Actual Expenditures, FY00 – FY04................................................................................................................................................54

Figure 20: Selected Priority Sector Expenditures, Difference between 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04................54

Figure 21: Recurrent Expenditure on Education, 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04..............................................................55

Figure 22: Recurrent Expenditure on Health, 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04..............................................................56

Figure 23: Recurrent Expenditure on Water, 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04..............................................................56

Figure 24: Recurrent Expenditure on Roads, 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04..............................................................57

TABLESTable 1: Financing of the Fiscal Deficit (% of GDP), FY98-FY04.................................................1Table 2: Revenue Performance (% of GDP) FY98-04....................................................................3Table 3: Composition of External Assistance (% of GDP), FY98-04.............................................6

Table 4: Government Expenditure (% of GDP), FY98-04..............................................................7Table 5: Civil Service Employment FY99-FY04 (December of each year)...................................8Table 6: Civil Service Average Salaries, FY98-FY04....................................................................8Table 7: Composition of Public Expenditures as % of GDP, FY96-FY04.....................................9Table 8: Functional Allocation of Recurrent Expenditure (actuals, % of GDP), FY96-FY04......10Table 9: Social Sector Recurrent Expenditures (actuals, % of GDP), FY96-FY04......................11Table 10: Sectoral Development Expenditures (actuals, % of GDP), FY96-FY04......................12Table 11: Utilization of the Contingency Under Vote 50, FY03...................................................14Table 12: Utilization of the Contingency Under Vote 50, July-December 2003..........................14Table 13: Trend on Priority Sectors (RE + DEV) in Absolute and Relative Terms Over the Years

................................................................................................................................................15Table 14: Budget Deviation at Sub-Vote Level (average as % original budget)..........................21Table 15: Summary of CAG Reports on Local Authority Accounts 1999 – 2002.........................32Table 16: Indicators of Weak Public Financial Management in LGAs 1999 - 2002.....................33Table 17: Total Local Government Expenditure (T.Shs.)..............................................................34Table 18: Total Local Government Expenditure (%).....................................................................34Table 19: Local Government Recurrent and Development Expenditure......................................35Table 20: The Budget Preparation Process....................................................................................42Table 21: Composition of Recurrent Expenditure.........................................................................50Table 22: Budget Deviation at Sub-Vote Level (average as % original budget)..........................52

INTRODUCTION

1. Since 1998, public expenditure reviews in Tanzania have been conducted on an annual basis, closely aligned with Government’s budget cycle and carried out under the direction of the Public Expenditure review (PER) working group, chaired by the Ministry of Finance (MoF) and including in its membership a wide range of stakeholders from Government, development partners and Tanzanian civil society. This approach has been consistent with the series of initiatives in Tanzania aiming at developing an open process of formulation of policy and budget strategy.

2. Without departing from the broad objectives of this approach, there are several recent developments, or emerging concerns, which present a new context for approaching the PER process and new questions on how it can best support national policy objectives. These developments suggest the need to look again at how the PER process works, how it relates to other processes in the policy and budget cycle, and in the interface between development partners and domestic stakeholders. Some of these factors may be briefly discussed here as background to the current year’s PER external evaluation, as described below.

Approaching a new PRS cycle: Review of the Poverty Reduction Strategy entails also taking a fresh look at the role of the PRS in the process of national policy formulation and its linkage with budget process. Government’s guidelines on preparation for the new PRS recognize the need to develop the role of the PER process as a means of interface with domestic stakeholders. In parallel with these institutional issues, the new PRS cycle calls for review of the public expenditure strategy and performance over the past few years and a contribution to a vision for the next cycle

Multiplication of processes: Since development of the annual PER process, a wide range of broad processes have been developed, engaging many of the same players on similar issues, including the PRBS, PRSC, and PRS processes. Processes have accumulated without adequately integrating requirements or streamlining the cumulative burden on transaction costs. Managing these multiple processes imposes severe demands on government, which may have diluted the quality of interaction, leading to a sense of frustration. Development partners have recognized the need for harmonization of procedures to reduce transaction costs and net burden on Government of donor procedures.

Developing accountability to domestic stakeholders: There is growing recognition of the risks that procedures, which support Government accountability to donors, can undermine the growth of accountability to domestic stakeholders. This raises the question whether processes of consultation between Government

i

and development partners can be designed, or redesigned, to meet the dual objective of fostering accountability to domestic stakeholders while also meeting development partner concerns.

Defining priorities: Both Government and development partners have expressed interest in moving beyond a definition of priorities in terms of ‘priority sectors’, a designation which has been highly influential in the PRS, in the budget process and in the focus of attention within PER work. Blanket definition of ‘priority sectors’ can tend to create disincentives for a strong focus on results and efficiency on both side of the dividing line between priority and non priority sectors: if priority areas are predefined with respect to inputs, incentives for ‘priority sectors’ are weakened, while non priority sectors consider their activities unrecognized. Focusing dialogue on the net movement in allocations to priority sectors may also distract from broader problems, including the important issue of consistency between policy and resource allocations across the whole of government. While the category of priority expenditures may be a necessary interim tool for purposes of monitoring and dialogue, there is a need to develop alternative approaches, which shift attention to priority objectives at the level of results, and aim to ensure that all government expenditure contributes to priority objectives.

Consistency through the policy and budget cycle: Rationalizing the process of consultation requires on the part of all stakeholders a high level of trust in the consistency of the process. A simpler less demanding interface with external stakeholders may be easier if participants have a high level of confidence that what is agreed at the level of policy will be carried through into budgets, that budgets will be implemented fully as approved and that monitoring and evaluation will provide clear accountability for results Stakeholders need to know that the framework discussed at one stage in the process will remain the basis for subsequent stages, with clear information on what has changed and why. A clear demonstration of consistency in the chain of steps between PRS, Budget Guidelines, MTEF, annual budgets, actual expenditure and subsequent results, is not only desirable in terms of transparency and accountability – it is also an important precondition for developing processes of consultation which are less demanding and more effective, both for development partners and domestic stakeholders.

3. Following review of these issues with Government and the PER working group, it was agreed that a work program should be developed which, while meeting the normal objectives of the review of fiscal performance in the period, would also provide the basis for tackling these longer term concerns with public expenditure management and the institutions for donor interface with the budget process. The work program was expected to cover a set of interrelated areas, including: analysis of problems in consistency between public expenditure planning, budget preparation and execution, institutional analysis of the existing budget process, including the role of donors, civil society and the legislature, M&E systems, sectoral public expenditure analysis, the strengthening of treasury systems, and local government expenditure. It was recognized that the required

ii

work program would extend beyond the current year’s exercise. The present report is therefore an interim report on work completed so far: some of the topics are the subject of ongoing work and others will be carried through in the next cycle.

4. First, a core concern was the analysis of the recent record on consistency through the budget process, including consistency between the PRS, Budget Guidelines/MTEF, annual budget and its implementation. Given the overlap between this analysis and the overview of fiscal performance normally carried in the external evaluation, these two topics form the first two sections of the report below. An annex is also included giving more detailed analysis of the issues of consistency.

5. A key area is institutional analysis of the budget process. Work is ongoing in this area, but preliminary conclusions and recommendations are included in section 3. Further work in this area is also ongoing on the specific topics of the role of civil society and the legislature in budget process and the integration of human resource planning in budget formulation.

6. In section 4, the report reviews issues in shifting towards a greater focus on outputs and current status and plans development of M&E functions. As noted in the last PER report, a large number of monitoring processes exist: the main task is of rationalization and strengthening systematic linkages to PRS objectives and to the budget

7. The final section reports on local government expenditure and development of capacity for public expenditure management at local level.

iii

1. OVERALL FISCAL PERFORMANCE

Aggregate Fiscal Discipline

1.1 In recent years, expenditures have been rising faster than domestic revenues leading to a higher fiscal deficit before grants, which is projected to rise to 10.5 percent of GDP in FY04. The increases in expenditures were made possible by increased availability of development assistance and, until recently, government refrained entirely from domestic borrowing in line with the objective of sustaining macro-economic stability. However, in FY03 and FY04, government had to supplement financing from foreign sources with domestic borrowing to close the gap between expenditures and domestic revenue. In FY03, domestic financing came to 0.4 percent of GDP and is likely to increase to 0.8 percent in FY04. While such limited access to domestic borrowing is unlikely to have any significant impact on macro-economic stability, it will nonetheless be necessary to monitor closely government access to domestic credit in order to avoid strangling the slowly re-emerging private sector demand for credit.

Table 1: Financing of the Fiscal Deficit (% of GDP), FY98-FY04

Central Government Operations FY98 FY00 FY02 FY03 FY04 FY04Actual Actual Actual Actual Budget. Proj.

Total Revenue 12.0 11.3 12.1 12.8 13.1 13.2Total Expenditure 15.7 17.4* 17.6 20.8 22.5 23.7

Balance before Grants -3.7 -6.2 -5.6 -8.2 -9.3 -10.5Grants 3.0 4.5 4.5 6.5 6.3 6.5Balance after Grants -0.7 -1.7 -1.1 -1.7 -3.1 -4.0Foreign Loans (net) 1.0 1.5 1.4 2.1 2.6 3.2Balance After Grants and Foreign Financing

0.3 -0.3 0.3 0.4 -0.5 -0.8

Domestic (net) -0.3 0.3 -0.3 -0.4 0.4 0.8 Bank -0.9 -0.4 -0.7 -0.1 0.4 0.4 Non-bank (net of amortization) 0.5 0.5 0.4 -0.3 0.0 0.3 Privatization Funds 0.1 0.2 0.0 0.2 0.1 0.0* Net of parastatal recapitalizationSource: IMF and Tanzanian authorities

1.2 The primary macro-economic concerns arising from Tanzania’s fiscal situation relate to the effects of increased aid flows. These are in two areas. Firstly, increased aid inflows bear the risk of leading to Dutch disease effects, which could hamper the competitiveness of the Tanzanian economy. Both international evidence as well as real effective exchange rate developments in Tanzania indicates that it is unlikely that on aggregate aid inflows have caused a real appreciation of the exchange rate of the Tanzanian Shilling.

1.3 Second, the share of public expenditures financed through foreign aid has increased from around 25 percent in FY98 to more than 40 percent in FY04. Even taking into account that part of this increase reflects better capture of aid rather than additional aid flows, it nonetheless highlights the aid dependency of Tanzania’s public sector and the related risk to the sustainability of current expenditure levels. While recent work on the cost of achieving the MDGs indicates that current MDG related public expenditures are still

1

insufficient, sustaining or expanding current levels of foreign aid will require a clear demonstration that both domestic and foreign funds are effectively used and tangible progress is made in moving towards the MDG targets.

1.4 A specific facet of the increasing degree of aid dependency is illustrated by Figure 1, which shows Tanzania’s ability to finance recurrent expenditure from domestic revenue. Until FY99, domestic revenue exceeded recurrent expenditure. However, since FY00, domestic revenue was insufficient to finance recurrent expenditure and the coverage has indeed been declining. In FY04, more than 20 percent of recurrent expenditure are foreign financed through budget support. The significance of this is that the ability of Tanzania’s government to carry out its ongoing business is now significantly exposed to fluctuations in budget support provided by donors. One the one hand, donor financing of recurrent expenditures reflects one of the benefits of budget support which is precisely to allow government to make decisions as to whether support should be used to enhance development or recurrent spending. In the past, the limitation of using donor support for development expenditures has often led to imbalances between recurrent and development expenditures and under-funding of recurrent expenditures necessary to maintain and operate investments in development projects. Nonetheless, the fact that more than 20 percent of recurrent expenditures are dependent on donor support underlines the need to raise domestic revenue. It also underlines a point raised in previous PERs, i.e., Government needs to carefully consider the use of budget support and refrain from automatically using budget support primarily for recurrent expenditures while relying on the financing of development expenditures primarily on project support.

Figure 1: Ability to Finance Recurrent Expenditure from Domestic Revenue(Domestic Revenue as a percentage of Recurrent Expenditure)

0%

20%

40%

60%

80%

100%

120%

FY98 FY99 FY00 FY01 FY02 FY03 FY04

1.5 The development budget has traditionally been financed through donor-supported projects. A positive development is that the share of development expenditures finance from domestic revenue and budget support has indeed been increasing since FY00, reflecting the increased availability of program support and debt relief.

2

Figure 2: Share of Development Expenditure Financed from Non-Project Support Resources

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%


Domestic Revenue

1.6 Improvements in tax administration lead to an increase in the revenue to GDP ratio. Improvements in tax administration such as the set up of a large tax payers department, curbing tax evasion related to the import of petroleum products, and reducing the opportunities for the abuse of tax exemptions through the use of voucher schemes have led to substantial increases in the revenue to GDP ratio. The introduction of a new income tax act in 2004 is expected to lead to further increases in the revenue to GDP ratio. TRA aims at increasing the revenue to GDP ratio to the range of 15 to 16 percent in the medium term.

Table 2: Revenue Performance (% of GDP) FY98-04

Item FY98 FY00 FY02 FY03 FY04 FY04Actual Actual Actual Actual. Budget Proj.

Total Revenue 12.0 11.3 12.1 12.8 13.1 13.2Tax Revenue 11.0 10.1 10.9 11.6 12.0 12.0 Taxes on Imports and Exports 3.5 3.2 4.7 5.0 VAT and Excise on Local Goods 2.7 2.6 2.5 2.4 Refunds 0.0 0.1 0.4 0.0 Income Taxes 2.9 3.0 2.6 2.9 3.0 2.9 Other taxes 1.9 1.3 1.1 1.2 1.3 1.1Non- tax Revenue 1.0 1.2 1.2 1.2 1.1 1.2

Source: IMF and Tanzanian authorities

1.7 While increasing domestic revenue is important to reduce Tanzania’s aid dependency, it is also important to keep the impact of Tanzania’s tax regime on growth and

3

investment in focus. A recently completed Investment Climate Assessment highlights potential constraints imposed by the current tax regime.

Findings from the Investment Climate Survey on the Tax System1

1.8 Almost 73 percent of enterprises in Tanzania rated tax rates as a major or very severe constraint on enterprise performance and growth – considerably more than rated any other obstacle as a major constraint. Although tax rates are rated as a major constraint on operations and growth in many countries – for example, enterprises were more likely to rate high tax rates as a major problem than any of eight other constraints in almost two-thirds of the countries in the 1999 World Business Environment Survey – enterprises appear to be especially concerned about high tax rates in Tanzania.2 Among the 35 countries for which investment climate assessments had been completed by mid-2003, enterprises were more likely to rate tax rates as a serious obstacle in only one country (Ethiopia).

1.9 Although many enterprises rate tax rates as a major problem, this does not imply that tax rates on businesses should necessarily be reduced. On average, rates for individual taxes in Tanzania do not appear to be significantly out-of-line with rates in other developing countries. For example, corporate tax rates are 30 percent in Tanzania, China, Kenya and Uganda and 40 percent in India. The base rate for the VAT is slightly higher in Tanzania than it is in the other countries (20 percent in Tanzania compared to between 14 and 18 percent in China, Kenya and Uganda), but is not far out of line with the comparator countries.

1.10 It does, however, stress the need for broadening the tax base so that the burden on formal enterprises can be reduced without compromising fiscal stability. For example, the VAT efficiency ratio (the ratio of VAT revenues to GDP divided by that VAT rate) is lower in Tanzania (0.20) than it is in other countries in sub-Saharan Africa (0.27). Although the benefits of some recent reforms in tax administration (e.g., the establishment of the Large Taxpayer Unit in 2001 and the introduction of taxpayer identification numbers in 2000) are likely to be felt in the future, continued improvement could reduce the burden of taxation on formal enterprises. A second related issue is that tax exemptions and incentives in some important sectors of the economy have taken a toll on tax revenues.

1 From “Investment Climate Assessment: Constraints on Enterprise Performance and Growth in Tanzania” RPED, The World Bank, preliminary draft, February 7, 20042 From “Investment Climate Assessment: Constraints on Enterprise Performance and Growth in Tanzania” RPED, The World Bank, preliminary draft, February 7, 2004

4

Figure 3: Percent of enterprises rating tax rates and administration as major or very severe obstacles

73%

34%46% 48%

68% 74%

56%

24%

46%36%

51%60%

0%

20%

40%

60%

80%

Tanzania China Pakistan Uganda Kenya Ethiopia

Tax Rates Tax Administration

Source: Investment Climate Assessments

1.11 Despite the recent reforms intended to improve tax administration (see above), results from the Investment Climate Survey suggest that tax administration remained a serious problem even in mid-2003. Enterprises in Tanzania were more likely to rate tax administration as a major or severe problem (56 percent of enterprises) than enterprises in Kenya (51 percent), Uganda (36 percent), China (24 percent) or Pakistan (46 percent) (see Figure 3) – although enterprises in Ethiopia were more likely to rate it as a major or very severe obstacle.

1.12 Quantitative data from the Investment Climate Survey also supports the idea that tax administration is particularly burdensome in Tanzania. Enterprise managers in Tanzania reported that they spent about seven days dealing with inspections or required meetings with officials from the tax inspectorate in 2002/03. In comparison, enterprise managers in Kenya, Uganda, and China reported only spending 2-3 days in meetings. In general, Tanzanian firms that are more productive spend more time meeting tax officials and with tax inspections. Although this might seem puzzling, one reasonable explanation for this is that officials target enterprises that are more productive. Despite the reforms in tax administration, enterprise managers did not generally report any improvement in 2002 – although 36 managers reported fewer meetings with tax officials in 2002 than in 2001, 59 managers reported more meetings and 149 reported the same number.

1.13 Corruption in tax administration is a problem in Tanzania. Despite recent reforms – including the reforms when the Tanzania Revenue Authority (TRA) was formed in 1996 that were intended to reduce corruption in tax administration – 21 percent of enterprises that had required meetings with tax inspectors reported that gifts or informal payments were expected or requested during the meetings.3 In comparison, 7 percent of enterprises that had required meetings in Uganda, 21 percent of enterprises in China and 38 percent of enterprises in Kenya reported that gifts or informal payments were requested. The median

3 Reforms included reforms designed to limit political interference in tax administration and to allow the authority to pay salaries that were higher than they could if the agency remained part of the civil service (see Fjeldstad, 2002)

5

value of the gift/informal payment was T.Shs. 400,000 (about US$400 in mid-2003) in Tanzania.

1.14 Although it is difficult to draw strong conclusions about tax evasion from investment climate surveys – enterprise managers are unlikely to be especially forthcoming with respect to tax evasion – tax evasion appears to be a more serious problem in Tanzania than in the other comparator countries despite the greater number of inspections and required meetings in Tanzania. To try to estimate the extent of tax evasion, the investment climate surveys asks enterprise managers ‘what percentage of total sales would you estimate the typical firm in your area of activity would report for tax purposes.’ The median estimate in Tanzania was 80 percent of sales. This was lower than in Uganda (90 percent), China (100 percent) or Kenya (100 percent).

Donor Assistance

1.15 Donor assistance to the Government of Tanzania in the form of grants and loans has been on the rise since the mid 1990s. FY03 and FY04 saw significant further inflows of aid, which are now estimated at around 9.7 percent of GDP and finance more than forty percent of Tanzania’s budget. The most significant development has been the increase in general budget support provided by a group of 13 donors which now accounts for more than 50 percent of development assistance received by Tanzania.

Table 3: Composition of External Assistance (% of GDP), FY98-04

FY98 FY00 FY02 FY03 FY04 FY04Actual Actual Actual Actual. Budget Proj.

Grants and Loans 4.0 6.0 5.9 8.6 8.9 9.7Grants 3.0 4.5 4.5 6.5 6.3 6.5

Program 0.7 1.4 2.2 3.1 3.4 3.6Project 2.3 3.0 1.6 2.7 2.1 2.1HIPC debt relief 0.0 0.1 0.7 0.8 0.8 0.8

Loans (net) 1.0 1.5 1.4 2.1 2.6 3.2Loans 2.1 2.8 2.2 3.2 3.9 4.2

Program 1.3 0.9 1.0 1.6 1.8 2.1Project 0.8 1.9 1.2 1.6 2.0 2.1

Amortization -1.1 -1.3 -0.8 -1.1 -1.2 -1.0Source: IMF and Tanzanian authorities

Government Expenditures

1.16 Government expenditure has risen dramatically since the late 1990s. FY03 and FY04 saw further increases in government expenditure, which now stands at an estimated 23.7 percent of GDP. Most of these increases have been devoted to enhancing funding of operations and maintenance, which increased from a meager 4.5 percent of GDP in FY98 to 12.3 percent of GDP in FY04. While in the late 1990s and early 2000s operations and maintenance had been severely under-funded and increasing allocations to this area was the top priority, current expenditure levels require closer attention to the structure and efficiency of public expenditures.

1.17 With respect to the structure of expenditures, strategic decisions are required with respect to the appropriate distribution of funds between wages and salaries, operations and

6

maintenance, and (locally funded) development expenditures. Concerning wages and salaries, current expenditure levels should provide scope for the accelerated implementation of the medium term wage policy. It is also of concern that with respect to the development budget government continues to rely primarily on foreign donor financing, while locally financed development expenditures have increased at a rather slow pace.

1.18 Concerning the efficiency of spending on operations and maintenance, it would be useful to review management practices for key expenditure components such as wage related expenditures, training, government assets (such as buildings and vehicles), or the rental of office space.

Table 4: Government Expenditure (% of GDP), FY98-04

FY98 FY00 FY02 FY03 FY04 FY04

Actual Actual Actual Actual Budget Proj.

Total expenditure and net lending 15.7 18.9 17.6 20.8 22.5 23.7 Recurrent expenditure 11.0 11.8 13.6 15.6 17.0 18.1 Wages and salaries 4.3 4.2 4.0 4.2 4.4 4.4 Interest payments 2.3 1.9 1.4 1.0 1.4 1.4 Domestic 1.0 1.2 0.7 0.6 0.7 0.7 Foreign 1.3 0.7 0.7 0.4 0.7 0.7 Other goods, services & transfers 4.5 5.8 8.2 10.4 11.2 12.3 Clearance of domestic arrears 0.8 0.1 0.7 0.0 0.0 0.0 Bank and parastatal recapitalization 0.0 1.5 0.0 0.0 0.0 0.0 Dev. expenditure and net lending 3.8 5.3 3.4 5.2 5.5 5.5 Domestically financed 0.5 0.3 0.6 1.0 1.3 1.3 Foreign-financed 3.4 5.0 2.8 4.2 4.2 4.2 Net lending 0.0 0.0 0.0 0.0 0.0 0.0 Expenditure float 0.0 0.3 0.0 0.1 0.0 0.0

Source: IMF and Tanzanian authorities

1.19 The wage bill increased to 4.4 percent of the GDP in FY04 reflecting the continued implementation of the medium term pay policy. Average salaries increased by about 25 percent in FY03 and by 18 percent in FY04.

1.20 After declining until FY02, public sector employment has started to increase since FY03. Some of the increases in staffing such as the hiring of additional teachers is directly related to the implementation of the PRS. However, since the implementation of the medium term pay policy is predicated on a tight control of civil service employment, it will be crucial to closely monitor and control employment levels in the public sector. In particular, the widening gap between employment targets under the medium-term pay reform strategy and actual employment numbers is of concern.

7

Table 5: Civil Service Employment FY99-FY04 (December of each year)

Salary Scale FY99 FY00 FY01 FY02 FY03 FY04

Staff in absolute numbers

TGOS 35651 34806 35001 37530 38829 40288

TGS 63787 63310 59994 56838 57095 58651

TGTS 122215 118868 119566 122148 130158 134498

TPSW + TGPSW 36190 34821 36448 39823 38208 38983OTHERS 12785 11381 8837 772 775 58

TOTAL 270628 263186 259846 257111 265065 272478

Target under Medium-term Pay Reform Strategy

258543 249,000 240,943 234,655

Percentage change

TGOS -2.40% 0.60% 7.20% 3.50% 3.76%

TGS -0.70% -5.20% -5.30% 0.50% 2.73%

TGTS -2.70% 0.60% 2.20% 6.60% 3.33%

TPSW + TGPSW -3.80% 4.70% 9.30% -4.10% 2.03%OTHERS -11.00% -22.40% -91.30% 0.40% -92.52%

TOTAL -2.70% -1.30% -1.10% 3.10% 2.80%

Note: Employment numbers are as of December 31 of the fiscal year, e.g. data for FY04 are as of December 2003.Source: CSD

Table 6: Civil Service Average Salaries, FY98-FY04

Salary Scale FY99 FY00 FY01 FY02 FY03 FY04

Average monthly salary (T.Shs.)

TGOS 37703 38212 47220 56633 65265 74116

TGS 57313 57625 78067 85267 100817 127681

TGTS 56314 57073 80162 92670 108844 127959

TPSW + TGPSW 60506 61400 70925 75669 85830 93091OTHERS 11587

8108011 167861 84149 483285 222155

TOTAL 54684 55387 74143 78878 98560 116484

Percentage Change

TGOS 1.4% 23.6% 19.9% 15.4% 13.5%

TGS 0.5% 35.5% 9.2% 18.2% 26.6%

TGTS 1.3% 40.5% 15.6% 17.5% 17.5%

TPSW + TGPSW 1.5% 15.5% 6.7% 13.4% 8.4%OTHERS -6.8% 55.4% -49.9% 474.3% -54.0%

TOTAL 1.3% 33.9% 6.4% 24.9% 18.1%

Source: CSD

8

1.2 STRATEGIC RESOURCE ALLOCATION

1.21 This section reviews developments in the allocation of resources in FY04 as well as actual expenditures during FY03.

Table 7: Composition of Public Expenditures as % of GDP, FY96-FY04

VOTE HOLDER FY96 FY98 FY00 FY02 FY03 FY04Actual Actual Actual Actual Actual Budget

Recurrent Expenditures 12.5 13.1 12.8 14.9 14.9 17.1Debt Service 3.7 4.9 4.2 3.2 2.8 3.9

Supply Votes 8.8 8.3 8.6 11.7 12.2 13.3 Recurrent Central 6.5 6 6.3 8.8 9.0 10.2 Recurrent Regions and

Districts* 2.3 2.3 2.3 2.9 3.2 3.1Development Expenditure 0.5 1.6 1.5 2.7 4.7 7.7Total Expenditure 13 14.7 14.3 17.6 19.6 24.8

* Transfers from the central government to the regions and local authorities.Source: Appropriation Accounts (FY96-FY03), Budget Books for FY04

1.22 The appropriation accounts show an increase in total expenditure (including debt service) from 17.6 percent of GDP in FY02 to 19.6 percent of GDP in FY03.4 Higher development expenditure accounts for this increase in spending, while recurrent expenditure remained at the same level as in FY02. The budget for FY04 projects a further increase in overall spending to 24.8 percent of GDP with development spending increasing from 4.7 to 7.7 percent of GDP and recurrent expenditure increasing from 14.9 to 17.1 percent of GDP. The increase in development expenditure witnessed since FY02 reflects both improved capture of ongoing assistance in the budget but also the scaling up of activities, especially in the infrastructure and energy sectors.

1.23 The share of recurrent expenditures used for debt service payments declined from about 3.2 percent of GDP in FY02 to 2.8 percent in FY03, reflecting debt relief received under the HIPC initiative. The budget for FY04 projects an increase in debt service payments to3.9 percent of GDP, partly reflecting provisions for the servicing of non-Paris Club debt. By February 2004, spending on debt services was significantly less than budgeted and about 25 percent of the original allocation (or about one percent of GDP) has been reallocated to other uses. The increase in recurrent expenditures was fairly evenly divided between spending at the central level and transfers to the local level (regions and districts). Spending at the central level increased by 0.2 percentage points to 9.0 percent of GDP, while transfers to the local level increased by 0.3 percent to 3.2 percent of GDP. The budgeted increase in recurrent expenditure for FY04 is focused on central government expenditures, while allocations to regions and districts decline by 0.1 4 The definition of recurrent expenditure in this section uses the government of Tanzania classification, which includes total debt service payments as part of Consolidated Fund Services, while the classification used in the previous section includes only interest payments but not amortization as part of recurrent expenditures. Expenditures funded from the education and health sector baskets are included under recurrent expenditure in the CGO tables but under development in the appropriation accounts. There is are also significant differences between the estimates of development expenditure in the CGO and development expenditure recorded in the appropriation accounts. While the figure for development expenditure shown in the CGO is an estimate based on commitments, the appropriation accounts only show development assistance which is actually recorded in the government books.

9

percent of GDP. Since transfers to local authorities provide resources for the delivery of decentralized services in the priority sectors covering education, health, water, and agriculture, the fact that these expenditures have increased at a much slower pace than expenditures by the central government is of concern. Starting in FY05, Government has decided to adopt a formula based approach to distributing resources for health and education to the district. An important provision of the introduction of the formula based approach is that no district should receive fewer transfers than under the previous system. In order to support service delivery at the local level, Government is encouraged to give appropriate priority to resource allocations to service delivery units as compared to the funding of central administrative structures.

Table 8: Functional Allocation of Recurrent Expenditure (actuals, % of GDP), FY96-FY04

Sector FY96 FY98 FY00 FY02 FY03 FY04Actual Actual Actual Actual Actual Budget

Administration 2.1 1.6 2.1 3.5 3.0 3.9Defense and Security 2.4 2.2 2.0 2.1 2.3 2.4Social Services 3.5 3.6 3.6 4.7 5.1 5.3Economic Services 0.2 0.4 0.7 1.0 1.2 1.1Productive Services 0.5 0.3 0.3 0.4 0.5 0.6Supply Votes 8.7 8.2 8.6 11.7 12.1 13.2Consolidated Fund Services 3.7 4.9 4.2 3.2 2.8 3.9Total Recurrent Expenditures 12.5 13.1 12.8 14.9 14.9 17.1Source: Appropriation Accounts (FY96-FY03), Budget Books for FY04

1.24 Table 8 presents the functional classification of recurrent expenditures funded from the central government budget for the period FY96-FY04. Expenditures on the social sectors had remained fairly constant at around 3.6 percent of GDP over the period FY96 to FY00. In response to enhanced HIPC debt relief and the introduction of the PRSP supported by budget support from the PRBS/PRSC facility, spending in the social sectors has increased in the following years reaching 5.1 percent of GDP in FY03 and is budgeted to increase further to 5.3 percent of GDP in FY04. It is also important to note that the appropriation accounts capture basket funded expenditures in the health and education sectors in the development budget rather than in the recurrent budget, which is likely to result in a quite substantial underestimation of recurrent spending on the social sectors. The increase in spending on the social sectors was facilitated by the overall increase in recurrent expenditure. As a share of spending on the supply votes, social spending has fluctuated between 40 and 44 percent in recent years. As spending on other sectors is rising faster than in the social sectors, the share of social services is budgeted to fall to 40 percent in FY04 after reaching 42 percent in FY03.

1.25 Expenditures on administration have fallen from 3.5 percent of GDP in FY02 to 3.0 percent in FY03. Spending in FY02 had been exceptionally high as it covered liabilities and the clearance of arrears to suppliers incurred by all ministries. The budgeted increase in spending on administration to 3.9 percent in FY04 is partly due to the introduction of the Treasury voucher scheme, were T.Shs. 31 billion (or 0.3 percent of GDP) have been set aside for tax expenditures. Administration also includes the vote for the President’s Office – Regional Administration and Local Authorities which in FY02 and FY03 contains large amounts (T.Shs. 14 and 20 billion, respectively) for transfers to the local authorities for the funding of capitation grants in primary education. In

10

addition, administration also includes expenditures for priority areas related to accountability in the public sector such as the OCAG, the Judiciary, the Civil Service Reform Department, and the Ministry of Lands.

1.26 Expenditures on defense and security have declined from 2.4 percent of GDP in FY96 to 2.0 percent in FY00. Since then, expenditures on defense and security have been on a steady increase and are projected to reach again 2.4 percent of GDP in FY04. Government justifies these increases in security related expenditures as a result of insecurity in the region and rising crime in Tanzania. As such, enhanced security is considered to be a pre-condition for sustaining economic growth and reducing poverty.

1.27 Expenditures on economic services, which include works, energy and mining, lands, housing, and urban development, and communications and transport, increased from 1.0 percent of GDP in FY02 to 1.2 percent in FY03, attributable to increased spending by the Ministry of Transport and communications of about T. Shs. 25.8 billion.

1.28 Expenditure on the productive services, which include agriculture and food security, cooperatives and marketing, industry and trade, and tourism, natural resources, and environment, has been increasing steadily during the recent years from 0.3 percent in FY00 to a budgeted 0.6 percent in FY04. Agriculture and food security as well as tourism, natural resources, and environment were the primary beneficiaries of these increases.

Table 9: Social Sector Recurrent Expenditures (actuals, % of GDP), FY96-FY04


Education 0.3 0.4 0.3 0.4 0.4 0.4Health 0.3 0.5 0.5 0.6 0.6 0.8Water 0 0 0 0.1 0.2 0.1Science, Technology & Higher Education 0.5 0.4 0.4 0.6 0.7 0.7Regions 2.3 2.3 2.3 2.9 3.2 3.1Total Social Services 3.5 3.6 3.6 4.7 5.1 5.3

Source: Appropriation Accounts (FY96-FY03), Budget Books for FY04

1.29 Recurrent expenditures in the social sectors have increased from 4.7 percent of GDP in FY02 to 5.1 percent of GDP in FY03. The budget for FY04 provides for a further increase in social spending to 5.3 percent of GDP. Most of the increase in social spending occurs at the regional and district level, where the responsibility for the delivery of basic services such as primary education and health care lies. At the central level, funding for tertiary education going through the Ministry of Science, Technology and Higher Education has been the primary beneficiary of spending increases.

11

Table 10: Sectoral Development Expenditures (actuals, % of GDP), FY96-FY04


Administration 0.1 0.4 0.3 0.9 1.5 2.6Defense and Security 0.0 0.0 0.0 0.0 0.1 0.1Social Services 0.2 0.4 0.5 0.9 1.7 2.0Economic Services 0.1 0.6 0.6 0.7 1.2 2.4Productive Services 0.1 0.2 0.2 0.2 0.2 0.6Total Development Expenditures 0.5 1.6 1.5 2.7 4.7 7.7

Source: Appropriation Accounts (FY96-FY03), Budget Books for FY04

1.30 Recorded development expenditures show a continuous upward trend increasing from 0.5 percent of GDP in FY96 to 4.7 percent of GDP in FY02. This increase in development expenditures is partly due to a greater share of donor-funded expenditures being captured in the appropriation accounts, but also to improvements in project implementation performance. The budget for FY04 projects a further significant jump in development spending to 7.7 percent of GDP. The increase in FY03 is primarily due to increased spending under the sector development programs for health and education (an important share of resources provided for health and education is shown under administration, as they are included in the budget of the President’s Office – Regional Administration and Local Authorities). Important contributors to the increase in FY04 are higher spending on infrastructure, especially roads and power and further increases in spending under the sector development programs for health and education captured under administration – PO-RALG.

12

Box 1: Supplementary Budget and Additional Expenditure Needs in FY04In February 2004, Government identified a range of emerging expenditure needs (8 items, amounting in total to TZS 157,260 mill., which were not (or not adequately) provided for in the approved 2003/04 budget estimates. These additional expenditures will be financed partly through draw down of reserves and partly through reallocation.Additional Expenditure Needs identified during FY04

Estimates in TZS bill Sub-vote / Sub-item Originally Approved Estimates

Additional Expenditure Needs

Vote 50 MOF Contingency(a) parastatal wages (and related pension contributions) 3,617 2,600(b) anti-corruption campaign 857(c) local government subventions 4,000 18,900(d) retrenchment 4,000(e) government wages (and related pension contributions) 34,516 12,990(f) contingent proper 8,686 12,200Sub-total 2001/261132 55,676

Vote 43 MAFSstrategic grain reserve, grain 5002/320103 2,210 15,000

Vote 58 MoEMEnergy & Petroleum, misc. grants & subsidies 3001/280714 20,087 41,300

Vote 61 Electoral Commissionvoters' register not specified not specified 14,270

Vote 62 MoTCTanzania Aircraft Auth. Transfers & subsidy 2001/280571 4,868 40,000

Total 157,260

Subsequently, in February 2004, a supplementary estimate of expenditure (No.1 2003/04) was approved by Parliament, as follows:

The supplementary budget is for an increase in budgetary expenditure of TZS 87.25 bill; Only 3 items of additional expenditure are mentioned, including strategic grain reserve, TANESCO subsidy and Government aircraft (i.e. 5

items listed in the Note are not included). Financing of the additional expenditure is composed of (i) additional domestic revenue: TZS 7.58 bill (ii) Budget support from PRBS 5 and

ADB SAL6: TZS 39.67 bill and (iii) draw down of government reserves with BoT: TZS 40.0 bill.

This leaves TZS 60.0 bill of the additional expenditure needs to be funded from reallocations within the original budget.

TZS 53.5 billion is being reallocated from amortization of public debt to recurrent expenditure. Secondly, an amount of TZS 36.0 billion has been mobilized as savings from most MDA votes. Combined with the increased resource envelope of the Supplementary Budget, this allows new allocations of TZS 176.7 billion. This amount exceeds the originally identified additional expenditure needs by TZS 19.4 billion7. Of this amount, TZS 3.5 billion will be allocated to MAFS in addition to the TZS 15.0 for the SGR. No information has yet been received on the use of the remaining TZS 15.9.

These budget enhancements and reallocations are quite substantial, representing 12% of total discretionary recurrent expenditure as originally approved. The budget reallocations are primarily targeted at Other Charges, where they constitute 18% of the originally approved estimates, and only about one third of the value of those amendments concern the drought related expenditure for the Strategic Grain Reserve and TANESCO.

The impact of these reallocations on the PRS priority sectors’ share of total discretionary recurrent expenditure is a decline from 46.4% (based on the originally approved budget including planned PE adjustments) to 43.1%. If the emergency related expenditures to SGR and TANESCO are excluded from the calculation, the PRS priority sectors’ share becomes 43.7%.

MoF has indicated that the cuts in allocations for the PRS priority sectors will affect only lower-priority items such as workshops, seminars and overseas travel. The question is whether this is realistic in view of the amounts involved (e.g. TZS 3.6 billion for Ministry of health) and the in-year budgetary adjustment where some of the original budget allocations may already have been spent. It is recommended that implementation of these expenditure cuts in the PRS priority sectors should be closely monitored.

The substantial additional expenditure needs related to the drought (at least TZS 56 bill8) compared to the ‘real’ contingency of only TZS 8.7 bill, raises the issue of how the government should plan to finance substantial emergency requirements as a result of external shocks. It is suggested that this could be done from domestic borrowing as long as the annually programmed level of such borrowing is kept at a very modest level, e.g. at the 0.5% of GDP which has been the standard in the past. The agreement with the IMF under the PRGF reviews during FY04 has led to an increase in the domestic financing of about TZS 40 billion, which is not far short of the emergency requirements. However, the budget frame for FY05 to FY07 count on steady increases in domestic borrowing. If that approach is taken, the scope for funding budget implications of external shocks from additional domestic borrowing without negative impact on macro-stability will be constrained.

5 Impact of actual average exchange rate being different from budgeted average exchange rate. 6 Disbursement of one tranche of SAL in FY04, which was originally expected in FY03.7 Total funds mobilized TZS 176.7 billion less additional expenditure needs of TZS 157.3 billion.8 SGR TZS 15 billion, TANESCO TZS 41 billion).

13

Use of the Contingency under Vote 50

1.31 In the budget for FY03, the Ministry of Finance retained T.Shs. 75 billion under vote 50 for reallocation to spending units during budget implementation. The bulk of this amount (T.Shs. 41 billion) was retained for salary adjustments). Other indicative uses include parastatal wage increases, debt swaps, the voucher system for non-religious NGOs, the anti-corruption campaign, and a HIPC relief adjustment. Table 11 shows the indicative and actual use of this retained amount [Analysis to be completed based on reallocation warrants].

Table 11: Utilization of the Contingency Under Vote 50, FY03

Intended Use T.Shs. Billion Actual Use T.Shs. BillionParastatal wage increase 6.0Contingency proper 8.0Debt Swaps 4.0Voucher system for non religious NGOs 13.2Salary Adjustment 41.3Anti-Corruption Campaign 0.8HIPC Relief Adjustment 1.7Total 75.0

1.32 For FY04, the Contingency voted under Ministry of Finance amounts to TZS 55.7 billions. Most of the contingency (TZS 47.0 bill for FY04) is in fact provisionally allocated for ‘Special Expenditure’ during the budget formulation process, but allocated to the executing MDAs into the budget year when details for final estimates have been obtained (Table 12).

1.33 Although the quarterly budget execution reports (BER) do not give comprehensive data for the use of the Contingency during quarters 1 & 2, the combined information from the BER Q2 and the Reallocation Warrant no. 1 for 2002/03 allows an (not necessarily complete) assessment to be made.

Table 12: Utilization of the Contingency Under Vote 50, July-December 2003

TZS billion Planned as per Budget Speech June 2003

Reallocations July-December 2003 (*)

Special Expenditure 95.380Parastatal Wages 3.617 3.600Anti Corruption Campaign 0.857 0.857Local Government subvention 4.000 4.000Joint Finance Commission 0.700 -Retrenchment 4.000 -Government wages 34.516 34.500Contingent Proper 8.686Allowances to MPs 3.924Election of Ward Officers 0.778Transport of food to deficit areas 0.693Pension contributions 0.500Student allowances, higher education 0.425Miscellaneous allocations 0.405Total 56.376 49.682 (*) Assessment based on Reallocation Warrant no. 1 and PE budget 2003/04 as per Budget Guidelines tables (January 2004) plus the (confirmed) assumption that special expenditure allocations for wages has indeed been used for that purpose, although this is not included in Reallocation Warrant no.1.

14

1.34 Among the Special Expenditure, the allocation for Retrenchment and the Joint Finance Commission (between Zanzibar and the Mainland) has not yet been utilized. According to the BER for Q2, the retrenchment expenditure for Q1&2 is nil. The commentary does not offer any explanation why no progress has been made in this area. With no reallocation towards the JFC it must also be assumed that little if any progress has taken place in the work of the JFC.

1.35 Some 77% of the Contingency Proper has been allocated for specific purposes. Of those reallocations about half went to allowances for MPs. The use of the Contingency proper for drought related expenditure has been very modest (10% of the allocations made.

1.36 It is reported that TZS 10.5 bill was spent on the Population Census in Q1, but not clear from what vote/sub-vote funds were allocated for that purpose. There is only a reference in a footnote to item reclassification.

1.3 PRIORITY SECTOR EXPENDITURES

Priority sector trends 1999/2000 – 2003/04

1.37 Tanzania’s PRSP identifies several sectors as priority sectors that would be the focus of budgetary allocations for poverty reduction. These sectors include (i) education (notably at primary school level), (ii) health (primary health care), (iii) agriculture (research and extension), (iv) roads (in the rural areas), (v) water, (vi) judiciary, and (vii) HIV/AIDS. Subsequently, lands has been added as another priority sector given the critical role of land for the agriculture sector and for income generation. Since 1999/2000 and in the budget for 2003/04 (see Table 13), expenditures in these sectors have seen significant absolute increases.

Table 13: Trend on Priority Sectors (RE + DEV) in Absolute and Relative Terms Over the Years

TZS bl and % 99/00 00/01 01/02 02/03 ¾Priority sectors (RE + DEV) (excl. Lands)

414.4 494.6 753.7 953.6 (+200 bn or 26.5%)

1,122(+ 168 bn or 18 %)

Priority sectors, share of total discretionary budget (excl CFS*)

44.8% 51.8% 58.1% 53.3% 52.1%

Priority sectors, share of total budget (incl. CFS)

34.7% 39.3% 46.3% 45.6% 43.03 %

Non-priority sectors (RE + DEV – including Lands), share of total budget (incl. CFS)

44% 37% 34% 40%(+ 291.5 bl or +53.7%)

41%(+ 242.8 bn or + 29%)

(Source: Ministry of Finance, Central Government Expenditure, 1998/99 – 2003/04) * Total discretionary budget in FY 2002/03 (in bn TZS): (total expenditure: 2,091.1) – (Consolidated Fund Services: 303.7) = 1,787.4.

1.38 Until FY02, the share of priority sector expenditures in both total discretionary expenditure (from 44.8 percent in FY00 to 58.1 percent in FY02) and total expenditure (from 34.7 percent in FY00 to 46.3 percent in FY02) has been on the increase. However, subsequently the expenditure shares of the priority sectors declined to 52.1 percent and

15

43.0 percent respectively. There are two complementary reasons for these recent developments. Firstly, government claims rightly that the priority sector defined in the PRSP are primarily focused on the social sectors and give insufficient weight to public expenditures that are supportive of economic growth such as infrastructure expenditures.

1.39 The Songo Songo gas development project was a large part of foreign development expenditure and T.Shs 67 billion was recorded as total project loans. So while there was a 38 percent year on year rise in foreign development expenditure, 17 percent of total foreign development expenditure (T.Shs 405.2 billion) in 2002/03 was Songo Songo project loans. This represents 3 percent of total expenditure (including CFS) in 2002/03. Since this comes under the energy sector, and hence is not captured in priority sector outturn, the relative share of priorities in total expenditure is lower in 2002/03. The exclusion of the Songo Songo project from the outturn of 2002/03 expenditure reveals that the priority sector share of total expenditure including CFS is 46 percent, hence there would be no relative decline compared to 2001/02.

1.40 Secondly, after the rapid increase in spending on the priority sectors until FY02, it was considered necessary to enhance funding for other sectors, which, even though their activities are not directly linked to poverty reduction, are nonetheless critical to a functioning government. The new PRSP, which is currently under preparation, is expected to revisit the definition of the priority sectors and to move from a rather mechanic link between defined priority sectors and budgetary allocation to a system that pays greater attention to poverty reduction results and outcomes.

1.41 The primary causes for the decline in the share of priority sector expenditures in the total budget (incl. CFS) between 2002/03 and 2003/04 from 45.6 % to 43.03 % are the following:

Wage reallocation. Priority sector expenditure in the approved budget does not include the amounts of T.Shs 34 billion for civil servant wages and T.Shs 3.6 billion for parastatal wages, which are currently under the Ministry of Finance vote 50 contingency. These will be reallocated across Government during the course of the fiscal year. It is estimated that the reallocations to priority sectors will amount to T.Shs 24.95 billion. This represents 1 percent of total expenditure (incl. CFS).

Foreign development reallocation. The approved budget overestimated the amount of foreign development expenditure to the energy sector by T.Shs 40 billion. This will be reallocated to the health sector (T.Shs 17 billion) and roads sector (T.Shs 23 billion). This represents 1.5 percent of total expenditure (incl. CFS).

Contingency. A contingent amount of T.Shs 8.7 billion has been kept aside under Ministry of Finance, vote 50 for unforeseen circumstances eg. food procurement in the event that the drought will affect food supplies in the country. When reallocated, it will raise the agriculture sector estimate. This represents 0.33 percent of total expenditure (incl. CFS).

1.42 In total, these items which will increase priority sector estimates relative to non-priorities, amount to 2.8% of total expenditure budget (incl. CFS), which when compared to the relative decline of 2.6% indicated in Table 13, it shows that the Government has

16

not placed a declining importance to priority sectors. In addition, there is an 18 percent rise in absolute terms in approved budget priority sector allocations between the two years. The increase for PRS priority sectors between 2002/03 and 2003/04 is somehow muted to accommodate the need for increased allocations to security agencies such as the police due to the rising crime experienced in the country.

1.43 There has been a rise in allocations to regions and local governments in the priority sectors from T.Shs 327.3 billion in 2002/03 to T.Shs 448.6 billion (37 percent increase) in 2003/04. The policy strategy is to fund priority programs, and to devolve expenditure as much as possible from MDAs and sector programs to local government authority and district-based programs. This shift in allocations is progressing.

Actual expenditures on the priority sectors during the first half of FY04

1.44 Starting in FY04, PRS priority sector expenditure is included in the budget execution report (BER) with full details according to GFS codes (in annexes E, F, G and H). This change of format greatly improves the usefulness of the BER as a monitoring tool in relation to PRS implementation for the development partners and the Tanzanian civil society at large (at least as far as the direction of the original PRSP was concerned).

1.45 Nonetheless, the assessment of expenditure performance is made difficult for two reasons:

Compatibility problems between the BER format and the budget estimates books. The presentation of recurrent expenditure break-down in the BER (Annexes A and C) does not match well with the structure of the budget estimates books, except for the priority sector expenditure (this is discussed in more detail in section 2.4 below); and

Assessment of performance on the development budget during the year before the annual appropriation accounts is not very meaningful as most donor-funded expenditure is typically only reported in the form of dummy vouchers at the end of the fiscal year. The performance of development expenditure during the first half of FY04 is thus not reviewed.

1.46 Salaries and Wages. The implementation of personnel expenditure is very much on track at 99% of estimates for the six months. It is worth noting that personnel expenditure for PRS priority sectors is ahead of estimates, mainly because PE releases to LGAs for basic education (i.e. primary school teachers) are 11% above estimates. This indicates strong commitment to implementing the programmed increase in the teaching staff under PEDP. The commentary of the BER needs to elaborate further on the developments in salaries and wages e.g. to which extent the year-on-year or quarter-to-quarter changes are caused by changes in remuneration rates or in staff numbers. This is particularly important for the education (priority) sector in order to follow the provision of primary school teachers, which constitute some 40% of government employees, a third of the consolidated government payroll and more than half of all subventions to LGAs.

1.47 Other Charges for Priority Sectors. The overall execution of OC expenditure in the priority sectors has made substantial progress compared to the rate of progress made for the corresponding period in FY03. Expenditure in Q2 was almost on target (at 99% of

17

half-year estimates, after 75% achievement in Q1). As in the case of salaries and wages, PRS priority sectors are more than fully funded in accordance with estimates for the six months and have received about 55% of the annual estimates. It is again the PEDP related subventions to LGAs that contribute to this picture. This is a positive development that indicates that procurement planning and implementation is improving, but perhaps also that the aggregate increase in budget allocations for OC in the priority sectors, compared to the previous year, has been less than in Other Sectors. However, the additional resources transferred to LGAs do not necessarily correspond to actual expenditure at LGA level (no consolidated LGA reports are available) and the shifting of SWAP basket funding and associated expenditure from the recurrent to the development budget classification further blurs the picture (and illustrates why a unification of the budget would be beneficial to budget formulation and execution monitoring).

1.48 Substantial differences in sector performance are noted among priority sectors and items, around the average expenditure of 47.5% of annual estimates. Within the priority items, education and water have achieved 59% and 57% respectively of the annual estimates, while HIV/AIDS is at the bottom with 26%. All other sectors have achieved 43-47%.

1.49 The BER mentions that additional funding of emergency expenditures related to drought is required and that some such allocations already have been made. This includes the Strategic Grain Reserve replenishment and procurement of fertilizers. The SGR actual expenditure (a priority item) for the six months is about TZS 0.9 bill above the budgeted six months estimate, while for the entire MAFS (vote 43), including non-priority items, the expenditure is about TZS 5.0 bill above six months estimate. This suggests that fertilizer procurement was been given top priority during those six months.

1.50 It is indicated that funding of the emergency needs has been made available by a (TZS 14 bill) reallocation from local development expenditure to recurrent Other Charges. About TZS 6 billion of this originates from planned allocations to development projects in the priority sectors. Whether there is an aggregate reallocation away from or into the priority sectors during Q1&2 is not possible to establish in detail, but the figures mentioned suggest that this reallocation has been fairly neutral for the priority sectors in general (though probably not for priority items).

1.51 Sectors other than PRS priorities. Overall, the expenditure performance was 122% of the half-year estimates. No breakdown is provided in the BER on the execution in these sectors, apart from an aggregate for all such votes. Earlier BERs included such a breakdown, which is important for monitoring when developments outside the priority sectors threaten to impact the execution of priority sector expenditure. This may be the case this year due to the large (more than 200%9) increase in subsidies to TANESCO as part of the emerging expenditure needs based on below average rainfall in 2003.10 The BER for Q2 highlights the changing expenditure needs, but is unspecific as to the

9 According to the MoF Brief Note for PRBS Partners, January 2004.

10 The purchase of a new government aircraft will be financed from additional resources through a supplementary budget and should therefore not affect priority sector expenditure in the current year, though the issue of amortization and interest charges for the additional debt in future years remain an issue.

18

financing of this expenditure, and particularly the potential for reallocations from priority expenditure votes and items (see further on budget revision in section 2.3 below).

1.52 The BER Q2 mentions that some of the new expenditure needs have already been partly catered for during the first six months of the year, but lack of details on ‘non-priority’ expenditure lines and the MoF contingency in text and tables (with the exception of the agricultural sector items) leaves an incomplete picture.

2. CONSISTENCY IN BUDGET PLANNING AND EXECUTION

1.53 A key objective of this year’s external evaluation is to assess the consistency of policy, planning, and budgeting in Tanzania. In particular, the mission focused on the consistency between the PRS, budget guidelines, the MTEF, the budget, and actual expenditure. Preliminary findings of this analysis are as follows:

1.54 The PRS (2000) contains explicit statements on areas of priority spending and expected expenditure trends. The FY03 PER carried out a detailed assessment of the consistency between the PRS and budgetary developments and found that spending trends are broadly consistent with the PRS. Subsequently, analysis carried out in the context of the PRBS/PRSC Annual Review raised concerns about Government’s continued commitment and the implementation of the PRS, as spending estimates for the priority sectors and areas in the FY04 budget suggested a stagnation or even decline of the share of spending going to these areas. Similarly, projections for FY05 and FY06 show no increase in the share of spending going to the priority sectors. Concerns about government’s commitment to the PRS have also been fueled by discussions on actual or considered large expenditure commitments that seem to lie clearly outside the PRS, such as the purchase of government aircraft, or the construction of a new Parliament building and a new Sports Stadium (which is expected to be financed through a grant from China).

1.55 A key element in the discussion of the consistency between the PRS and budgetary developments is the definition of priority sectors and items, which was provided in the PRS and subsequently refined. The usefulness of assessing current budgetary developments in relation to the priority sectors and items identified in the PRS has come under attack from two sides. Firstly, there is an increasing realization that specific areas of expenditures, which are not included in the PRS priority sector definitions, especially those related to infrastructure, are critical to growth and poverty reduction. Secondly, there is also the argument that rather than focusing on expenditure shares the focus needs to be much more on results. Indeed, a overly strong focus on expenditure shares may indeed undermine the budget process and incentives for effective and efficient service delivery. While such a results focus is clearly desirable, it is also recognized that at present a clear statement by government (in the form of a credible MTEF) on how the PRS will be translated into public expenditures would still be the most meaningful benchmark against which domestic stakeholders and donors can assess the consistency of public resource use with the PRS. This should thus be a priority in the preparation of the new PRS currently underway.

1.56 The review of the consistency between expenditure estimates in the budget guidelines and the budget indicates that at the aggregate level, consistency has increased until very recently. However, more severe consistency issues arise with respect to

19

expenditure allocations within sectors and votes. The analysis focused first on the consistency of various statements on revenue and expenditure for a given budget year and subsequently focuses on the consistency and quality of outer year projections.

1.57 Resource estimates contained in the budget guidelines typically underestimate resources actually allocated in the budget, however, the difference has become smaller in recent years declining from 12 percent in FY02 to 6 percent in FY06. An underestimate of around 6 percent is most likely indicative of conservative budgeting rather than severe problems in preparing accurate resource projections at the time of the budget guidelines.

1.58 The main contributing factor is an underestimation of the amount of forthcoming program support in the budget guidelines. However, while in FY00 the underestimation of program resources was about 60 percent, it has decline to the range of 10 to 20 percent in recent years, which is attributable to the combined efforts of government and budget support donors to enhance the predictability of support. Underestimation of domestic revenue has been in the range of five percent in recent years.

1.59 On the expenditure side, the underestimates of resource availability are primarily reflected in under-estimates of allocations to operations and maintenance (Other Charges). In FY00, OC estimates in the budget were about 20 percent higher than the estimate in the budget guidelines. In FY04, the difference was less than 10 percent.

1.60 The consistency between budget execution and the budget suffers from two weaknesses in the budget process. Firstly, while estimates of domestic revenue are fairly accurate, the are still significant differences between the budgeted and actual amounts of budget support. In FY02 actual budget support was 20 percent below the budgeted amount while in FY03 it was 40 percent above the budgeted amount. For FY04 it is expected that the difference is less due to the development of the PRBS/PRSC modality. Another source for budget deviations is the practice of retaining a fairly large amount in the budget of the Ministry of Finance for reallocations during the budget year. However, in recent years this source of uncertainty has diminished and retained amounts are now primarily for salary adjustments and true contingencies. On the positive side, government has consistently protected expenditure allocations to the priority sectors from cuts during budget execution. The possibility that this protection of priority sector expenditures might not be maintained in FY04 is thus of particular concern.

1.61 Projections of resources and expenditures for the outer years have tended to be relatively conservative and discussions during the mission suggested that the preparation of the outer year estimates is generally seen as a bureaucratic rather than a strategic exercise. Given the continuously expanding resource envelope in recent years, overall resource allocations as well as allocations to the priority sectors typically exceeded significantly the amounts that were indicated in the MTEFs one or two years before.

1.62 In looking at the degree of consistency between approved budgets and actual expenditure, it is important to note that the level deviation observed varies greatly with the level of aggregation of the analysis. Many of the sharpest concerns on effective implementation of the budget concern expenditure at the level of individual programs rather than broad sector aggregates. Performance on implementation of individual programs can be examined, but it is valuable to draw these results into a more general

20

picture of how well the budget predicts actual expenditure at this level of detail. What is needed is analysis which draws on a lower level of aggregation and yet provides a picture of what is happening in terms of overall budget implementation, rather than a commentary on individual programs or sub votes.

1.63 For a preliminary analysis of budget execution at lower levels of aggregation, it seemed most useful to focus on expenditure data by sub vote. Data are available for three years from FY01 to FY03. An index of budget deviation was calculated as the sum of absolute differences between approved budget and actual expenditure at the sub vote level, expressed as a percentage of the total budget. For each subvote the budget figure under reference is from the budget, book as approved by parliament, taking no account of any reallocations subsequent to parliamentary approval. The table below shows preliminary summary results for these years, for central government recurrent expenditure.

Table 14: Budget Deviation at Sub-Vote Level (average as % original budget)

Percentage deviationFY01 22.6FY02 26.3FY03 12.3

1.64 Some observations can be made on these preliminary results. First, it is important to consider the implications of the degree of deviation observed. Deviation at subvote level of the order reported is clearly a significant problem for effective management of resources. Given that the level of analysis is still fairly aggregate, it must be assumed that deviations at the level of individual spending units are even higher, implying a serious problem for spending managers in planning activities. Variances can arise from a wide range of possible sources and it should not be assumed that deviations should optimally be reduced to zero. Nevertheless, deviations of the order observed imply reallocation at a level, which must be expected to have adverse effects on the efficiency of expenditure.

1.65 Secondly, there appears to be a significant improvement in performance during FY03. Tracing through the potential causes of the pattern in observed deviation will require further analysis, but there seem to be good grounds for associating the improvement with a tightening of the approval process for reallocations within year, starting FY03. Performance may also have been assisted by the reform of budget classification in the previous year, allowing better alignment between expenditure classification and the organizational structure.

3. STRENGTHENING BUDGET PROCESS

1.66 One of the central concerns of the FY04 PER is the need to enhance consistency within the budget process, between the broad policy statements such as the PRS and Government’s translation of these commitments into expenditure plans, annual budgets and actual expenditure. A key element of this process of translation must be a budget strategy, which should have the following characteristics:

Clear linkage to policy: Sufficiently detailed to capture the key elements of Government’s principal policy commitments in the PRS and other major policy

21

statements (including core elements of MDA Strategic Plans and performance agreements). This implies a multi year horizon and allocations disaggregated sufficiently to establish the link to program objectives and targets, and related monitoring systems.

Consistent with aggregate resource framework Rolled over annually and adjusted to reflect changes in the aggregate resource framework over the planning horizon.

Prepared and approved with the full engagement of policy makers: Approved by Cabinet, and prepared according to a timetable allowing policy makers to establish consensus that the strategy adequately translates existing policy commitments within the constraints imposed by aggregate resource availability.

Binding for the annual budget Decisions taken by Cabinet on strategic allocations are binding for the subsequent stage of detailed budget preparation.

Integrated within a realistic timetable for the policy/budget cycle Prepared sufficiently early in the budget cycle to allow for subsequent preparation of detailed annual budgets.

1.67 Among the various planning instruments prepared in the Tanzanian policy and budget cycle, none exactly fit this description, although the Budget Guidelines comes closest. In principle, the Budget Guidelines meets criteria 2-5 above. However, it does not in present form adequately meet the first criterion, and in practice there are some significant issues with the other criteria, as discussed further below. Nevertheless, as a development of the existing system, strengthening the Budget Guidelines would probably be the easiest way for Government to build a stronger formulation of budget strategy.

Strengthening the Budget Guidelines

1.68 Suggestions on key areas for strengthening the Budget Guidelines can be grouped under the headings of the five criteria above.

1.69 Linkage to policy Currently the Budget Guidelines does not provide a clear translation of policy commitments in the PRS and other statements of Government policy. The degree of disaggregation in the Guidelines needs to be sufficient to capture the critical policy commitments in the PRS and other policy documents, and to make it possible over time to link budget allocations with agreed program targets and related monitoring systems. This implies bringing forward some of the sector analysis currently carried out under the MTEF and incorporating it within the Budget Guidelines – this will also require timely completion of the sector PER work. The Guidelines also needs to have a more credible framework for the medium term, though not at the same level of detail, allowing a realistic expectation that year 2 numbers will be the starting point for budget preparation in the next cycle and a useful focus for consultation in the current cycle.

1.70 Aggregate resource framework: Adequate forecasting of external financing is one of the key problems in this area. This year’s Guidelines has shown some deterioration in this respect, in that the estimates of future donor support are based on information received on confirmed financing, rather than as in the past, a central estimate of likely aggregate financing. The Macro working group was in the past instrumental in assisting to compile such estimates and it may be that revival of the functioning of this group could

22

help to address this problem in future. As noted further below in comments on the donor interface with budget, Government could respond proactively to incomplete forward information on donor financing with its own plan for external financing, as a routine annex to the annual Budget Guidelines, for consolidation in subsequent annual consultations.

1.71 Cabinet approval process: While the Budget Guidelines is submitted to Cabinet, the document is frequently issued before Cabinet have discussed and made revisions to the document. It may be useful in future to allow space in the timetable to complete revision of the document prior to the next stage of budget preparation, making a clear distinction between the Guidelines as draft proposal and the final approved version, with the latter providing the basis for both external consultation and the parallel process of detailed budget preparation. (As a matter of terminology it may also be helpful to distinguish between the budget framework itself – which is in practice the core of the MTEF – and specific directions for MDA submissions for the detailed annual budget, as budget guidelines).

1.72 For the Budget Guidelines to play the role in budget strategy envisaged above, it is very important that Cabinet review of the Guidelines is adequately informed and that policy makers are deeply engaged in the lead up to the Cabinet review of the Guidelines. This reinforces the point above on greater clarity on policy linkages and also underlines the importance of redesigning the budget cycle to ensure that sector inputs are used effectively in Guidelines preparation.

1.73 Link to subsequent budget preparation: In the current year’s Guidelines, there are several factors, which imply that key framework numbers will require modification prior to finalizing the annual budget. There is a significant volume of unallocated expenditure (in part covering potential wage increases, which might more appropriately be decided directly in the context of overall budget decisions). The majority of PE and OC allocations have been derived by uniform changes based on the previous year’s allocations, and it seems implausible that these allocations will not need to be adjusted if the final budget is responsive to policy priorities and varying scope for efficiency gains. Estimates of project financing show major increases which may not be realized: given that this financing is included within sector envelopes, there must be some uncertainty on the impact of adjustments in this area on sector financing.

1.74 Budget timetable: Key inputs at sector level, particularly the sector PERs and sector reviews, have mostly not been completed in time to inform preparation of the Guidelines. This has obviously reduced the potential effectiveness of the exercise. Getting these elements of the budget cycle better aligned is an important component of strengthening the Guidelines. However, it should also be recognized that several sector PERs are now recurrent annual exercises and past work on the sector can still provide useful material for Guidelines preparation.

1.75 If the Budget Guidelines are strengthened by bringing forward some work now covered in the later MTEF preparation, (as suggested above and discussed further below) the implication for the budget timetable may be that the remaining work within the cycle is restricted more to fairly mechanical preparation of the detailed annual budget. This

23

might open the possibility of the Guidelines being finalized a little later, giving longer lead times for preparation of sector inputs to the Guidelines.

Implications for other planning instruments

1.76 MTEF. Strengthening the Budget Guidelines along the lines suggested above amounts in practice to bringing forward the more strategic aspects of the work on the MTEF, which is currently carried out in the later stages of the budget cycle. Terminology is perhaps a little confusing here: in principle the Budget Guidelines is already the core of an MTEF and the strengthening proposed above would do more to translate that into practice. With this development, there could also be savings in the work done in the later stages MTEF preparation. The present level of detail in forward budgeting through the later stages of MTEF preparation does not appear to be very useful and does not in practice seem to provide a reliable base for the following year’s budget preparation.

1.77 Medium term plan for Growth and Poverty Reduction: PRS. The Budget Guidelines as a statement of Government budget strategy should be a combined translation of all core Government policy statements into the form of coherent sustainable expenditure plans, reconciled as required with the aggregate resource constraint. It should be clear that the Guidelines provide the final reconciliation prior to detailed budget preparation, even though policy documents may also include a perspective on the link between policies and resource constraints. This is true of both the PRS and recently produced Medium term plan for Growth and Poverty reduction (which covers a wider spectrum of Government policy than the PRS).

1.78 Strategic Plans, annual performance agreements: An increasing number of MDAs have prepared Strategic Plans and annual performance agreements under the ongoing public service reform program. Currently, the linkage between these activities and the budget process has not been developed to ensure that these plans are strongly linked to resource constraints approved without lo to the budget process. A strengthened Budget Guidelines exercise could support a means of reconciling key features of MDA strategic plans and performance agreements with proposed expenditure plans.

Implications for donor consultations on the budget

1.79 A budget strategy document strengthened along the lines described above, could make a very significant contribution to improving the quality of dialogue on public expenditure issues between Government and the donors. Earlier work on harmonization of donor procedures in this area has essentially assumed Government’s core budget process to be unchanged. With the strengthening proposed, there may be much greater scope for rationalization of donor procedures.

1.80 First, the Budget Guidelines could play an effective role as the single focal point of donor dialogue, allowing donors interface with the overall budget process to be streamlined around discussion of one core document. To the extent that the strategy is firmly linked to Government policy commitments and binding for the subsequent annual budget, the present concerns on consistency through the budget process would be alleviated and the focus on a single overall budget strategy would provide the basis for

24

rationalizing donor interface with the budget. Specifically, it may provide the basis for consolidating the dual processes of PRBS budget review and the PER process.

1.81 Second, as the authoritative translation of Government policy into public expenditure plans, the Budget Guidelines would become a key reference point for the donor community in establishing whether subsequent budgets and actual expenditure are indeed in line with Government’s prior commitments. At present, budgets and actual expenditure in individual sectors may be interpreted as inconsistent with PRS commitments, but it is in practice very difficult to establish this in the absence of a single approved expenditure framework for the whole of Government, which credibly represents Government’s three year expenditure plans.

1.82 Third, a more effective Budget Guidelines document could provide the basis for Government to develop a more proactive role with respect to coordination of external financing. The Budget Guidelines could itself include in addition to the budget strategy, an outline plan for external financing in support of the strategy. This could be developed in subsequent cycles to provide guidance to the donor community on the level and mode of financing required to meet the objectives of the strategy. Such a financing plan would provide a very specific focus for consultation and a significant contribution to the development of coordinated assistance strategy.

4. STRENGTHENING A RESULT-BASED APPROACH IN POLICY MAKING AND BUDGETING

Introduction

1.83 Tanzania has several parallel and complementary monitoring systems that, while providing a wealth of information, fail to provide a framework for policy and budgetary choices. The government is taking steps to strengthen the result-based approach for policy-making and budgeting. This includes the shift from a sector-based to an outcome-based PRS. The next PRS would identify the combination of activities, inputs and outputs required to achieve desired objectives and spell out underlying assumptions; and develop a plan to align monitoring and evaluation efforts to the needs of a result-based framework. The shift from sector-based to outcome-based PRS will provide the necessary basis for strengthening the result-based approach. This will need to be accompanied by greater program orientation in the budgeting process to facilitate the translation of the outcome-based PRS into budgetary allocations. In addition, the government is strengthening performance-compatible incentives structures and performance monitoring within the context of the civil service reform. Further efforts will need to be made to link these efforts to more result-oriented budgeting.

Monitoring systems

1.84 Tanzania has a large number of parallel systems to monitor government policies and programs, but their capacity to deliver outputs that are useful for improving budget allocations and policy effectiveness is weak.

1.85 The Poverty Monitoring System (PMS) is aimed at generating usable data over the short term to inform policy-making and budgeting in real time; and assessing the effectiveness of policies and programs on poverty reduction (see box). But while the

25

PMS is relatively effective in generating data to monitor inputs, implementation processes, and outcomes, it is less effective at feeding the budgeting process with usable analysis of which policies and programs are most cost effective in achieving outcomes. Its routine data component, which should provide policy makers with a source of data that integrates financial and sector data, has so far failed to come up with an overall strategy to generate and feed routine data through the planning and budgeting cycles. And while survey data have been addressing reviewing, planning and budgeting needs, they can only do so at a frequency that is not consistent with annual planning and budgeting processes.

1.86 Annual public expenditure reviews by the Ministry of Finance and annual sector public expenditure reviews by line ministries in priority sectors have served several purposes: i) to analyze past expenditures, ii) to assess sector performance in delivering outputs, and iii) estimate spending requirements for the next budget cycle. Sector-based PER, however, seldom provide the Ministry of Finance with a menu of choices to inform budgetary allocations.

1.87 Public Expenditure Tracking Surveys (PETS) are used to establish the relationship between budget allocations and actual expenditures and determine how much of the allocations actually reach the intended beneficiaries. PETS have been undertaken in Tanzania since 1999 in the education and health sectors (three completed and one ongoing). They find that significant amounts of funds are diverted or misappropriated, and that compliance with policies and regulations is weak. They estimate for example that about 57% of “other charges” in education and 88% in health were diverted (1999) and that available receipts underreported “other charges” by councils by 4-36% (2001) (Sundet 2004). The findings prompted the government to advertise in the media central government transfers to the districts. They have not, however, been incorporated into the policy dialogue or impacted significantly on the level of transparency and accountability.

1.88 Service delivery surveys are being used to obtain feedback from users of public services. The exercises are being carried out as part of Public Service Management’s performance agreements with MDAs. Eleven MDAs completed 29 service delivery surveys, which they used as a basis for developing their performance agreements. They also started rolling out their performance monitoring systems (URT 2003a). Several other service delivery assessments are being undertaken by various NGOs to, among others, evaluate progress in PRS implementation, monitor standard of service in the health and education sectors, monitor budget, and evaluate projects financed by the Tanzania Social Action Fund. Because of subject specificity and lack of uniformity in methodological approach, this body of data does not result in clear recommendations of universal validity. The impact of these initiative is yet small. Efforts to standardize methodology to improve comparability, and to compile and disseminate results would go some way to increase these initiatives’ impact on public discourse.

1.89 Finally, the medium term expenditure framework, which provides an estimate of resource allocations in the medium term, also includes quarterly monitoring of budget execution, and progress in delivering outputs. Quarterly monitoring has been used as a mechanism for mid-course resource reallocation.

26

1.90 The outputs of these monitoring initiatives include routine, survey and census data, the Annual Report on Poverty and Human Development, an annual public expenditure review, annual priority sector public expenditure reviews and at least one non-priority sector PER, several expenditure tracking exercises particularly in education and health, several service delivery surveys, a plethora of reports, and analytical studies and policy briefings.

1.91 The capacity of these outputs to affect policy-making and budgeting is weak. The factors that contribute to it include capacity to extract and operationalize recommendations, and the emphasis placed on adding processes and producing reports rather than providing tailored and useful information to the users. The Ministry of Education and Culture, for example, complains that too much attention is placed in having consultants complete annual education sector PER rather than building analytical capacity in the ministry and affording the ministry sufficient time to incorporate results in their sector strategy by lowering the frequency of the PER exercise.

1.92 Monitoring outputs tend to track inputs and processes or outputs or outcomes, seldom drawing the link between them to measure efficiency or attribute causality. Measures of relative efficiency and effectiveness would be most useful for budgeting and policy-making purposes.

Developing a Result-Based Tanzania PRS

1.93 The Poverty Eradication Division (PED) in the Vice President’s Office is in the process of coordinating the preparation of the new Poverty Reduction Strategy (PRS), which will validate poverty reduction objectives and update the strategies for their achievement. This is the policy document that is somewhat removed from the constraints of MDA-specific and priority sector budget politics, and can provide an integrated vision of how cross-sectoral reforms and programs contribute to the achievement of growth and poverty reduction objectives.

1.94 Current discussions in PED are leaning towards modifying the priority sector-based approach of the previous PRS into a priority outcome-based approach for the next PRS. Such a shift would have the PRS focus on achieving priority objectives in the most feasible and cost-efficient manner, while keeping open the possibility that sectors that were not previously identified as priority may be critical to improve the returns of priority sector investments. Instead of assigning responsibility for poverty reduction to priority sectors, it would invite all sectors to strengthen their contribution to poverty reduction.

1.95 Policy discussions also lean toward expanding the understanding of poverty reduction outcomes to include, in addition to the current focus on social services, more economic measures of wellbeing such as pro-poor employment generation and economic growth.

1.96 The inter-sectoral and strategic vision embedded in such PRS would need to be translated into a monitorable result-based framework that specifies the linkages between policies and programs and the stated objectives. A result-based framework would lay out the developmental priors and assumptions of the strategy. Its principal objective would not be to champion specific approaches to development but to validate the effectiveness

27

of those approaches through the use of information and analysis, and modify them over time to improve the overall effectiveness of the strategy.

1.97 The framework would reduce long-term MDG commitments to yearly and medium-term achievable targets, to both measure progress overtime with respect to those long-term commitments and make explicit the level of resources required for their achievement. For this task, PRS preparation would need to work interactively with the sectors to estimate the expected contributions of each sector activities to developmental outcomes and make assumptions on the synergies of combined inter-sectoral activities in the achievement of each specific objective.

1.98 PED is considering the implications of shifting from a sector-based to an outcome-based PRS. These would need to be thought through and incorporated in the strategy. They might include implications for PRS preparation, monitoring and evaluation, and implementation.

1.99 PRS preparation: Preparing a sector-based PRS could be easily accomplished. Sector-based strategies, studies, PER, PETS and service delivery surveys would offer a ready array of inputs into its preparation. Preparing an outcome-based PRS requires: i) clarity on priority outcomes; ii) comparative analysis of results from different sectors in reaching common outcomes; iii) significant cross sectoral collaboration; iv) understanding of developmental model and efforts to isolate the inter-sectoral synergies and linkages as well as identifying the missing links; and v) significant effort in costing targets in a cross-sectoral setting.

1.100 PRS monitoring and Evaluation: In the current monitoring systems, as laid out above, most monitoring is done at the sector level. While sector-based monitoring will continue to be required to measure the efficiency of MDAs in delivering outputs and services, an outcome-based PRS requires a certain capacity to monitor and evaluate the combination of cross-sectoral activities that contribute to a specific outcome. Two specific implications can be derived: i) the need to develop an integrated management information system that combines financial data with core routine data from all sectors for day-to day monitoring; and ii) the need to redirect a medium-term research agenda to evaluate the impact of policies and programs on desired outcomes.

1.101 Routine data: The routine data component of the Poverty Monitoring System has encounter serious implementation difficulties. It has failed to build public sector capacity for the generation of consistent data required for annual monitoring needs of the PRS. Furthermore, the component has not succeeded in developing a consistent strategy for providing the government with functioning and shared management information systems. Week implementation capacity, lack of coordination between different initiatives across sectors and levels of government (IFMIS, PlanRep, LGMD, etc.), and lack of incentives and capacity for collecting quality data have contributed to this failure. There is a clear need to simplify the approach and consolidate efforts with the aim to develop one integrated routine data system for use across sectors and levels of government.

1.102 A consolidated initiative would focus on developing a management information system that would: i) integrate financial as well as sector information for a well defined set of core output and outcome indicators; ii) establish data collection formats and

28

information sharing protocols as well as improved incentives, iii) eliminate duplication in the demands for information, clarify the flows of information from the locality to the central government and from the central to the local governments; and iv) focus on building data collection capacity at the local authorities and data processing capacity at the finance and line ministries level to improve data quality and analysis.

1.103 An integrated management information system would provide PED and MoF with day-to-day cross-sectoral information required to monitor progress of the PRS and the budget. It would significantly diminish the need or the frequency for ad hoc studies, such as sector PER and PETS.

1.104 Evaluate impact: To improve result-orientation, the PRS review process would focus on measuring the effectiveness of selected programs in achieving their targets, instead of attempting to ascertain on an annual basis the effectiveness of the overall government program.

1.105 To do this, the PRS preparation would tailor the PSM medium-term research agenda to the need to build knowledge on the effectiveness of key PRS programs. The agenda would identify the key programs to be evaluated, establish a timeframe for evaluation, coordinate with data producers to match studies data requirements with the timing of survey, and build technical capacity for the analysis. Logical candidates for evaluation would be large programs that absorb significant fiscal resources (e.g. UPE, agricultural extension) as well as smaller programs targeted for future expansion (e.g. microfinance).

1.106 Results from this analysis would form the basis for: i) estimating a program cost-effectiveness in meeting specified targets, and ii) recommending shifts in resource allocation toward highly effective programs and away from programs with low cost effectiveness. As such, the result-based framework would be the living part of the PRS that would provide for an update of yearly targets and programmatic shifts in the strategy.

1.107 PRS Implementation: The greatest difficulty in an outcome-based PRS is in coordinating the implementation of cross-sectoral initiatives to maximize complementarities and synergies. Specifically, the lack of incentives for working in a collaborative manner across MDAs must be recognized. It would be thus necessary to carefully consider options to increase incentives for collaboration.

1.108 No less difficult will be translating cross-sectoral approaches into MDA-specific performance agreements and budgets, and into MTEF/budget guidelines. The programmatic nature of an outcome-based strategy need to be recognized. The sector overhead component in the national budget, for instance, is not reflected in an outcome-based PRS. This means that the costing of targets in an outcome-based framework will be an estimate of the marginal cost of affecting a certain change in an indicator, but will not be reflective of the overall budgetary effort. Strengthening budget classification along programmatic lines will be required to operationalize the translation of the outcome based approach into the planning and budgeting stages. A program-based budget will also facilitate the monitoring of consistency between the budget and PRS policies.

29

Performance-compatible incentive structures

1.109 A growing number of MDAs have prepared strategic plans and performance agreements within the context of the civil service reform. Performance agreements are informed by service delivery surveys that measure levels and changes in satisfaction with services delivery. By 2003, eleven MDAs had completed service delivery surveys (SDS), and prepared annual plans and performance agreements. The quality, usefulness and comparability of these service delivery surveys will need to be reviewed. By 2004, 42 MDAs had prepared performance agreements. The remaining 144 “vote holders” (including local government authorities) are expected to come on performance agreements in the next two years. For now, these performance agreements are linked to human resource management and incentives systems and are not linked to a system to provide budgetary incentives for achieving results.

1.110 Strengthening performance-compatible incentives will require the alignment of performance agreements with budgetary resource envelop constraints; the establishment of performance-compatible incentive systems for budgetary allocation, including greater decentralization of budget execution; and MDA reporting focused on measuring the efficiency of operations (input to output) rather and solely based on tracking use of inputs.

5. PUBLIC EXPENDITURE MANAGEMENT ISSUES AT LOCAL GOVERNMENT LEVEL

1.111 Status of Public Expenditure Management at Local Level: Overall, government has continued to implement comprehensive reforms at local government level with some improvement in budgeting, public financial management and accountability. However, a number of concerns remain.

1.112 PER and budget preparation for local government: So far the PER process for local government covers only PO-RALG. The main constraint to the roll-out has been capacity limits in LGAs, Regional Secretariats and PO-RALG. In this regard, GoT is encouraged to explore the possibility of using the recently introduced zoning system to undertake PER for LGAs by zone. The use of GFS codes in the preparation of LGA budgets is also a welcome development. PO-RALG in collaboration with PO-PSM is also developing a planning and reporting manual (‘Plan-Rep’) to bring together strategic planning and budgeting. Other notable recent developments include the implementation of harmonization of fiscal year of the local authorities with that of the central government from July 2004. The newly approved population-based formula for allocation of grants to local authorities is also due to start from FY04/05 for sectors of education and health, followed by other sectors in the coming years. Although these developments are generally welcomed, it is suggested that capacity for planning and budgeting need to be strengthened for harmonization of the two levels of processes as well as financial management in the local level. In order that this happens, the Budget Guidelines for the central and local levels would need to be unified. The Government is commended for continuing with the practice of putting quarterly allocations to LGAs in the public domain via local news papers, national website and public notice boards. However, there is still room and need to make the information more widely available and accessible at local level so as to further enhance transparency and accountability.

30

1.113 Status of IFMS in LGAs: During 2002 and 2003 the local government reform program focused attention on getting the first 32 LGAs to reach the final stage of IFMS instead of rolling out the IFMS to more LGAs. Currently 28 out of the 32 are fully using IFMS as of January 2004. All the 32 councils have been instructed to stop using the manual system come December 31, 2004. With regard to preparation for eventual roll-out of IFMS to other LGAs, the Accountant General’s Department has begun training of fresh graduates on the IFMS, and sponsored some in the areas of information technology, accounting and materials management in various institutions in and outside the country to those earmarked for posting in LGAs. Increased effort has also been put on strengthening support systems to implement IFMS. A Systems Development Unit for LGAs housed in ACGEN is now in place. 25 specialists are being trained in the EPICOR software and are to be deployed to the zones from July 01, 2004. Other support by ACGEN include, defining a chart of accounts for LGAs, and supervision of 5 LGAs support zones. LGRP is also in the process of identifying an additional 24 LGAs to start implementing IFMS.

1.114 The Government is commended for the effort made so far and is encouraged to work further toward making all LGAs to operate on IFMS in due course so as to improve public financial management at the local government level. This is now being given high priority under the LGRP. Although a significant effort is being made to train local government staff, there is need to improve coordination of the training supported by ACGEN with that done or planned under the LGSP/LGRP. With regard to recruitment of certified accountants and systems development specialists to beef-up necessary competencies of LGA staff, GoT is encouraged to intensify training of LGA staff, undertake staff rationalization and also focus on making local government a good employer, which attracts and retains qualified staff. Provision of the requisite working tools (especially PC facilities) and physical conditions, as electricity is also key. There is also need to carry out systematic training, taking into account both levels of technical staff and geographical equity. Government is also encouraged to contain/limit political pressure to transfer council directors at will.

1.115 Financial reporting: To-date there is no mechanism/procedure for collection, collating and analysis of local government revenue and expenditure to arrive at a general view of government operations. This situation has constrained government ability to monitor/track expenditure in the PRS sectors and ensure that LGAs are fully accountable for spending at the local level. Under the circumstances, it is also not possible to analyze the consistency between policy, budgets and implementation.

1.116 Nevertheless, there exists some fragmented data in various reports - scattered in different MDAs (Councils, RAS, PO-RALG, sector ministries, ACGEN, and Budget Department in MoF). The range, quality and scope of the data/reports also differ quite substantially. A government initiative to obtain and consolidate revenue and expenditure data using a common reporting format has apparently stalled partly because of the excessive level of detail demanded, the shear large number of LGAs, frequency of reporting (monthly, quarterly, semi-annual, and annual), and capacity limits in the RAS offices/PO-RALG. There were also delays in submission of the reports from LGAs. There are two sets of bottlenecks in this regard. One has to do generally with weak capacity (lack of skilled personnel and equipment) in LGAs/RAS/PO-RALG to undertake timely preparation and compilation of data and reports. It was noted that LGAs that are already using IFMS are able to produce better and more timely reports than those that are

31

not on IFMS. The other problem seems to be the lack of a central focal point in MoF where all reports from LGAs are submitted. Currently reports from LGAs go unilaterally to the PS, ACGEN, CPAD, and Budget department in MoF and even to sector ministries and PMO. Each council prepares and submits separate implementation reports to the various sector ministry (education, agriculture and health) for a particular sector program and utilization of related baskets fund. To deal with the latter problem the GoT could consider issuing a circular or make it explicitly in the financial regulations that no releases will be done for any quarter to any council before an expenditure report for the quarter preceding the previous quarter is received by the Commissioner for Budget (e.g. exchequer release for the third quarter to be issued only after the first quarter report is received by CB). But equally important, is the need to finalize the development of the planning and reporting manual (‘Plan-Rep’) to facilitate consolidation. As regards weak capacity, the longer term solution is capacity building for LGAs/RAS/PO-RALG. However, in the interim GoT could consider compilation of LGA data by zones to reduce overload on the Regional Secretariats.

1.117 Reporting of direct donor support at local government level: Although there was an initiative in PO-RALG to solicit and compile data on direct donor support to LGAs, it was noted that it has become an exceedingly difficult endeavor on the side of PO-RALG particularly in the case of area-based programs. In such cases quite often the magnitude and direction of support is not disclosed/revealed, more so when the support is channeled through private agents and NGOs. This has made planning and reporting on such support very problematic. The donors and NGOs are therefore encouraged to make information on donor support more transparent to inform Government planning and targeting of interventions. Harmonization of donor support and reporting mechanism at local government level is also important.

1.118 Auditing: There has been a steady improvement in this area in the last three years. Based on the Report of the Controller and Auditor General on Local Government Authority Accounts for the Year ended 31st December 2002, all local government authorities (LGAs) were able to submit their final accounts for audit. The total number of LGAs awarded a Clean Certificate amounted 17 (equivalent to 15% of all LGAs) compared with only 12 LGAs (out of 114 that submitted final accounts) in 2001. The number of qualified reports increased by 17% from 59 to 69, while the number of adverse reports declined by about 33% from 43 to 29 (Table 15). Improvements are also reported in the categories of revenues not accounted for and reduction in revenues collected not banked.

Table 15: Summary of CAG Reports on Local Authority Accounts 1999 – 2002

Year Adverse Qualified Clean Not Submitted No Certificate Total1999 51 46% 51 46% 10 9% 0 0% 0% 1122000 75 65% 23 20% 16 14% 1 1% 0 0% 1152001 43 37% 59 50% 12 10% 3 3% 0% 1172002 29 25% 69 59% 17 15% 0 0% 2 2% 117

Source: LGRP

1.119 The improvements summarized in Table 15 above are attributed to ongoing local government financial management reforms, improved accounting particularly in the 32 councils that are now fully running on IFMS, training of Council Directors and

32

Treasurers, as well as stricter supervision and monitoring of LGAs by PO-RALG made possible by the zoning system.

1.120 However, all is not rosy. The CAG report still points to weak budget discipline and control in LGAs as evidenced by losses of cash, questionable payments, stores losses, and unsatisfactory accounting and banking of revenues (Table 16). These weaknesses are partly on account of low capacity in LGAs particularly in terms of finance and accounting skills to carry out appropriate accounting and financial management, lack of equipment and possibly outright theft of public resources/fraud. Most of the LGAs also lack effective internal audit units. LGRP is encouraged to complete the production of an internal audit manual as soon as possible. It was also noted that audit work by CAG has so far focused almost exclusively on councils’ own revenues and not on subventions from central government, which constitute the biggest share of total LGA resources. This therefore requires immediate attention given the significant increase in subventions to LGAs. More also needs to be done to further improve on dissemination and follow-up of the CAG report recommendations.

Table 16: Indicators of Weak Public Financial Management in LGAs 1999 - 2002

Problem 1999(T.Shs. Mill.)

2000(T.Shs. Mill.)

2001(T.Shs. Mill.)

2002(T.Shs. Mill.)

Revenue not accounted for 516 372 162 1033Revenue collected but not banked 208 212 97 500Short falls in revenue collection 18,639 12,602 9,569 49,172Unauthorized expenditure 745 0 0 505Unvoutchered expenditure 2,257 1,335 1,661 4,345Improperly voutchered expenditure 3,864 5,337 3,335 9,293Irregular payments 240 401 321 713Source: CAG Reports for 1999 - 2002

1.121 Sustainable local Government financing: The abolition of nuisance taxes (including development levy) in the 2003/04 budget, has proved a challenge for LGAs (many of which were receiving over 50% of their revenues from development levy) in spite of the toping-up provided by MoF. It is therefore key that an overall strategic frame is developed to guide future decisions on specific local taxes and inter-governmental transfer arrangements. In this regard, government is urged to move ahead with the proposed study on local government revenue and fiscal decentralization issues being coordinated by the Policy Analysis Division in MoF.

1.122 Consultation on Local Government Issues: Two main options for undertaking consultations on local government issues at the national level have been suggested. One is simply to expand LGA representation during the annual PER national consultative meeting. The idea would be to have more participants from across the LGAs, Association of Local Authorities of Tanzania (ALAT) and other stakeholders so as to obtain feedback from best and worst performing councils, rural and urban councils, some average performing councils, and those that are/not fully running on IFMS. However, this would entail having a special session focusing on local government issues. The other option could be to organize separate PER consultations for LGAs by zone OR make use of the annual meeting of local governments organized by ALAT annually to include deliberations on burning public financial management issues at the local level. GoT should consider implementing the first option this year.

33

1.123 Review of Actual Local Government Expenditures: Preliminary Findings 11

Table 17summarizes total expenditure by LGAs for 2001 - 2003. The information indicates that the bulk (61.5%) of total annual expenditure by LGAs goes to finance education, followed by administration (14%)! Other social sectors (health and water) receive 12% and 3.7% respectively. Average annual expenditure on roads is also relatively small, amounting to about 5.5% while expenditure on agriculture claims a mere 3.5%.

1.124 Table 18 also shows that total expenditure (nominal) at the local government level has been increasing over the past three years. The increase was highest for education (274%) reflecting implementation of the PEDP, followed by health (46%), roads (41%), and administration (19%). The only exception was spending on agriculture, which declined by 73% over the same period. Increased spending for the social sectors most likely reflects the poverty reduction strategy (PRS) focus.

Table 17: Total Local Government Expenditure (T.Shs.)

2001 2002 2003Education 128,165,484,577 193,704,882,413 479,182,152,569Health 35,924,648,180 43,741,029,752 52,393,382,860Roads 17,071,470,428 20,100,494,209 24,098,140,809Water 10,133,566,755 8,685,664,140 27,967,159,317Agriculture 20,860,665,912 4,829,684,946 5,546,423,836Administration 44,859,526,919 51,111,640,269 53,192,204,876 Total 257,015,362,771 322,173,395,729 642,379,464,267

Table 18: Total Local Government Expenditure (%)

2001 2002 2003 Average

Education 49.9 60.1 74.6 61.5

Health 14.0 13.6 8.2 11.9

Roads 6.6 6.2 3.8 5.5

Water 3.9 2.7 4.4 3.7

Agriculture 8.1 1.5 0.9 3.5

Administration 17.5 15.9 8.3 13.9

Total 100.0 100.0 100.0

Recurrent vis-vis Development Expenditure

1.125 A comparison of annual recurrent and development expenditure categories over the period 2001 to 2003 shows that recurrent expenditure claims about 85.5% of total expenditure compared to only 14.5% for development activities. A decomposition of annual recurrent expenditure suggests that about 65% is spent on education compared to only 12% for health. By contrast, administration accounts for as much as 14%. Other sectors receive only between 2% to 3.5% of total recurrent expenditure. The data also

11 The actual expenditure data used here is as reported by individual local government authorities (LGAs) to the Local Government Reform Program (LGRP) Secretariat and consolidated by The World Bank. The data is for 108 LGAs, which is about 93% of the 116 LGAs. However, there are still gaps in the data, particularly for development expenditure. As such, the emerging picture should be interpreted with caution.

34

indicates that with the exception of education (whose share increased by 50%), all other sector shares in total recurrent expenditure declined by between 41% and 93% between 2001 and 2003. Analogously, data on development expenditure at the local government level indicates that an annual average of 41.3% is spent on education followed by roads (18.6%). Administration and water each claim about 12% of annual development expenditure ahead of health (10.8%) and agriculture (4.9%). Except for education and water sectors, the share of other sectors in development expenditure has declined since 2001 particularly for agriculture (67%), and roads (59%).

Table 19: Local Government Recurrent and Development Expenditure

2001 2002 2003Recurrent Development Recurrent Development Recurrent Development

Education 119,257,933,121 8,907,551,456 163,927,156,640 29,777,725,773 441,508,665,751 37,673,486,818Health 31,993,980,517 3,930,667,663 38,368,242,615 5,372,787,137 43,193,142,040 9,200,240,820Roads 8,799,308,353 8,272,162,075 10,478,782,985 9,621,711,224 12,674,921,103 11,423,219,706Water 7,773,894,802 2,359,671,953 6,080,478,323 2,605,185,817 6,415,380,027 21,551,779,290Agriculture 18,220,694,296 2,639,971,616 2,738,317,994 2,091,366,952 3,110,971,838 2,435,451,998Administration 38,785,250,406 6,074,276,513 45,305,305,668 5,806,334,601 46,308,616,955 6,883,587,921 Total 224,831,061,495 32,184,301,276 266,898,284,225 55,275,111,504 553,211,697,714 89,167,766,553

Figure 4: LGAs – Sectoral Shares of Recurrent Expenditure 2001

Fig. 1a: LGAs - Sectoral Shares of Recurrent Expenditure 2001

Education54%

Health14%

Roads4%

Water3%

Agriculture8%

Administration17%

35

Figure 5: LGAs – Sectoral Shares of Recurrent Expenditure 2003

Fig. 1b: LGAs - Sectoral Shares of Recurrent Expenditure 2003

Education80%

Health8%

Roads2%

Water1%

Agriculture1%

Administration8%

Expenditure on Personal Emoluments and Other Charges at LG level

1.126 Based on exchequer issues for 2003 where data is available (Fig. 6), expenditure on personal emoluments (PE) accounted for about 78% and the remainder 22% was for other charges (OC). A further decomposition of the 2003 exchequer issues for OC indicate that about 50% went to cater for education, followed by health (19%), toping-up local government revenue (16%) occasioned by the abolition of nuisance taxes in the FY04 budget, and water (7%). Exchequer issues to OC for agriculture, administration and roads were in the range of 2 to 3 percent.

36

Figure 6: LGAs – Sectoral Shares of Exchequer Issues to OC 2003

Fig.2: LGAs - Sectoral Shares of Exchequer Issues to OC 2003

Agriculture3%

Education50%

Health19%

Roads2%

Water7%

Administration3%

Toping LG Revenue16%

Figure 7: LGAs – Sectoral Development Expenditure 2001

Fig. 3a: LGAs - Sectoral Development Expenditure 2001

Education28%

Health12%

Roads26%

Water7%

Agriculture8%

Administration19%

37

Figure 8: LGAs – Sectoral Development Expenditure 2003

Fig.3b: LGAs - Sectoral Development Expenditure 2003

Education42%

Health10%

Roads13%

Water24%

Agriculture3%

Administration8%

Emerging Issues from the preliminary analysis of actual LG expenditure

Dominance of education in local government spending: Education claims the largest share of both recurrent and development expenditure, more so following implementation of the Primary Education Development Program (PEDP).

Relatively large expenditure on administration: Recurrent spending on administration is the second largest after education, exceeding recurrent expenditure on health! Administration is also the third largest component of development expenditure – a level that is at par with the water sector. This raises the issue of consistency and adherence to PRS priorities at local and central government levels.

Limited infrastructure maintenance at local level: Recurrent expenditure on roads, water is quite low (less than 3.5%) which suggest that only minimal infrastructural maintenance takes place. The decline in development expenditure on roads is of concern given the PRS focus on rural roads.

Negligible and declining spending on agriculture: Recurrent and development expenditure on agriculture is lowest and declining. This could be pointing to very limited focus on key services (especially extension services) necessary for agriculture to grow.

Wide disparities in recurrent spending across local government authorities: The data exhibits very wide variations in recurrent spending across LGAs most likely driven by physical facilities (number of schools, health centers/ dispensaries etc.) available in a particular local government authority. Other factors - historical, political, cultural may also have a role. This situation provides support to the proposed introduction of a formula-based system of allocating grants to local government authorities so as to foster transparency in grant allocation and equity across LGAs.

38

High volatility of development expenditure by sector: Except for education and water sectors whose shares in LGAs development expenditure rose markedly between 2001 and 2003, sector shares for the remainder of the sectors declined sharply. The observed high volatility of development spending is likely to be due to the practice of treating this component as a residual, making it to suffer heavily in the event of a resource crunch.

Need for further work: As already alluded to, the actual expenditure data used here is still partial and therefore needs to be completed and validated so as to obtain a more robust picture of the status of public expenditure management at the local government level.

39

ANNEX 1 PRELIMINARY REVIEW OF SECTOR PERS

As in previous years the PER process has included extensive work at sector level in the form of sector PERs. This annex provides a brief summary issues identified in the sector PERs carried out this year for the environment, roads, education, health, HIV/Aids, and agriculture. The current round of sector PERs will be used as preliminary input for the next Budget Guidelines exercise.

A.1 ENVIRONMENT

The intention of the environment sector PER is to review the policies and programs in the sector, the levels and distribution of environmental expenditure and revenue, assess the level of expenditure required in relation to the country’s environmental priorities and poverty reduction strategies and make recommendations to improve the overall performance of the sector. This is the first PER on the environment sector in Tanzania and has been constrained by limited availability of data.

Review of environment policies and programs

The National Environment Policy (NEP), 1997 is the main policy document addressing environmental issues in Tanzania. The main roles of the NEP are to promote sustainable development, develop integrated multisectoral policies and to integrate environmental concerns in sectoral policies and investment decisions. NEP identified six major environmental problems in the country: land degradation, lack of access to improved water sources, environmental pollution, loss of wildlife habitats and biodiversity, and deterioration of aquatic systems and deforestation. In order to combat the above problems the following plans and programs are being implemented at the national level: National Conservation Strategy for Sustainable Development (1988), National Environment Action Plan, National Action Program to combat Desertification (1999) and other programs like Biodiversity Conservation, program to phase out ozone depleting substances and National action plan on climate change. In addition to the above national plans the country is also implementing sector strategies and programs like National Land policy, wildlife policy, Tourism policy, Mining policy, Forestry policy, Fisheries policy and water policy. At the local level the environmental improvement programs include Land management program and Participatory forest management program.

The national PRS recognizes the important linkage between poverty and environment as about 50% of the cash income in rural areas comes from the sale of forest products like charcoal, honey, wild fruits and firewood.

Although the Government has been implementing a number of policies at the national, sectoral and local government levels, the PER found a few weaknesses/gaps. These include lack of an environment module in Tanzania Social and Economic data system, inadequate technical support and capacity building for integrating environment into local

40

and regional development plans, lack of plans or strategies to monitor and assess various community based environmental management projects, lack of programs to strengthen institutional and legal framework for environmental management activities in rural and urban areas and lack of institutional framework for environmental monitoring.

Revenue from environmental resources

Revenues from environmental resources come from taxes on hunting activities, sales of charcoal, sales of drinking water, sales of gemstones, etc. Although the contribution of environmental resources to the national and sectoral revenues is high, under the current system of accounting and tax collection in the country it is almost impossible to identify the revenues at the national level.

At the sectoral level revenues came from wildlife, fisheries, forestry, tourism, water and lands. Among the above sectors, revenues were the highest in the wildlife sector, the main sources of revenues being hunting and exports of live animals (animals and birds, export of Hippo teeth, live tortoises and crocodiles). The revenue from wildlife sector increased by about 9.9 percent in the year 2002/03. However, the limited capacity to collect revenues has prevented it from realizing the full potential.

Among the environmental resource sectors, the fisheries sector is number two in revenue collection. Royalties from fish sales, licenses from fishing vessels/vehicles and export licenses are the main sources of revenue in the fisheries sector. Fisheries sector revenue increased by about 10.5% in year2001/02. Tourism is another sector, which has significant potential for revenues. Most of the income from tourism is from the national parks. The Serengeti national park and Ngorao Park together contributed most of the earnings from tourism. The most important sources of revenue in the forestry sector are timber/logs, flooring wood, Ebony wood, bee wax, semi finished sown timber, carving and arts, honey, tree seeds and licensing. Revenue from forestry increased by about 30.9% in 2002/03 compared to that in 2001/02. In the water sector the main sources of revenues are fee for water rights and fees for construction of wells and sewage disposal. In the livestock sector, sales of cattle and goat hides generated most of the revenues. The environmental sector PER however found that government has not been able to realize significant revenues from fisheries, forestry, tourism and wildlife sectors. There is scope to increase revenues significantly through better pricing that reflect the true economic costs.

Environmental expenditure in Tanzania

The PER found that with the exception of donor funded projects, money allocated by the government and the actual expenditure for environmental improvement projects has been inadequate. Donor funding plays an important role in environmental management in Tanzania. Any expenditure that addresses any one of the key environmental problem areas identified in the country is defined as an environmental expenditure. We do not have accurate data on the allocations and actual spending in the different sectors.

41

In the fisheries sector, expenditure covered activities like surveillance and control of marine and aquatic ecosystems, creation of marine reserves or marine national parks, management of aquatic ecosystems and tourist attractions, community fish farming or aquaculture, research and training organization of sector consultative forums and review of fisheries policy and legislation. The government programs in the forestry sector include forest administration, operationalization of forest and bee keeping policies, conservation and management of natural forests, implementation of afforestation measures, management of forest plantations, capacity building, involvement and participation of community in the management of natural forest and community based organization and participatory forest management. In the wildlife and tourism sectors, the government expenditures addressed loss of wildlife habitats and biodiversity. In the water sector the government programs addressed the lack of accessible good quality water. The programs include surveillance of water quality, provision of water supply to domestic and industrial uses, and to ensure water supply for hydroelectric power.

The PER found that there are significant resource gaps for financing environmental expenditure over the medium term. The MTEF projections for the three years from 2004/05 to 2006/07 shows resource requirements of about 3 billion Tsh for the VPO-DOE and 12 billion Tsh for the NEMC. The spending in relation to the revenues from the environmental sector is inadequate for sustainable use of these resources. Most projects that address key environmental problems are donor driven.

Capacity building for environmental management

The PER found that there is need for capacity building for the National Environmental Management Council (NEMC) as it will be charged with the responsibility of implementing the Environmental Management Act of 2004. The current staffing pattern shows a high ratio of supporting staff to technical staff. Additional staff may be needed for environmental and poverty monitoring at the district and ward levels, environmental law and inspection at the district and regional levels, project management and environmental sanitation, logistical planning and wildlife management.

In addition to the NEMC, the Department of Environment (DoE) under the Vice-Presidents office and headed by the Director of Environment also addresses environmental issues. Under the director of environment there are three assistant directors with responsibilities for environmental pollution and control, policy planning and research and environmental impact assessment. The PER found that there is some duplication of effort between the DoE and NEMC. Although DoE and NEMC has been undertaking some initiatives for capacity building, it is found to be grossly insufficient in relation to the demands. Additional capacity may be developed through recruiting additional staff, imparting management development training and through short-term contractors with an estimated expenditure of Tsh. 91.5 million, Tsh. 96 million and Tsh. 30 million respectively. Low levels of staff remuneration in the sector, low morale, lack

42

of good working environment and long and slow recruitment process are some of the constraints in human capacity building.

Summary recommendations for improvement

In the light of the above findings, the PER made the following recommendations for better performance of the environmental sector.

1. Since the environmental issues are cross cutting, there is a need for all other sectors to address environmental issues in their policies, strategies, guidelines and programs.

2. In order to ensure sustainable utilization of environmental resources, the government may put in a pricing mechanism that reflects the true economic costs.

3. Given the significance of the role of environmental resources for poverty reduction, the future policies for exploitation should put in place mechanisms for their sustainability.

4. The government should invest more resources in environmental management as it reaps revenue from this sector. Such investments should be in proportion to the revenue. A database on all the environmental issues, revenues and expenditure in all sectors need to be developed.

5. Better coordination and eventual merger of NEMC and DOE will avoid duplication of efforts.

6. The PER also recommended reform of the salary structure, recruitment process, retirement plans and ratio of technical to supporting staff. In order to retain and attract better staff incentives to increase morale of the staff may also be provided.

A2 ROADS

The road sector is an important area of focus in Tanzania to achieve the long-term objectives set out in the vision 2025 and the short to medium term targets in the PRSP. The objectives of the current PER are to review the actions taken on the recommendations of the last PER, analyze the budget performance on different kinds of roads, analyze the institutional reforms in the sector and identify key performance indicators of MTEF and to review the progress in dealing with cross cutting issues in the road sector.

Major findings of the 2003 PER and action taken and pending issues

The major findings of the last PER, the actions taken on the recommendations and the pending issues are presented below:

Actual road network length and condition: As noted in the last PER, the actual road network in Tanzania is not known. Efforts have been made to establishing the actual length of TANROADS. But there is no clear record and inventory yet on the exact length of roads and the condition of the network with the PO-RALG.

43

Establish criteria for identifying the roads for maintenance: Consistent with the recommendations of the last PER, a roads maintenance management system is being put in place as is expected to address the issue of criteria of identifying the roads for maintenance.

Time lag before the Performance Agreement (PA) between RFB and the agencies: The PER identified this problem of time lag. An interim Performance Agreement is now signed to allow better flow of funds.Mismatch between timing of funds release and the best time for carrying out roadwork: The above problem identified in the last PER still persists although the RFB is considering the possibility of having pre-financing agreements with the banks.Problems related to Procurement Act and Procurement Expertise: The problem of delay in procurement process caused by the requirements of procurement Act No.3 of 2001 and New Regulations of march 2003 have been partly resolved through decentralization of tendering activities. However there still are concerns on thresholds for procurement.VAT on road works: The VAT on road works has been affecting progress. This has partly been resolved but there are still problems resulting from interpretation of the government decision to abolish VAT on road equipments with effect from July 1, 2003.

Road network in Tanzania

The actual length of the road network in Tanzania is not known. Three systems are being developed to capture the road condition under TRANROADS and LGAs. The first is the Road Maintenance Management System (RMMS), which will be fully operational by 2004, will give more realistic information for budgeting and decision-making on the trunk and regional roads. The second is the District Road Maintenance Management System (DROMAS) being developed by the PO-RLAG which is a comprehensive data management application to assist engineers at the district council and road fund management unit in managing district roads maintenance works. The third system is the RMMS for urban council funded by the IDA.

The government is making efforts to identify new low cost technologies that can improve road conditions. Use of Otta seal, material mixtures and natural pozzolana are some of the low cost technologies tried.

Road sector revenue and expenditures

The main sources of funds for development and maintenance of trunk and regional roads are the Road Fund Board (RFB) fund, treasury and donor funds. The revenue for the RFB comprises four road user charges- fuel charges, transit charges, overloading fees and heavy vehicle license fees. Of the above four user charges, fuel levy accounts for 94 percent of the total revenue collection in the year 2002/03 followed by overloading fees, transit charges and then heavy vehicle licensing fees at 3.1, 2.5 and 0.2 percent respectively. Over the years revenue from all the above four sources have been increasing. While the transit charges are collected at eight entry points, the overload fines are collected at weighbridge stations located along the trunk roads. In addition to

44

the above four sources, the development partners have also been providing resources targeted at trunk and regional roads development. These are meant to support development and maintenance activities. The projected receipts in 2003/04 will meet only 49 percent of the expenditure for road maintenance and spot improvement. Hence there are substantial financing gaps in the road sector.

Sector performance

The overall physical and financial performance for the financial year 2002/03 under TANROADS has been impressive. In the case of trunk roads the physical performance was 105 percent and financial performance 63 percent. In the case of regional roads the physical and financial performance were 71 and 75 percent respectively. However the LGA performance in the road sector has been quite unimpressive, physical performance at 51 percent of the planned activities and financial performance was only 63 percent of the planned budget. But the periodic maintenance of district roads has been quite impressive at about 83 percent.

The PRS outlined two targets to improve the rural road network in Tanzania. The first was to rehabilitate 4500 km of rural roads under the Urgent Roads Rehabilitation Program (URRP) in 12 most poor regions by 2003. The second target was to undertake routine and periodic maintenance promptly on all rural road networks. As the system of accounting does not define clearly the regions where the roadwork is done, it has been difficult to evaluate the progress in relation to the above targets. There has been only a slight increase in the proportion of the total Road Fund allocations to the identified poor regions, from 43 percent in 2000/01 to 45 percent in 2002/03. The Village Travel and Transport Pilot Program (VTTP), under the PO-RALG, a key element of Integrated Roads Project II (IRP) have been implemented in seven districts. Under this project, the districts have implemented a number roads related activities and the financial progress has been impressive.

Institutional reforms

The PER reviewed the progress in firming up institutional framework, establishment and reorganization of specific road sector institutions and capacity building including staffing and equipment services.

(a) Firming up the institutional framework: Although progress has been made by the road sector to firm up institutional framework through drafting Roads Act in 2003, the PER found the following three major issues, which need to be addressed:

1. Autonomy of the TANROADS: The PER recommended that autonomy of TANROADS with a Board of Directors who can make decisions may be needed

2. The PER recommended better division of responsibilities between road sector institutions, specifically separation of policymaking and regulatory functions from those of implementation among MOW, TANROADS and PO-RALG.

45

3. Separation of urban, district and feeder roads from regional roads will have the advantage of addressing problems of coordination and management.

(b) Establishment or reorganization of road sector institutions: The PER found that the road sector has made significant progress in institutional reforms relating to establishment and reorganization of priority institutions like Road Fund Board, TANROADS, MOW and PO-RALG. Further, the Road Fund Board with a wider representation of the stakeholders in the Board of Directors and an internal auditor as recommended by the last PER is in place. However, the PER found that there still are issues relating to the establishment of a sustainable funding mechanism, performance agreement and more efficient collection and flow of RFB revenues. The growth in road fund revenues is low compared to the maintenance requirements, which is an issue that needs to be corrected. In order to increase performance and monitoring, it is recommended that a list of performance indicators be included in the performance agreements. The PER further recommended the formation of a task force comprising of TRA, RFB and treasury to explore the feasibility of resolving problems of delays and inefficiencies in collection and flow of RFB revenues. As regards TANROADS the PER reports that TANROADS has been reorganized and has adopted a lean structure, which gives greater autonomy to the regions. However, this agency face problems of absorption capacity largely due to inadequate procurement expertise and delayed disbursement of funds from RFB. Problems arising from delays in procurement of goods and services reported in the last PER have been addressed by the current amendment of the procurement act that promotes decentralization. MOW, another agency involved in the road sector, has reorganized its structure with fewer staff. But, the PER reports that it still has to revisit its organization to ensure appropriate balance between core staff and supporting staff. The PO-RALG also faces problems with regard to staffing and institutional capacity. Overlapping and unclear mandates between PO-RALG and MOW, absence of a single comprehensive framework for rural roads, absence of a rural roads agency to coordinate implementation of district and other rural roads, lack of transparency in allocation of road funds, cumbersome procurement procedures and LGAS absorption capacity, and lack of donor coordination are other institutional problems identified by the PER.

Cross cutting issues

The PER reviewed sector performance in mainstreaming cross cutting issues including gender, HIV and AIDS, environment and Labor Based Technology (LBT) based on planned targets for FY2002/03. (a) Gender: On gender issues, the following recommendations were made:

a. Revisit performance agreement of RFB to capture information that will indicate the extent to which road sector expenditure has impacted women

b. Strengthen monitoring and reporting on gender performance by sector institutions.c. RFB may allocate resources specifically aimed at capacity building targeting

gender

46

d. Establish road sector database containing information relevant to measure sector performance on gender and encourage learning on good practices in integrating gender in road works

e. MOW set targets to ensure at least 30 percent women’s representation in decision-making body of sector institutions.

(b) HIV and AIDS: The last PER found that HIV and AIDS issues were accorded very low priority in the road sector programs. The current PER does not report any significant progress on specific HIV and AIDS related activities. Since HIV/AIDS is an important issue, the PER makes the following recommendations on HIV and AIDS issues:

a. MOW should concentrate on policy related responsibility for combating HIV and AIDS and capacity strengthening activities including awareness raising workshops be left to PO RALG and other regional and district level institutions

b. Stakeholders in the sector should come up with an agreed set of indicators to monitor HIV and AIDS activities in the sector and an HIV/AIDS monitoring system should be developed.

(c) Environment: There is no information on the progress made in mainstreaming environmental issues in the road sector. The PER recommended that efforts should be made to finalize the establishment of a sound sector environment management and monitoring system and to ensure that adequate resources are allocated for the proposed targets for environmental management in the road sector. (d) Labor Based Technology (LBT): LBT envisages use of low cost locally available technologies and inputs (including labor) to reduce costs of roadwork. Although the last PER reported that there were numerous intended road projects and interventions to support use of LBT, there is not enough information on actual implementation of LBT for roadwork. In the light of the above observations the PER recommended to promote the use of LBT by LGAs and TANROADS through improved PA. The immediate task should be to revise the RFB performance agreement to contain an addendum of list of indicators, which will monitor sector performance in integrating LBT and other cross cutting issues.

A.3 AGRICULTURE

The specific objectives of this year’s agriculture PER are to review progress on implementation of recommendations of the previous PER, review sector performance for the past three years and for the current fiscal year and analyze the recurrent and development budget performance for the above period, review the existing plans and strategies and refine costing of interventions over medium term and assess their impact on PRS targets, identify performance and service delivery on cross cutting issues like HIV/AIDS, environment sector. In addition to the above objectives an analysis of the medium term plan and budget framework is also attempted. Agriculture sector covers three line ministries namely, the Ministry of Agriculture and Food Security (MAFS), Ministry of Water and Livestock Development (MWLD) and the Ministry of Cooperatives and Marketing (MCM).

47

Agriculture sector continues to be the predominant sector in the economy accounting for about 50% of the GDP, 63% of the rural employment and 70% of all export earnings. This sector has since 2001 formulated an Agricultural Sector Development Strategy (ASDS) for achieving the PRS objectives consistent with the stable macroeconomic framework envisaged in the Development Vision 2025 and the MDG The government is implementing the ASDS through an Agricultural Sector Development Program (ASDP). The District Agricultural Development Plans (DADP) were started in 2003/04. The long-term strategies for growth in this sector call for macroeconomic and sector policy reforms which include structural adjustment policies, local government reform program, medium term expenditure framework and public sector reform program.

Implementation of recommendations of the last PER

The main recommendations of the last PER were to set up realistic targets with regard to delivery of public services by all Agriculture Sector Line Ministries (ASLM) and to maintain a database regarding the delivery of such services, inform the general public/clients about the clients’ service charter of the line ministries, increase the overall funding of the sector with larger share for the local government, better inter-ministerial coordination, setting up a block grant that would be accessed only by the Local Authorities (LA) that produce District Agricultural Development Plans (DADP) and to set up a monitoring and evaluation mechanism to assess the performance of LAs.

In line with the recommendations of the last PER, the three line ministries, implemented few reforms. MAFS revised the sub sector performance and public delivery targets, client service charter and established a baseline data to assess performance of services delivered in 2003/04. It also started publication of quarterly progress reports for better monitoring and evaluation. With the implementation of the ASDP, the government increased the allocation of resources and the local component of the donor funded projects have been increased. Harmonization of taxes at the local and central government level, better enforcement of regulatory standards for export crops through crop boards, and improved coordination are other reforms implemented by the MAFS. The MWLD has made some effort in establishing a database on livestock in the country and an early warning system on livestock diseases and to inform the general public/clients of the existence of a client service charter . Efforts are under way to put in place a client feed back mechanism. The funding for MWLD has increased by 12.1% in 2003/04 and task forces to identify constraints in the realization of ASDP objectives have been formed. The MCM, in response to the recommendations of the last PER distributed copies of service client charter to the stakeholders and the general public. It has also reviewed some targets to make them more realistic. The MCM is implementing a member empowerment program in cooperatives and have trained 105 change agents. The ministry also facilitated inspection and auditing of 972 primary societies and 22 cooperative unions.

48

The main constraint in the implementation of the recommendations is lack of funds. Although the government increased the budget allocation to the sector, it was not enough for the line ministries to implement the reforms and deliver public services.

Sector performance

Although agricultural sector recorded an increasing trend since mid 1990s, the actual performance fell short of the targeted growth to achieve reduction in poverty from the current 27% to 14% in 2010. Agricultural sector growth in 1999, 2000, 2001 and 2002 were 4%, 3.4%, 5.5% and 5% respectively.

Crop sector: Production of both cereals and non-cereals fell in 2002/03, the growth rates were -9.8% and -9.6% respectively. Production of cash crops have been very erratic. All cash crops except pyrethrum recorded positive growth rates in 2002/03, ranging from 0.4% (sisal) to 33.5% in the case of cashew nuts. Production of pyrethrum fell by 14.3% in 2002/03. However, it may be noted that production of most cash crops recorded high negative rates of growth in 2001/02 The high growth rates recorded in the year 2002/03 could be due to effects of the steep declines in the year 2001/02. The causes of such large fluctuations in cash crop production need to be analyzed for corrective policy interventions.

Livestock and milk: In the livestock sector, overall meat production is expected to increase by about 2.3% in 2003/04. Production of pork is expected to increase by 13 percent in 2003/04; chicken, mutton and beef by 3.3%, 1.7% and 0.8% respectively. It may be noted that growth in beef, which contributes the largest share to overall meat production, is the lowest. In 2003/04 milk production from traditional cattle is expected to increase by 31% and that of dairy milk by 1.8%. The increasing trend in urbanization and population growth is likely to increase demand for milk and hence there is a need to increase milk production.

Eggs Hides and skin: Production of eggs is expected to increase by 15% in 2003/04. Hides and skins is a major source of foreign exchange. Hides of cattle, goat and sheep are expected to increase by 14.3%, 25% and 9.1% respectively in 2003/04.

49

Constraints to sectoral performance

Poor performance of the agricultural sector is primarily due to the low levels of productivity in the sector. It may noted that for most crops the actual productivity is far less than the potential (Table A.1)

Table A.1 Actual, targeted and potential productivity of selected crops (tons/ha)Crop 2002/03 actual Target(2009/2010) PotentialPaddy 2 (36) 3.5 5-6Maize 1.2 (24) 3 5Cassava 2.6 (12) 5 20-25Sorghum 1 (50) 2 NACashew nuts 0.69 (49) 1.4 NACoffee 0.2 (33) 0.7 0.6Cotton 0.6 (43) 0.7 0.6Sugar 7 (61) 11.5 NATea 1.8 (51) 2.5 3-5Tobaco 0.7 (39) 1.5 1.8Sisal 1 (67) 2 1.5Pyrethrum 0.3 (30) 1 NAFigures in parentheses are actual yields as a percentage of potential or target yieldSource: MAFS

For food crops only 12 to 50% of the potential yields are actually realized. For cash crops it ranged from 30 to 67%. Such low productivity is due to the use of traditional farm implements like hand hoe, low adoption of improved seeds and fertilizers, poor infrastructure facilities in distribution and inputs, poor marketing facilities, dependence on rainfed agriculture and adverse weather conditions and inadequate extension services. Low levels of public and private investment, inadequate capacity in managing agricultural projects and limited access to financial services are also affecting sectoral performance. In the livestock sector, severely undeveloped markets for both live and livestock products, limited processing capacity, inadequate number of qaulified staff and livestock keepers and high initial capital investment requirements are the major constraints.

Service delivery

The year 2002/03 was the first year of implementation of the ASDP. The MAFS with the implementation of the ASDP is concentrating on the public delivery of irrigation, improved seeds, agricultural equipment, research and extension services, crop protection and inspection services, grain storage and farmer training services. Seventeen irrigation schemes which brought 6065 hectares under irrigation have been constructed. In addition 26 irrigation schemes which cover a total area of 7639 hectares are under various stages of implementation. The government is also working on an irrigation master plan which envisages bringing about 29.4 million hectares of land under irrigation. Distribution of

50

quality seeds is another activity of the MAFS. In 2002/03 the actual supply of improved seeds was only 35% of the requirement of which the government seeds farms contributed about 1%. About 70% of the land in Tanzania is tilled with hand hoe. The MAFS in 2002/03 implemented various measures to improve mechanization of agriculture. These include import and distribution of tractors, tillers and other machinery, oxenization and training of farmers in the use of farm machinery. Agricultural research in 2002/03 emphasized on the development of breeder seeds for paddy (nine new breeder seeds for paddy), sorghum, beans, pigeon peas, soybean, wheat, sweet potato and cotton. Breeder seeds for the above crops were produced and distributed for farmer use. In order to improve distribution of inputs, the MAFS through Agricultural Inputs Trust Fund has provided loans and subsidies on interest rates to importers and distributors of inputs. Promotion of low cost technologies, crop protection services, crop inspection services and farmers’ training are other activities undertaken by the MAFS.

The MWLD has not been able to provide services like animal health services, improvement of genetic potential of livestock, supply of water, and establishment of livestock markets. There have been some progress in the development of rangelands through construction of charco dams. The ministry has also initiated research on improving the genetic stock of livestock and to improve the quality of feeds and forage. The MCM provided services in the area of crop marketing, establishment and registration of farmer organizations, accessing affordable financial services to the poor in rural and urban areas and promotion of non-traditional crop export.

Sector Budget performance

Public spending in the agricultural sector include allocations to the MAFS, MCM and MWLD with recent emphasis on decentralization and increased allocations to LAs. Planned and actual Sectoral expenditure trends over the period 2000/01 to 2003/04, presented in table A.2, shows a gradual decrease in expenditure both in absolute and as a proportion of the total government expenditure from 4.3% in 2000/01 to 2.5% in 2002/03. The planned spending in the agriculture sector as a percent of the total government spending in 2003/04 is 3.7%. These shares are much below the average of 6 to 7 % in less developed countries where agriculture sector is the dominant sector in the economy. There is also a large discrepancy between actual and planned expenditure in 2002/03, as high as 52.2% in 2002/03.

51

Table A2 Agriculture expenditure trends (Mill. Sh)

2000/01 2001/02 2002/03 2003/04Agricultural Expenditure

Planned

Actual Planned Actual Planned Actual Planned

Recurrent 24644 33797(137.1)

33719 31964 (94.8)

64090 27602 (43.1)

47647

Development 26708 21362(74.4)

25074 20843(89.1)

38114 16551 (43.4)

35213

Total agricultural 53352 55158(103.4)

58293 52797(81.8)

84541 44153 (52.2)

82,860

Government Expenditure

954700 1296100 1787400 2198600

Agricultural/Total Govt. (%)

4.3 4.0 2.5 3.7

Figures in parantheses are actual as a percent of the planned expenditures

Source: PER 2002/03 Table 3.1 and MAFS and MCM, Appropriation Accounts. MLWD not included

The internal and external factors that lead to poor fiscal performance and thus deviations between actual and planned expenditures indicates the poor fiscal performance of the government. Over the period considered, the deviations between planned and actual expenditures have been increasing.

Budget performance in the MAFS has been impressive in the last two years (Table A3).

Table A3: MAFS total expenditure (domestic and foreign)Expenditure 2000/01 2001/02 %change

2000/01/2001/02

2002/03 %change 2001/02/2002/03

2003/04 %change 2002/03/2003/04

Planned

Actual Planned

Actual Planned

Actual Planned

Recurrent 10,315 10,314 9,560 8,702 -7.3 14,659 15,672 53.3 22,629 54.4Development (L&F) 15,807 13,006 15,957 14,661 1.0 18,604 14,533 16.6 31,518 69.4Total Rec. and Dev. 26,121 23,320 25,517 23,364 -2.3 33,262 30,205 30.4 54,147 62.8

Source: Appropriation Books of Accounts and other financial statements of 2001, 2002, 2003 and MTEF of 2003/04- 2005/06.

The expenditure, recurrent and development from both domestic and foreign sources, allocated to the MFAS increased by 30.4% and 62.8% respectively in 2002/03 and 2003/04. The deviations between planned and actual expenditures were also insignificant. The budget allocation and performance in terms of ratio of actual disbursement to allocation for the MCM also improved over the last three years. Both the allocation and actual expenditure increased over the last three years. The MCM is implementing two projects, the Coffee/Cotton Markets Development and Trade Promotion and the Development of Rural Markets. However, project implementation is constrained by the scarcity of funds allocated to the ministry. Revenue collection on the other hand improved significantly between FY 02 and FY03.

52

Expenditure and PRSP: The PRS identifies agriculture as the priority sector for poverty reduction. In 2001/02 and 2002/03 the proportion of actual expenditures to the proposed PRS expenditures were 64.1% and 74.2% respectively (Table A.3).

Table A.3 PRSP proposals and actual expenditure for agriculture sector (Mill. Shs)

Sn. Budget Item 2001/02 2002/03 2003/041 Planned Expenditure 25,517.21 33,262.49 54,146.812 Actual Expenditure 23,363.7 30,205.58 -3 PRSP Proposed Budget 36,455.9 40,727.9 -4 Percentage of the Planned (local contribution

only) to PRSP Proposed Expenditure 30 40.5 -5 Percentage of the Actual (local and external

contributions) to PRSP Proposed Expenditure 64.1 74.2 -Source: MAFS 2003

The analysis of the sector budget shows that support in areas like agricultural research and extension, irrigation, technological innovation, training and disease control has been increasing over time. However, overall budget support does not reach the 10% recommended level to achieve the desired poverty reduction. Further, the deviations between planned and actual expenditures continue to persist.

Policy interventions in agricultural sector

In order to increase productivity in the agricultural sector the three line ministries have increased actual and budgeted expenditures on line items like development of policy, regulatory and institutional framework; sector information; strengthening training, advisory and technical services; research services; and private sector development (Tables A4, 5 and 6).

Table A4: Trends in recurrent expenditure in priority areas of activity under MAFS (million T.Shs.)

Activity 2001/02 (actual)

2002/03 (actual) 2003/04 (Budget)

Policy, regulatory and institutional framework

859.32 890.50 (3.6) 1093.75 (22.8)

Agril. Sector information 31.94 53.77 (68.3) 146.57 (172.6)Training advisory and tech. services 4615.16 9711.05 (110.4) 17102.18 (76.1)Agril. Research services 220.08 370.45 (68.3) 588.28 (58.8)Private sector development 4.63 7.79 (68.3) 33.57 330.9)Cross cutting issues 38.66 44.26 (14.5)Figures in parentheses are growth rates over the preceding year

The annual percentage growth rates show significant increases in actual and budgeted expenditures for all priority activities under the MAFS. All priority activities under the MWLD show significant increase in expenditures in the year 2002/03 (Table A5). Actual expenditures for all activities except private sector development shows significant growth

53

in the year 2002/03. The government increased the spending on livestock information significantly. The MCM also increased its spending on priority areas(Table A6) like policy and regulatory framework, technical services, agricultural finance, strengthening information, etc.

One of the innovative approaches in the ASDS is the consideration of cross cutting issues like rural infrastructure, electrification and communication, HIV/AIDS, malaria, gender, empowering youth, and environmental management in planning and implementation. The ASLMs will start with programs for gender, HIV/AIDS and environment in their budget in 2004/05. The Land (amendment) act submitted to the parliament in 2003 is expected to provide an environment for private sector development in investment activities that require land. Although HIV/AIDS is affecting of the work force in agriculture sector thus reducing productivity, ASLMs have taken only a low profile in the fight against this epidemic. The PER recommends voluntary counseling and testing for employees, access to up-to-date information on HIV/AIDS issues, leadership commitment on HIV/AIDS issues, non-discriminatory environment and provision of condoms under all the LMs. Financing of HIV/AIDS interventions within the ASLMs have been erratic due to the absence of well defined strategies.

Table A5: Trends in recurrent expenditure in priority areas of activity under MWLD (million TShs)


2002/03 (actual) 2003/04 (Budget)

Policy and institutional framework 308.13 348.45 (13.1) 557.37 (60.0)Livestock information 18.50 257.94 (1294.3) 87.30 (-66.2)Research and training 742.65 778.05 (4.8) 1200.48 (54.3)Production, extension & rangeland development

879.31 1554.66 (76.8) 1530.00 (-1.6)

Delivery of veterinary services 659.87 1958.55 (196.8) 1406.00 (-28.2)Figures in parentheses are growth rates over the preceding year

Under the environmental initiative, the MAFS implemented soil and water conservation measures and measures for the control of water hyacinth weed in the river Kagera and Lake Victoria. In 2004/05 the MAFS plans to implement irrigation schemes such as protection and demarcation of catchment areas and land use planning.

54

Table A6: Trends in recurrent expenditure in priority areas of activity under MCM (million TShs)


2002/03 (actual) 2003/04 (Budget)

Policy and regulatory framework 80.89 306.77 (279.2) 1064.68 (247.1)Training and advisory services 979.56 1357.04 (38.5) 983.32 (–27.5)Technical services 450.21 1166.55 (159.1) 1018.32 (-12.7) Strengthening information system 300.00 463.05 (54.4) 1000.00 (116.0)Private sector development 1860.50 629.07 (-66.2) 100.00 (-84.1)Research services 660.30 690.00 (4.5) 824.68 (19.5)Advisory services 500.00 685.64 (37.1) 600.00 (-12.5)Agricultural finance 46.48 510.00 (997.2) 200.00 (-60.8)Figures in parentheses are growth rates over the preceding year

Summary of major observations and recommendations of the PER

The agriculture sector PER made the following observations and recommendations in the agriculture sector to achieve the overall objective of poverty reduction by 50% by 2015.

a. Agricultural growth rate in 2002 (5.2%) was lower than the planned growth rate of 6% to achieve the desired poverty reduction. The PER suggested to increase the budgetary support for agriculture to at least 10% of the total government expenditure.

b. The levels of investment and output growth are low in the agriculture sector. The PER recommends an increase in public investment (such as construction of rural roads and marketing infrastructure), to provide more budget support to MCM and putting in place policies and measures that would attract private investment.

c. In adequate delivery of research and extension services is another constraint limiting agricultural production. Research to improve the technology and seeds to suit local conditions and more extensive extension support are recommended.

d. Inadequate planning and implementation capacity at the district level is affecting the implementation of projects. Improved civil service deployment at the agricultural sector in the local government is recommended.

e. As regards HIV/AIDS strategies and gender mainstreaming, the ASLMs should implement the activities that have been proposed.

f. There still are problems with the devolution of responsibilities to LGAs. It is recommended to strengthen the inter-ministerial coordination and devolve ASDS/ASDP activities earmarked for LGAs.

g. The PER further recommends to put in place a monitoring and evaluation mechanism for performance assessment.

55

A.4 EDUCATION

The fiscal year 2003/04 education sector PER builds up on the previous PER updates. It first summarizes the recommendations of the last PER and actions taken on those recommendations and reviews sector plans, strategies and performance. It then tracks the budget performance and expenditures at the sector and sub-sectoral levels in the last three financial years. The PER then analyzes sector targets and achievements, progress in cross cutting issues like gender and HIV/AIDS, projections of resource requirements and availability in the medium term and finally makes recommendations for policy reform in budgeting financial management and accountability.

Developments since the last PER

Since the 2002/03 PER progress has been made on a number of outstanding issues. These include: Completion of a Secondary Education Development Plan in early 2004. A National Education Fund was established to cater to the needs of disadvantaged

children. An education database unit has been established to monitor the progress in

educational achievements In order to enhance transparency and accountability in the use of PEDP funds the

government has started to publish allocation of funds by councils and schools Primary school enrollment has been made compulsory and the attendance rules of

2002 has been enforced

Review of sector plans, strategies and performance

The Education Sector Development Program (ESDP) developed in 1996 outlines the framework for development of the education sector. The ESDP covers basic education, higher education and vocational education.

Primary education: Under the ESDP a Primary Education Development Plan (PEDP) is being implemented to improve enrolment and quality of education at primary school level. The PEDP implementation resulted in significant increase in enrollment in 2003, though not in tune with the increase in 2002. The second component of the PEDP has been to improve the quality of primary education. There has been significant increase in the quality of primary education through improvement in teaching and learning materials, pre-service and in-service training programs, improvement of school environment and better monitoring and evaluation systems. However increased enrolment has affected quality as improvements in capacity building did not kept the pace with increases in enrolment. For instance only 80 percent of target could be achieved in classroom construction and construction teacher’s houses. There has also been a decrease in in-service training.

56

Non-formal education: In 2003 the government developed the medium term strategy and action plans for adult and nonformal education. Three separate programs have been developed for adult and nonformal education; COBET Cohort one for children aged 11-13 years, COBET Cohort two for youth aged 14-18 and Adults aged 19+. However, implementation was constrained due to lack of enough facilities, instructors and instruction materials.Secondary education: The education sector is finalizing a Secondary Education Development Plan (SEDP). The gross enrollment continues to be low though there was a small improvement in 2003. Enrolment at lower secondary (Forms I-IV) is well below 10 percent and the problem is more pronounced at upper secondary level. Lately, there is a general deterioration in transition rates, drop out rates and repetition rate. Inadequacy of qualified teachers and managers, quality of student intake and financial constraints are the main reasons for poor performance. The number of schools increased by 6 percent.

Higher and technical education: The government launched the Higher and Technical Education Master Plan (HTEMP) in 2003 but the implementation has not yet started. Data from the universities and technical colleges show that there has been significant increase in both undergraduate and graduate enrolment. There has also been significant increase in teacher-student ratios. However data from the universities show that there is considerable cohort wastage in higher learning institutions.

Sector policies, performance indicators and PRS targets

The PRS (2000) had set the following specific targets for the basic education sector to be achieved by 2003:

1. raise primary enrolment to 85 percent2. increase transition rate from primary to secondary from 15 to 21 percent, 3. reduce drop-out rate in primary school from 6.6 to 3 percent, 4. raise net primary enrolment from 57 to 70 percent, 5. increase number of students passing standard seven from 20 to 50 percent, 6. expand adult education programs, and 7. to monitor changes in literacy rate of 15-24 year olds.

The assessment of performance in relation to PRS targets have shown that there has been significant increase in primary and secondary enrolment though the validity of the data on enrolment rates is questionable. The gap between the gross and net enrolment rates continues to be high which shows that the over and under pupils together account for 17 percent of total enrolment. Annual drop out rates at the primary level are about 5 percent and the retention rates shows only modest signs of improvement. Similarly the pass rates in Primary School Leaving Examination were low but a significant improvement has been achieved in 2003. The PRS target of 50 percent pass rate in PSLE by 2005 could be achieved only if the pass rate observed in 2003 could be sustained. Although the literacy rates among 15-24 year olds are high there are significant regional and gender disparities. The ratio of girls to boys at primary schools has improved significantly in the past three years largely due to the measures under the PEDP. Though the gender gap at the secondary school has declined it continues to be significant. In sum, the PRS targets regarding gross and net enrolment at primary level and transition rate from standard VII

57

to Form I has already been achieved. It is highly likely that the targets regarding percentage of students passing PSLE, net enrolment rate in secondary school and girls/ boys ratio in primary and secondary levels will be achieved by 2005. However, the targets for reduction in primary school drop out rate and secondary gross enrolment rate are highly unlikely to be achieved.

The 2003 PER noted that the current PRS targets for the education sector focus on primary education alone and hence there is a need to include targets for secondary and tertiary levels of education. Further gender related indicators in line with the MDGs are needed at the secondary and tertiary levels.

Review of Education Budget and Expenditure

Expenditure on education continues to increase over the period from 2000/01 through 2003/04 (Table 1). As a percentage of GDP, total spending on education has been showing an increasing trend from 3.3 percent in 2000/01 to a budgeted 5.1 percent in 2003/04. Share of education sector in the expenditure on priority sectors ranges from

Table 1. Government expenditure on education, 2000/01-2003/04 (Billion T.Shs. current prices)

2000/01 2001/02 20020/03(Budget)

2003/04(Budget)

Total Education Expenditure 254.9 344.9 474.3 570.6 Recurrent 189.2 282.1 400.0 464.4 Recurrent as % of total 74.2 81.8 84.3 81.4 Development 65.7 62.8 74.3 106.2 Development as % of total 25.8 18.2 15.7 18.6Total govt. expenditure on priority sectors 494.7 749.8 999.9 1258.1Education expenditure as % of exp. on priority sectors

51.5 46.0 47.0 45.4

Education expenditure as % of GDP 3.3 4.0 4.9 5.1Source: URT (2003k)

51.5 percent in 2000/01 to 45.4 percent in 2003/04. The above data shows a gradually declining trend in the share of education expenditures in total government spending on priority sectors. Share of recurrent expenditure in the total increased from 74.2 percent in 2000/01 to a budgeted 81.4 percent in 2003/04 with an average of 80.4 percent over the four year period considered.

58

Recurrent expenditure

Sub-sectoral analysis of recurrent spending on education shows that the government is putting increasing emphasis on primary education, followed by higher and technical education. The share of primary education in total recurrent spending increased from 65.4 percent in 2000/01 to a budgeted 71.4 percent in 2003/04 (Table 2). Such high shares for primary education is a reflection of the resources required for wages and salaries and material supplies in this sub-sector. Share of higher and technical education is showing a gradually decreasing trend from 21.9 percent in 2000/01 to 16.7 percent in 2003/04. The pattern of recurrent expenditure seems to show an increasing bias towards the primary education with falling emphasis on secondary and higher education. This may result in dearth of qualified manpower with superior skills and the labor market crowed by labor force with basic education only.

Table 2: Distribution of recurrent and development expenditure by sub-sectors,2000/01-2003/04 (%)

Sub-sector 2000/01 2001/02 2002/03 2003/04Recurrent Develop. Recurrent Develop. Recurrent Develop. Recurrent Develop.

Primary education

65.4 75.1 61.3 77.4 73.9 67.5 71.4 NA

Secondary education

6 10.2 8.2 9.6 5.8 10.6 6.6 NA

Teacher education

1.4 0.8 2.0 1.4 1.3 2.1 1.6 NA

Higher and technical education

21.9 6.8 22.8 4.6 15.2 13.0 16.7 NA

Administration and others

5.3 7.1 5.7 6.9 3.8 6.7 3.7 NA

Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 NASource: URT, (2003m)

Personal emoluments and other charges account for a major share of the recurrent spending on education, the latter accounting for 75 percent in 2002/03 and the former 25 percent of the recurrent expenditure. Among the sub-sectors, personal emoluments account for a major share of recurrent expenditures except in the case of higher and technical education. Other charges accounted for more than 98% of the recurrent expenditures in the case of higher and technical education. Analysis of the recurrent budget performance shows that in 2001/02 and 2002/03 about 99 percent of the approved recurrent budget was disbursed to the two ministries and out of the total disbursements, an average of 98 percent was spent.

Development expenditure

Both foreign and domestic funds have been supporting development spending in the education sector, the former being the main source of funds for the development budget. The foreign component of the development budget increased from 87.5 percent in 2000/01 to 95.6 percent in 2003/04. The corresponding share of the domestic funds

59

increased from 12.5 percent in 200/01 to 16.6 percent in 2001/02 but then fell to 5.7 percent in 2002/03. The PER notes that there are significant resource inflows that are not properly accounted for in the government accounting system. Development expenditures by sub-sectors show higher emphasis on primary education followed by secondary education, higher and technical education (Table 2). Teacher training got the least share. The above sub-sectoral expenditure pattern is similar to the pattern followed in the case of the recurrent expenditures. Data on approved budgets versus actual expenditure shows that in 2002/03 the actual expenditure by MOEC was only 68 percent of the approved budget (76.7 % in 2001/02). In the case of MSTHE, the actual expenditure was only 90.8 percent of the approved budget in 2002/03. This could be because a major share of the development budget is foreign funded and that donors are holding the release funds as the government fails to meet counterpart budgetary commitments.

PEDP expenditure analysis

A major share of the funds that are allocated to education sector in general and primary education in particular, is expended through the implementation of the PEDP. The sources of funds for PEDP and their trends are presented in Table 3.

Table 3: PEDP Expenditure analysis (Billion TSh.)

2001/02 2002/03 2003/04A. Source of fundsGovernment financing 167.1 (71.9) 175.1 (60.4) 130.3 (46.6)External financing 65.2 (28.1) 114.6 (39.6) 149.4 (53.4)Total 232.3 289.7 279.6B. Government financing of PEDPRecurrent expenditure 165.2 (98.8) 173.8 (99.3) NA PE 132.2 138.4 NA OC 33.0 35.4 NADevelopment 1.9 (1.2) 1.3 (0.7) NATotal 167.1 (100.0) 175.1 (100.0) NA

Source: URT, (2003b)Figures in parentheses are percentage of the respective total

The share of government financing for PEDP shows a decreasing trend from 71.9 percent of the total in 2001/02 to 46.6 percent in 2003/04. During the above period the share of external funding nearly doubled from 28.1 percent to 53.4 percent. The data on donor funding, however, do not capture the off-budget funding. The dominance of government funding in PEDP recurrent budget is a reflection of the interest of development partners in developments programs as opposed to recurrent spending. Government spending accounted for 98.8 percent of the government funding for PEDP in 2001/02 which increased to 99.3 percent in 2002/03. A further breakdown of government funding for PEDP recurrent expenditure shows that while PE accounts for 80 percent of the total, OC accounts for only 20 percent.

60

Financial management and role of local government authorities

The local government authorities have been entrusted with the task of implementing the PEDP. Resources for PEDP fall under three categories: capitation grant, other charges and development or capital funds. The funds move from the Ministry of Finance to the Accountant General, District Councils and then to the schools. In general there has been an increase in funds allocated to districts during the first two years of PEDP implementation. Lately there have been problems in the flow of funds largely due to resource gaps. The PER found that there are substantial variations in per capita allocations among the districts and there is no clear criteria for allocations among the councils. The assessment and practice of managing PEDP funds show that though there has been a general improvement over the first year of PEDP implementation the flow of funds from the center to the districts depends much from which window the funds come and that funds managed by donors are not easily ascertained in terms of disbursement schedules. It is recommended that school accounts be computerized for more efficient flow of funds. Delays in transfer of funds from the center to districts and then to schools, misappropriation of funds at school level, lack of transparency and schools receiving less than allocated amounts are other problems identified.

Cross cutting issues

The main cross cutting issues in the education sector are HIV/AIDS and gender issues.

HIV/AIDS and education sector

The main effect of AIDS on education services are reduction in supply of experienced and newly recruited teachers, depletion of resources due to increased AIDS related expenses and transaction costs associated with transportation and burial of teachers who die of the disease. The Ministry of education undertook the following activities to deal with HIV/AIDS problem:

1. Strengthening management and coordination of programs on AIDS activities-An AIDS coordination unit has been established to coordinate the activities.

2. Reducing vulnerability of school youth to HIV/AIDS/STI. This has been achieved through integration of HIV/AIDS education component in the curriculum, distribution of learning materials related to AIDS.

3. Reducing unprotected sex among men and women: MOEC leaders and workers and peer educators have been trained to conduct workplace education sessions in departments and institutions.

4. Promoting the culture of values and norms in the society that encourage positive attitudes and decision making about sexual matters. A number of cultural activities have been designed and carried out with this objective.

5. Research to appraise the impact of AIDS on the education sector and to identify the needs for social behavior and communication on HIV/AIDS and promoting use of health and social services on HIV/STIs have also been conducted to devise appropriate policies to combat this epidemic.

61

Issues that need attention: MOEC and PO-RALG share the responsibilities of HIV/AIDS education, MOEC developing the guidelines and teaching materials and PO-RALG the implementing them at the district level. However the modalities of sharing these responsibilities have not yet been worked out. Although substantial funds have been spent on AIDS related activities, it has been hard to track the expenditures on HIV/AIDS related interventions. Due to the fragmented nature of HIV/AIDS interventions in the sector, it has been hard to verify whether the identified activities are fully funded or not. Further, the funding of HIV/AIDS related interventions are not fully captured in the national accounting system. Participation of local communities and NGOs is important in order to have a sustainable and coherent HIV/AIDS program. The 2003/05-2006/07 education sector MTEF underlined the need for strengthening school AIDS committees that are composed of teachers, students, pupils, parents and community members including faith organizations. Schools that have not developed such programs should be encouraged to pursue this. The MTEF identified programs to cover children in non-formal school system, adult learners and students and teachers with disability. There is a need to ensure sufficient budget commitment for these activities. Resources should also be directed towards programs that mitigate the impact of AIDS, like programs to support vulnerable HIV/AIDS orphans.

Gender and education

Primary education: There are apparent gender differences in the education sector as revealed by the Gross and Net Enrolment rates. Although the gross enrolment rates have improved significantly in the past three years, difference between the male and female enrolment tends to increase; from 2 percentage points in 2001 to 7 percentage points in 2003 (Table 4). The PER found that although an equal number of boys and girls are enrolled at the age 7, older girls are finding it difficult to start schooling than their male counterparts. Similarly, the percentage of girls passing the primary school examination seems to be significantly lower than males. These trends in enrolment seem to indicate the PRS and MDG target for parity in overall enrolment of girls and boys in primary education by 2005 may not be achieved.

Table 4: Gender disaggregated performance indicators at the primary level

Indicator 2001 2002 2003Net Enrolment Ratio (%)Female 65 79 87Male 66 82 90Gross Enrolment ratio (%)Female 83 96 102Male 85 101 109Pupils Passing PSLE (%)Female 21 20 --Male 36 34 --Girls/Boys in Primary SchoolGirls/Boys: All standards 0.97 0.96 0.95Girls/Boys: Standard one 0.93 0.94 0.94Source: URT, 2003f

62

Secondary education: The gender gap in transition rates from Standard IV to Form I are not large despite the fact that boys tend to outperform girls in the primary school leaving examination. However, the gender gaps in enrolment rates are large (Table 5). The ratio of girls to boys in secondary education tends to fall at higher levels of secondary education. These results suggest that elimination of gender disparity in secondary education by 2005 would require more serious interventions to reduce dropout rate of female students in the secondary school.

Table 5: Secondary education Gender indicators

Indicator 2001 2002 2003Transition from standard seven to Form IFemale 19 19 18Male 21 22 20Girls/Boys in Primary SchoolForm I-IV 0.86 0.84 0.84Form I 0.96 0.88 0.93Form IV 0.85 0.78 0.80Form VI 0.51 0.47 0.50Source: URT, 2003f

Higher education: Gender disparity is much higher at the technical and higher education levels. A number of programs like pre-entry science program for female students and female undergraduate scholarship program have been implemented.

Costing of interventions

The projected resource requirements for the education sector during the MTEF period (2004/05-2006/07) are presented in this section. The proposed policies concentrate on interventions for secondary education, Adult and Non-formal Education (AE/NFE), and HIV/AIDS related interventions.

Secondary Education Development Plan (SEDP)

The overall objective of the SEDP is to increase the number and of youth in Tanzania who complete secondary education with acceptable learning outcomes. In order to achieve this SEDP has outlined programs to improve access, ensure equity of participation in undeserved areas by geographical location, remove gender and income inequalities, improve the performance of students by increasing the passing rate and increase efficiency of secondary education through management reforms and devolution of authority. The projected resource requirements to implement the above programs are presented in Table 6.

63

Table 6 Projected Resource Requirement and available resources for the years 2004/05-2006/07 (Tsh. Millions)

REQUIREMENTS 2004/05 2005/06 2006/07Recurrent1] Teacher Salaries 13,162 14,955 17,5092] Non Teacher Salaries 2,632 2,991 3,5023] Non Salary Recurrent Expenditure 11,523 13,508 14,5994] Government Schools Student Capitation Grant

7,682 9,006 10,814

5] Non- Government Schools Student Capitation Grant 2,078 2,362 2,7246] Government Schools Fee Subsidy 4,878 6,004 7,5707] Targeted Bursaries 1,097 1,351 1,7038] Form V-VI Supplementary Capitation 390 421 4629] Disadvantaged Schools Capitation 576 675 81110] Exam Fee Waiver 1,000 1,385 1,63511] Secondary Teacher Training 6,494 9,848 11,02912] Total recurrent 51,512 62,506 72,359Development1] Public Financing of Government Secondary

Schools22,157 31,790 44,008

2] Government Secondary Schools Additional School Capital

6,132 9,054 7,549

3] Total Cost for Teachers Training Institutions 1,398 1,628 2,5074] Total cost for University Teacher Training 389 267 5375] Total development 30,076 42,739 54,602TOTAL REQUIREMENT 81,588 105,245 126,961

AVAILABLE RESOURCES

Total Recurrent Resources 54,799 64,663 73,716Total Development Resources 14,575 16,616 18,942Total 69,374 81,279 92,658Resource Gap 12,214 23,966 34303Resource Gap as % of Projected Resource Requirements

15 23 27

Comparison of the resource requirements in relation to the available resources shows a gap of 15 percent in 2004/05 which increases to 27 percent in 2006/07. It may be noted that the SEDP is giving increasing importance to the private sector in its plan.

64

Adult and Non Formal Education (AE/NFE)

The objective of AE/NFE is to target out of school children, youth and adults, especially girls, women, disadvantaged groups and nomads. The projected resource requirements for the Adult and Non Formal Education (AE/NFE) over the medium term (2004/05 – 2006/07 are presented in Table 7.

Table 7: Projections and resource requirements for AE/NFE for the years 2004/05-2006/07 ((Tsh. Millions)

Category of the Budget 2003/04 2004/05 2005/06 2006/071] 11-13 out of school 1,221.3 3,640.0 3,066.4 81.92] 14-18 out of school - 1,980.2 4,459.8 4,039.23] Adult Education Learners 1,868.9 5,936.9 7,677.3 4,997.64] Post Literacy Group 43.3 228.3 793.3 793.35] Other Indirect Costs (Capacity Building) 772.0 2.2161.3 2,176.7 1,121.96] TOTAL 3,905.4 13,946.7 18,173.4 11,033.9

There is a drastic increase in resource requirement between years 2003/04 and 2004/05 as the project envisages initial enrolment of a significant number of out of school children.

HIV/AIDS related interventions

Resource requirements for HIV/AIDS related activities are projected based on the assumption of projected increase in enrolment during the medium term and substantial expansion of HIV/AIDS related activities.

In order to improve access to secondary education the SEDP plans to complete existing schools, increase enrolment, increase the number of teachers in teacher deficient schools, relocate teachers based on a pupil: teacher ration of 30:1.

Main findings and recommendations

The main findings of the education sector PER are summarized below:

There are marked improvements in the implementation of PEDP program and enrolment ratios. However, the challenge remains to keep the newly enrolled children in school until they reach standard seven. Gender issues in education remain a concern, even though gender balance has almost been achieved at primary level and progress is required at the secondary level. Measures to reduce dropout of girls particularly at higher levels and to enhance the performance of girls should be put in place.

Structure of the education budget is dominated by the recurrent component at both national and education sector levels. Overall, both the recurrent and development budgets have been increasing over time. This upward trend is an indication that the activities related to education are expanding. Compared

65

with the other priority sectors, education sector has been receiving the largest proportion of the total budget.

Although the total development expenditure on education has in general been increasing during the period under review internal resources to education have been fluctuating with an overall decreasing trend. The external resources to education have recorded an increasing trend.

The budget structure for education sector favors primary education. One notable implication of such a bias is low transition rate to secondary level.

There is a noticeable mismatch between PEDP resource requirements and the actual budget allocations. This has adverse effects on planned PEDP activities. The budget mismatches between PEDP resource requirements and the actual budget allocations have affected planned PEDP activities and subsequently PRSP objectives and targets.

HIV/AIDS interventions activities have been found to be fragmented. This renders tracking of spending on HIV/AIDS interventions difficult. The newly introduced HIV/AIDS GFS code and Coordinating Unit will ease tracking expenditures on HIV/AIDS

In light of these findings, the report makes the following recommendations:

I: Sectoral Plans and Strategies: The PER recommended sectoral plans and strategies like establishment of more centers for Integrated Community Based Education and Complementary Basic Education, preparation of micro-plans by all schools, extension of coverage of school mapping to all districts, extension of Education Management Information System to all districts, and establishment of Education Database Unit within Ministry of Science Technology and Higher Education (MSTHE).

Though PEDP implementation has been largely a success, actions need to be taken to solve outstanding problems identified with respect to PTR variation across regions, recruitment of teachers, construction of teachers’ houses, procurement of textbooks, dropout rate, low pass rates and improving predictability of resource flows.

Expansion of secondary education will result in increased private sector involvement. Mechanism to distribute projected resources to the private sector has to be worked out. In order to minimize chances of disputes, it is recommended to support the private sector in terms of reward and/or appreciation to the service delivered rather than pre-support system.

Non-formal education has shown some success in terms of enrolment. However, the delivery system needs action in terms of expanded facilities, instruction materials, and recruitment of instructors.

66

Availability of textbooks is the only quality indicator that has been quantified. More quality indicators such as indicators related to inspectorate system, curriculum, attendance of teachers and pupils and school environment are needed.

II: Education Expenditure: In terms of funding, the areas that need improvement are: containing expansion of administration component, improving predictability of (donor) development funding, and increasing government component of development budget. In order to avoid double budgeting for some activities, a mechanism to capture flow of funds to schools should be developed.

Although the Secondary Education Development Plan (SEDP) that is intended to operationalise the Secondary Education Master Plan (SEMP) has been finalized, the role and modalities for private sector participation need to be spelt out clearly and cautiously so as to avoid conflicts and wastage of resources. For example, support of the private sector spelt out in the SEDP could be in form of rewarding good practices rather than advance provision of resources.

III: Financial Management and Accountability: The full implementation of the CAG report will increase effectiveness of funds utilized, enhance accountability and transparency. A system of following up the responses in relation to raised queries should be instituted.

School committees should be granted greater autonomy in executing their budget according to their material needs. Accounting for used money should be reinforced. Cases of misappropriation and other irregularities should be prosecuted with speed in order to pre-empt other would-be offenders. IFMIS should be used to track source and flow of funds to school level.

IV: Cross Cutting Issues: In addition to preventive interventions against HIV/AIDS measures to mitigate their impact, for example, by supporting access to education of orphans and vulnerable children, should be designed. Formation of functional school HIV/AIDS committees for schools that do not have such committees should be encouraged. A disaggregated accounting of cross cutting issues to reflect each item separately is recommended.

Gender related targets such as proportion of girls enrolled, especially at upper secondary schools, drop out rate and pass rate of girls and proportion of girls joining tertiary level of education are essential in order to monitor gender related aspects of education access and achievement at secondary and tertiary level institutions.

A.5 HEALTH

The objectives of the health sector PER were to review the findings and actions taken on the last PER and identify the follow-up actions required, analyze the recurrent and development budget performance for the past three years, analyze the expenditure trends at sectoral and sub-sectoral levels, review the existing plans and strategies for the sector

67

and costing priority interventions over medium term and to compare the financial requirements for meeting PRS targets with the resource availability. The main findings of the PER are summarized below. A summary of key issues and recommendations is also presented.

Review of the findings of the last PER and actions taken

Health sector has been identified as a priority sector in the Tanzania Poverty Reduction Strategy. The key findings and actions taken on the recommendations of the last health sector PER is summarized below.

68

Finding/recommendation Action takenSlow increase in on-budget allocation to the health sector in recent years

A Joint Health Finance Committee (JHFC) comprising of representatives from MOH, PORALG, MOF and Partner representatives have been identified. This committee will analyze in detail the level and type of resources coming into the sector as well as to set future targets, and to lobby the Ministry of Finance12. The PER found that although the members have been identified the Committee has not yet met13.

Downward trend in share of domestic funding to health sector within on-budget expenditure

The JHFC was supposed to take up this issue but has not yet met.

There is a reversal in trend of subventions to local government level as a share of the sectoral total and in health share of overall LGA grants

As above, the JHFC was assigned the responsibility to examine the downward trend of health subventions to Local Government. As noted this Committee has not yet met

Finalization of resource allocation formula as a means of strengthening the equity in the distribution of devolved funds, both from the basket and the block grants for Other Charges

Encouraging shift in spending towards Preventive Services

Emphasis towards Preventive Services has been further reinforced by the development of the medium term Health Sector Strategic Plan (2003-2008) with one critical objective to increasingly allocate funds towards priority areas and programs of the Sector The priority actions included strengthening immunization services, improvement in the availability of drugs and medical supplies, provision of quality health services through the delivery of the essential health package).

Improved capture of external resources through MOF database

Development Partners have been encouraged to provide regular quarterly financial updates to the Treasury for Projects/Programs that do not use the Exchequer System. However, despite this, there still remain significant gaps in the data that is currently available on this Database making it difficult to obtain a clearer picture of off-budget expenditure. The Health Sector, in preparation for the development of the MTEF FY2005, has circulated the latest figures from the database to each development partner for further checking and updating.

Improvement in central level absorption capacity

Although budget performance continues to improve at the Central Level in FY03 particularly in terms of the development budget, there is still room for further improvement.

Significant increase in spending on drugs and supplies from central level

This has been maintained in FY03, particularly through Basket funding. However, problems remain in determining their allocation.

Weak capture of spending at Local Government level

There have been continued efforts in the past financial year to strengthen capacity at the local level, through training of the CHMTs in recording income and expenditure, and wider circulation of the financial and implementation report.

In addition, there has been general government strengthening of the budgeting and financial management system through extension of the GFS coding system to councils, enabling clearer identification and comparison of key sub-items.

Inadequate MOH staff time dedicated to PER process with resulting over-reliance on consultant inputs

Discussion is required between MOH and MOF regarding realistic assignment of roles between MOH officials and external consultant support.

12 Para 14 of Side agreement, Annex x to 2003 JRM Report13 Progress against agreed milestones, as of 18th February 2004

69

The above summary of actions taken shows that although the government has initiated some actions the desired results have not yet been achieved.

Trends in budget performance and expenditure Overall budget performance

An analysis of the budget performance during the period 1998-99 to 2003-04 (Table 1) shows that the actual expenditure fell short of budgeted expenditure in all years except 2000/01. Actual expenditure ranged from 94.2 percent of the budgeted expenditure in 1999/2000 to 99.4 percent in 1998/99. In 2000/01 the total expenditure exceeded the budgeted expenditure (101 percent of the budget expenditure).. In all the years considered the actual off budget expenditures were higher than the budget expenditures. Similarly, the actual development expenditures were significantly less than the budgeted expenditures in all years considered.

Table 1. Health sector expenditure in Tanzania, FY99-FY04 (Billion TSh)

1998/99 1999/2000 2000/01 2001/02 2002/03 2003/04

Budget Actual Budget

Actual Budget

Actual Budget Actual Budget Actual Budget

Recurrent 62.2 62.2(100.0)

67.3 59.3(88.1)

92.0 85.5(92.9)

123.9 117.5(94.8)

154.6 141.8(91.7)

167.9

Domestic funds 62.2 62.2 60.7 58.0 75.5 74.9 100.6 95.9 121.3 109.5 150.7 Foreign funds 6.52 1.36 16.5 10.6 23.3 21.6 32.3 32.3 17.3Development 26.8 17.9

(66.8)21.5 12.0

(55.8)26.8 17.7

(66.0)36.1 23.9

(66.2)40.8 33.9

(83.1)33.1

Domestic funds 2.6 0.9 4.6 2.8 5.1 5.1 5.3 5.0 6.1 6.1 6.6

Foreign funds 24.2 17.0 16.9 9.2 21.7 12.6 30.8 18.8 34.7 27.8 26.4

Total on budget 89.0 80.1(90.0)

88.8 71.4(80.4)

118.8

103.3(87.0)

160.0 141.3(88.3)

195.4 175.6(89.9)

201.0

Total off budget 35.6 43.9(123.3)

52.3 61.5(117.6)

59.4 76.9(129.5)

66.1 80.6(121.9)

49.3 60.8(123.3)

70.7

Domestic funds -- 1.1 -- 1.5 -- 1.9 1.2 -- 1.7 1.7

Foreign funds 35.6 42.8 52.3 60.0 59.4 75.0 66.1 79.4 49.3 59.1 69.0

Total health expenditure

124.6 123.9(99.4)

141.1

132.9(94.2)

178.2

180(101.1)

226.2 221.9(98.1)

244.7 236.4(96.6)

271.7

Health expenditure as % of GDP

-- 2.2 -- 2.1 -- 2.5 -- 2.7 -- 2.6 2.7

Health expenditure as % of total government expenditure

-- -- -- 7.5 -- -- 10.4 9.2

Deflated Health expenditure* 149.6 148.7

157.1 148.0

187.4 189.4 226.2 221.9 233.5 225.6 248.4

Figures in parentheses are actual expenditure as a percentage of budgeted expenditure* Deflated using CPI at 2001 prices

Trends in expenditure

The health expenditures show an increasing trend during the period considered. The nominal health sector spending increased from Tsh. 132.9 billion in 1999/2000 to 236.4 billion in 2002/03 and a budget allocation of Tsh 271.7 billion in 2003/04. In 2001/02 prices, the actual expenditures increased from TSh. 149.6 billions in 1998/99 to

70

TSh. 225.6 billion in 2002/03 at an average annual rate of 11.6 percent. The per capita health sector expenditure at constant prices increased by 2.5 percent only between FY03 and FY04. As a percentage of GDP the total health expenditures increased from 2.1 percent in 1999/2000 to 2.6 percent in 2002/03 and is budgeted to be around 2.7 percent in FY04. Similar trends were observed in health expenditure as a percentage of the total government expenditures, increasing from 7.5 percent in FY00 to 10.4 percent in FY03 to a budgeted 9.2 percent in FY04. Although there have been increases in health expenditures in absolute terms and as percentage of total government expenditures, the percapita expenditure continues to be low, $5.39 in FY03 and it is budgeted to be US$5.24 in FY04. These percapita expenditures are relatively low in relation to increasing needs of this priority sector. The costs of providing health services have also been increasing lately. The PER recommends increased budget allocation to this important sector.

Domestic funds drive most of the recurrent spending while most of the development expenditure came from foreign funding. Basket funding with the foreign funding has increasingly played a significant role in supporting day-to-day operations with in the health sector through both recurrent budget support and grants to local government.

The PER found that there was an increasing trend in health expenditures for preventive and primary level services till FY01, which was reversed between FY02 and FY03 largely due to an increase in administrative expenses of the MOH. The trends in recurrent and development expenditures are discussed below.

Recurrent and development expenditures

The share of recurrent expenditure in total expenditure continues to be high (Table 2), 82 percent of the total expenditure in 2002/03. It is budgeted to be 50 percent of the total in 2003/04. This is largely because of the increasing share of the donor basket and the fact that majority of the basket spending is accounted for through the recurrent budget.

Table 2. Composition of recurrent and development expenditures in total expenditures

1998/99 1999/2000 2000/01 2001/02 2002/03 2003/04

Actual Budget Actual Budget Actual Budget Actual Budget Actual Budget Budget

Recurrent 77.7 74.3 84.9 68.9 83.9 70.3 83.8 74.2 82.0 81.1 50.0Development

22.3 25.7 15.1 31.1 16.1 29.7 16.2 25.8 18.0 18.9 50.0

Spending by activity type: The trend in recurrent spending for administrative, preventive and curative categories shows (Table 3) that the biggest increase between FY02 and FY03 was in expenditures for administration (41.7%) followed by hospital services (13.0%) then preventive services (11.6%). In FY03 preventive services accounted for 45.2 percent of the total recurrent spending compared to 47.8 percent in FY02. Spending on administration increased from 9.5 percent of total recurrent expenditure in FY02 to 13.7 percent in FY03. Similarly the spending on hospital services fell from 42.7 percent

71

of the total recurrent expenditure in FY02 to 41.1 percent in FY03. It may be noted that there has been a slight fall in expenditures on preventive services as percentage of total expenditures in the last fiscal year while the expenditures for administration has increased.

Table 3 Trends in spending by level/categories (Tsh billion)

FY01 FY02 FY03PE OC Total PE OC Total PE OC Total

MOH Admin 3.79 7.95 11.75 NIMR 1.59 0.35 1.95 TFNC 0.66 0.22 0.88 MOH Admin, NIMR and TFNC 2.90 3.69 6.58 3.25 5.24 8.50 6.05 8.52 14.57 HospitalsMuhimbili National Hospital 3.79 1.20 4.99 4.78 1.72 6.51 5.36 2.02 7.38 Muhimbili Orthopaedic Institute 0.54 0.30 0.85 0.56 0.38 0.94 1.53 0.93 2.46 Ocean Road Cancer Institute 0.28 0.23 0.51 0.29 0.43 0.72 0.35 0.48 0.83 Bugando Medical Centre 0.71 0.86 1.57 0.72 0.41 1.13 0.94 1.27 2.21 Kilimanjaro Christian Medical Centre 1.20 0.83 2.03 1.33 0.71 2.04 1.48 0.77 2.25 Referral hospitals, MoH * 1.74 4.06 5.81 2.03 0.32 2.35 0.23 0.87 Regional hospitals 4.43 2.07 6.50 5.14 2.40 7.54 5.69 3.27 8.96 District hospitals 3.53 5.06 8.59 5.22 4.39 9.62 6.54 5.34 11.88 Designated District Hospitals 3.07 1.27 4.33 2.94 2.23 5.16 3.43 1.74 5.17 Voluntary Agencies - Hospital 1.91 0.23 2.14 1.94 0.09 2.03 2.32 2.32 Total hospitals 21.20 16.11 37.31 22.93 13.07 38.03 27.65 16.05 43.70 Preventive/Primary health careMoH preventive services 0.27 3.55 3.82 0.27 6.01 6.28 0.30 6.21 6.51 Regional preventive services 0.13 0.08 0.21 0.25 0.05 0.30 0.15 0.15 0.30 Council preventive 14.11 12.86 26.97 18.50 17.45 35.95 23.40 17.89 41.29 Total Preventive/Primary 14.51 16.49 31.00 19.02 23.51 42.53 23.85 24.25 48.10 Total Health recurrent 74.90 89.06 106.37

On-versus off budget expenditures:

The share of on-budget expenditure continued to increase from 53.7 percent in 1999/2000 to 74.3 percent in 2002/03 and a budgeted 74percent in 2003/04 (Table 4). This has been made possible as a result of the concerted efforts by the MOH and MOF by improving information on external funding, capturing a higher proportion of the total within budgetary estimates and through improvements in the planning and budgeting process, as the MTEF becomes a more familiar and applied tool.

Table 4. On-budget and off-budget health expenditure

1998/99 1999/2000 2000/01 2001/02 2002/03 2003/04

Budget Actual Budget Actual Budget Actual Budget Actual Budget Actual Budget

Total on budget 89.0(71.4)

80.1(64.6)

88.8(62.9)

71.4(53.7)

118.8(66.7)

103.3(57.4)

160.0(70.7)

141.3(63.7)

195.4(79.9)

175.6(74.3)

201.0(74.0)

Total off budget 35.6(28.6)

43.9(35.4)

52.3(37.1)

61.5(46.3)

59.4(33.3)

76.9(42.7)

66.1(29.2)

80.6(36.3)

49.3(20.1)

60.8(25.7)

70.7(26.0)

Total health expenditure

124.6(100)

123.9(100)

141.1(100)

132.9(100)

178.2(100)

180(100)

226.2(100)

221.9(100)

244.7(100)

236.4(100)

271.7(100)

Analysis of spending

Recurrent spending: Recurrent expenditures both from domestic and foreign funding sources continue to increase during the period. Total recurrent expenditures increased by

72

17.1 percent in 2002/03. While recurrent expenditures from domestic funding increased by 12.4 percent the increase in foreign funded recurrent expenditures were 33.1 percent.

Most of the recurrent spending in the health sector has been for drugs and supplies, and on HIV/AIDS. Between FY 02 and FY03 the expenditure on drugs and supplies increased by 21 percent. A further 20 percent increase has been budgeted for FY04. The hospital services department continues to be the largest consumer of drugs and supplies. Incomplete estimates of per capita expenditure on drugs and medical supplies shows that it increased from $0.64 in FY02 to $0.69 in FY03 and to a budgeted $0.77 in FY04. The PER found that there are substantial variations in per capita allocation of drugs and medical supply to the different regions.

Spending on HIV/AIDS spans over different sectors and ministries with funding from government and basket funding. Since different ministries and sectors are involved, it has been difficult to get a correct estimate of the total. At the central level, the total MOH expenditures increased from Tsh. 1.52 in FY02 to Tsh. 7.29 billion in FY03. It is budgeted to be Tsh. 8.63 billion in FY04, a 4 percent increase from FY03 level. At the local government level all LA s have assigned funds for HIV/AIDS in their budget, although at a low level. HIV/AIDS spending at the LA level as a percentage of the total health OC spending increased from 1.3 percent in FY02 to 2.6 percent in FY04. Though there has been an increase in HIV/AIDS spending at all levels it may however be noted that such increases are small in relation to the extent of HIV/AIDS problem in the country.

Development expenditures: Development projects in the health sector have been implemented at the central, regional and council levels. At the central government level the MOH continued to implement reforms with the objective of improving efficiency. The Health Plans and Management program is one such program implemented in FY03. Other development projects at the central level focused on more traditional development activities such as construction, rehabilitation and purchase of equipments. Development projects at the regional level are funded by both donors and the government, the former accounting for about 80 percent of the actual expenditures in FY03. This pattern is expected to continue in FY04 also. At the council level majority of the development projects were to improve the infrastructure through capital investment in health centers, dispensaries and district hospitals

Local government budget and spending in relation to PRS objectives

According to the PRS local government spending in priority sectors represents a major component of the pro-poor activities. There is evidence of continued government commitment to devolution of responsibilities to local authorities (Table 5).

73

Table 5. Government subventions to LGAs, FY01-FY04

Information on releases in relation to the budget provided by MOF indicates that 100 percent of all planned OC block grants for health was released to each council by the end of the financial year in FY03. The PER did not analyze the geographical variation in allocation and expenditures in the health sector at the local government level

Fiscal decentralization and allocation

In order to achieve equitable allocation of local government funding across geographical areas the MOH has approved a new formula for fiscal decentralization and allocation in the health sector. This will be applied to both block grants and council basket funding from July 2004. This formula will use population, mileage traveled within the region, poverty level, and under-five mortality with different weights to determine the funding.

Health sector performance

The national health goal of the PRS is to first arrest the decline in life expectancy and then raise it to 52 years by the year 2010. The health sector has also set specific targets in terms of infant mortality rate, under-5 mortality, maternal mortality, malaria mortality, and access to safe water, to be achieved in the coming years. There is no data available to measure the progress in achieving these specific targets. Data presented in Table 1 show significant increase in health sector expenditure in nominal terms and in constant prices in the last three years.

Spending on priority items

In FY 04 the Ministry of Finance has made distinction between expenditure on items falling within the priority items in the health sector and spending on HIV/AIDS. This has created a break in the series between FY03 and FY04 in reporting health sector budget and expenditures.Recurrent spending on PRS priority items increased from Tsh. 52.4 billion in FY02 to Tsh. 63.3 billion in FY03 and is budgeted to be Tsh. 79.4 billion in FY04 (Table 6)

74

FY01 Approved estimates FY02 Approved estimates FY03 Estimates FY04 EstimatesPE OC Total PE OC Total PE OC Total PE OC Total

Urban 4.47 1.57 6.04 4.90 2.16 7.06 5.71 2.53 8.23 6.59 3.29 9.89 District 16.58 6.47 23.05 18.83 9.59 28.42 21.41 11.07 32.48 25.00 13.51 38.51 Total 21.05 8.04 29.09 23.73 11.75 35.48 27.12 13.59 40.71 31.59 16.80 48.40

Table 6. Recurrent spending on PRS priority items, FY02 to FY04 (Tsh. Million)

Spending on priority sector as percentage of overall sector spending is around 45 percent. Although the above figures do not account for funds from health basket funds and value of directly procured drugs, it may be noted that more than 50 percent of the total recurrent spending is on priority items.

In the case of development spending, PRS priority projects accounts for only about two thirds of the development spending. GOT funding although less than 13 percent of the total development budget, is primarily targeted at non-PRS priority projects. Similarly, the foreign basket development funding allocated to PRS priorities is very low (25 percent of the total). However, about 94 percent of the foreign non-basket development funding supported priority projects.

Financing performance

The per capita allocations and spending in the health sector at the central, regional and district levels have been showing improvement since FY01 (Table 7). It increased by

Table 7. Selected indicators of financial performance (in US $)

Level Baseline FY01

FY 03 FY04 BudgetBudget Actual

Total GOT public allocation to health per capita Central 1.49 2.12 1.77 2.58Regional 0.21 0.24 0.24 0.26District 1.02 1.31 1.33 1.37Total 2.72 3.67 3.34 4.21

GOT and donor allocation (budget & off-budget) to health sector per capita

National average

6.12 7.28 7.04 7.26

Per capita GOT recurrent expenditure by levels of government

Central 0.23 0.42Regional 1.29 1.26District 1.07 1.39Total 2.59 3.07

22.8 percent from $2.7 in FY01 to $3.3 in FY03 and is budgeted to be around $4.21 in FY04 (Table 7). The increase has been higher at the central and at the district levels. Total health sector spending percapita, budget and off budget, from government and donors increased from $6.1 in FY 01 to $7.0 in FY03 and is budgeted to be at $7.3. The percapita GOT recurrent expenditures increased from $2.6 in FY01 to $3.1 in FY03. Recurrent spending at the regional and district levels are significantly higher than that at

75

FY02 FY03 FY04Total subvention to LGAs 35,393 43,250 48,856 Preventive service subvote at RAS 302 304 310 LGA drugs budgeted under MOH 9,108 12,478 15,000 MOH HQ Preventive services subvote 7,574 7,253 15,187 Total Health Priority items 52,376 63,285 79,353 Total Priority sector spend/budget 117,473 141,746 167,966 Priority items as % sectoral spend 45% 45% 47%

the central level, which shows the government to a large extent provided health services at the local government levels.

The analysis of opportunities for cost sharing of revenues show that the revenues generated could make a substantial contribution to daily running expenses with in a given facility with revenues reaching as much as 50 percent of the expenditures on drugs. The PER further found that there is substantial variation in both revenue generation and absorption of such revenues between hospitals. The PER however suggested that further analysis might be needed to ensure that revenue generation objective does not affect the accessibility of the poor to the health services.

Future costs and revenues

Costing for the health sector is done based on a costing study done in 1998 in the MOH prior to the PER exercise. The MOH is using a figure of $9 per capita as a target for spending in the health sector. The PER found that this cost estimate and the baseline has to revalued as the country is facing increasing costs from new first line treatment of malaria, introduction of new vaccines, expansion of services such as VCT and TB prophylaxis and introduction of ARV treatment for AIDS. Poor costing of interventions required continues to be a weakness of the PER.

Future sources of revenues are government funding including GBS, basket funding and external project funding. Based on the budget guidelines issued by the government the total GOT real allocation to the sector is projected to increase by 24 percent over the three-year period from FY04 to FY07. This increase is entirely due to a large (about 218%) increase in foreign non-basket funding to the sector in FY05. But the total local funding is expected to fall by 3 percent during the above period.

There has been significant increase in the basket funding for the health sector in FY05 due in large part to the contribution from the World Bank under the new Health Sector Development Program. The PER expects significant increase in flow of external funding through donor project and basket funding in the health sector. Such funds are expected for the Global Fund for HIV/AIDS, tuberculosis and malaria control, and the HIV/AIDS Care and Treatment plan with the assistance of the Clinton Foundation.

Key issues and recommendations

The following key issues and recommendations emerge from the main sector report: Although there is an increasing trend in health sector budget and expenditures their real growths are much less than the demands due to the increasing health problems and higher costs of delivering a basic package of services to the population. There is an apparent decline in the share of health sector expenditure in the total government expenditure in the FY04 budget allocations. There is a need for the JHFC to take up this issue and to secure better budget allocation to the health sector.

76

The analysis of spending by level and by category (preventive versus curative) could be strengthened by inclusion of basket funding and those elements of the development budget, which contribute to recurrent spending. Further work is needed to disaggregate central spending on behalf of LGAs, but would be useful in terms of obtaining a more accurate measure of resource availability at that level.

Currently there is no analysis of actual expenditures in relation to budgets. The MTEF is still incomplete in terms of external support to these areas. It is therefore recommended that the FY05 update include analysis of spending in at least two such programmatic areas, to be defined in advance. This should include both a central and LGA analysis.

Analysis of budget performance and absorption capacity is incomplete. IFMIS permits comparison both of releases against budget, and of expenditure against release, thereby enabling comparison both of performance by GOT (and partners) in meeting their stated budgetary commitments, and of the spending agency in using resources. However, currently this is undertaken for MOH HQ only due to lack of detailed expenditure data at the Regional and LGA levels. For the past three years actual expenditures have been consistently less than budgets. Although the nominal level of funds allocated to the local government level has increased between FY03 and FY04 the budgeted health sector share of LGA funds has been falling in recent years, from an initial baseline of 18% to 16.6% in FY04. It is not possible to say whether this is due to the presence of the council basket funds. The allocation between urban and rural councils raises some queries, as the growth of both PEs and OC allocations is higher for urban councils. The PER FY05 update may analyze the geographical distribution of basket and/or block grant funding between Regions using the FY03 data as a baseline. IFMIS may be extended to the council level to obtain actual expenditure data from LGAs. Thi s may be undertaken prior to the FY05 PER update process in order for the results to be used, and should cover both PE and OC funding

This PER update indicates that the although the contribution of cost sharing to the overall sectoral resource envelope is limited, the existing cost-sharing schemes are creating barriers to access for the poorest contrary to the objectives of overall development policy as articulated in the PRSP. There is little clarity on sources and uses of funds under the cost-sharing program. The FY03 PER update had raised the issue of lack of reporting in this area, but little action appears to have been taken. The PER further recommends that there should be better information on the Community Health Fund. There are no incentives in the NHIF to ensure cost-effective or equitable use of heath services, and therefore the extent to which it might ultimately be considered pro-poor is questionable. Although there has been increase in total on-budget spending on drugs and supplies at MOH HQ the available data do not show the distribution of these funds, both by level of the health system and by geographic regions. The whole area of spending on drugs and supplies within the sector in the framework of the MTEF merits a more detailed study. The PER further recommended a systematic study on supply of drugs and services at the sectoral and at various levels of the government. Such a study should:

77

quantify the total drugs and supplies with in the sector including purchases by projects and programs, analyze the centrally procured items and their share of regional and local governments and analyze the geographical allocation of drugs and supplies in order to comment on the existing formula and justify any difference with that used for block grants and basket funding for councils.

The PER found that there is inconsistency in budget and expenditure information collected through the sectoral PER and MOF data. The changing definitions of priority sectors and items complicate calculations of time series data. For the revision of the PRS, it is recommended that definitions of priority sectors and items be clearly stated, together with targeted allocations, in order to give a basis for comparison in future PER updates.

Problems with the IFMIS within the MOH were a particular problem for this PER update, both delaying proceedings and necessitating repeat analysis of the same items. Such problems in accounting database affect the future inflow of funds. Unavailability of skilled staff, multiplicity of formats are other problems in the management of the database.

Immediate steps

The PER recommended the following immediate steps:a. In order to take up the issue of health sector financing and budget allocations, the

Joint Health Finance Committee agreed in 2003 need to reactivated.b. The preparation of the FY05 PER should begin as early as possiblec. The PER coordinator should meet with all stake holders and formats should be

prepared for quarterly reports so that annual reporting will then become less burdensome.

HIV/AIDS Expenditure Review

The objectives of the multi-sector public expenditure review conducted in 2003 are to review the major recommendations and actions taken on the last PER, review the public expenditures on HIV/AIDS, expenditures by sectors and local governments, budget performance and the proposed spending pattern and to make recommendations for improved performance of multi-sectoral approach to control HIV/AIDS.

HIV/AIDS has very severe impact on the economy and human development in Tanzania, through its impact on productivity, increasing dependency ratio and increasing costs for the care of infected individuals. The country has responded to the HIV/AIDS challenge through the formulation of the Tanzania Commission for AIDS (TACIDS) and a National Multi Sector Strategic Framework (NMSF) to operationalize the national policy on HIV/AIDS. Since HIV/AIDS affect all sectors, interventions in to the work plans of all sectors and ministries with object of a rapid acceleration of preventive programs and mitigation of its impact have been formulated and implemented.

78

Review of recommendations and actions taken

The main recommendations of the last PER and actions taken on those recommendations are summarized below:

1. HIV/AIDS should be coded at activity level in the IFMS budget coding structure- this recommendation has been accepted for 2004/05 budget.

2. Budget guidelines should provide more detailed guidance on planning and budgeting for HIV/AIDS programs- this has been proposed in the 2004/05 budget guidelines

3. TACIDS should strengthen financial staff and capacity and develop stronger ties to MOF and PORLAG- number of accounting staff has been increased and a procurement specialist is being recruited.

Expenditure on HIV/AIDS

There is no internationally accepted definition as to what constitutes public expenditure on HIV/AIDS, as it has not been possible for practical reasons to separate out HIV/AIDS element of public expenditures. However, spending programs that are targeted to HIV/AIDS interventions are considered in this report as public expenditure on HIV/AIDS. There is no good accounting of expenditures targeted to HIV/AIDS in Tanzania and hence data on expenditures are incomplete. The best estimates of the expenditures on HIV/AIDS for 2001-02 and 2001-03 are presented in Table 1. The total expenditure on HIV/AIDS increased by 89 percent from Tsh. 16.8 billion in 2001-02 to 31.7 billion in 2002-03. All the individual line items show significant increase from 2001/02 to 2002/03. While the recurrent government spending increased by Tsh 4.75 billion (207%) the aid flows to public sector increased from Tsh.9.9 billion to Tsh. 16.4 billion (65%) during the above period. The total aid flows to NGOs increased by 82 percent from Tsh. 4.5 billion to Tsh. 8.3 billion. The significant increase in total public sector spending (91.6%) and government recurring spending for HIV/AIDS show its commitment to the HIV/AIDS problem in the country. There has also been significant increase in aid flows.

HIV/AIDS expenditure plans

The existing and the proposed public expenditure plan for HIV/AIDS over the medium term are presented in Table 2. The budget envisaged an increase in expenditure from Tsh. 11.6 billion in 2002-03 to Tsh. 42.6 billion in 2003-04. Although this is a 267 percent increase, the outlay for 2003/04 accounts for only about 40 percent of the requirements based on international evidence. The total resources in the two subsequent

79

Table 1 Expenditure on HIV/AIDS*

2001-02 2002-03Recurrent, Government 2296 7050 (207.1)Development, Government 0 0Aid via Govt. systems NA 4460Total spending, as per Accountant general 11510Aid direct to public sector recipients 11950Total ODA to public Sector 9948 16410 (65.0)Total public sector expenditure 12244 23460 (91.6)ODA to NGOs 4546 8267 (81.9)Total expenditure 16789 31727 (89.0)The data presented in the table are incomplete as it is difficult to identify expenditures specific for HIV/AIDS Figures in parentheses are percentage increase from the 2001/02

years are higher than the budget guidelines, total resources are about 183 percent and 161percent of the budget guidelines in 2004/05 and 2005/06 respectively. It may however be noted that the actual donor commitments usually fall short of the actual spending and hence this should be treated with skepticism. The shares of different sources funds for HIV/AIDS activities show that external finance accounts for more than 75 percent of the total resources. Funding from government sources account for about 20 to 25 percent of the total resources. A significant share of the external financing goes to NGOs.

Table 2: Estimates of available and required HIV/AIDS expenditures, 2002/03-2005/06

2002/03 2003/04 2004/05 2005/06Estimates of required resources 2003 Budget guidelines 11.6 (36.6) 42.6 (72.2) 42.9 (54.5) 43.9 (61.9)Estimates of resources available Government resources 7.0 (22.1) 14.5 (24.6) 16.3 (20.7) 16.6 (23.4) External finance 24.7 (77.9) 44.5 (75.4) 62.4 (79.3) 54.3 (76.6) Budget aid 4.5 (14.2) 8.6 (14.6) 30.1 (38.2) 44.2 (62.3) Other aid to Government 11.9 (37.5) 18.9 (32.0) 16.4 (20.8) 6.5 (9.2) Total to public sector 16.4 (51.7) 27.5 (46.6) 46.5 (59.1) 50.7 (71.5) Aid to NGO 8.3 (26.2) 17.0 (28.8) 15.6 (19.8) 3.6 (5.1)Total resources 31.7 (100) 59.0 (100) 78.7 (100) 70.9 (100)Total resources as % or budget guidelines 138.5 183.4 161.5Figures in parentheses are percentages of total available resources

The figures reported tend to understate the actual level of resources devoted to HIV since they cover only specific HIV/AIDS spending programs. They do not capture the share of Peps accounted for by staff spending part of their time on HIV/AIDS, costs of HIV/AIDS patients being cared for in medical facilities, reproductive health costs that could be attributed to HIV/AIDS prevention and costs of such joint programs as TB/HIVAIDS etc.

80

The UNAIDS estimates that Tanzania need s to spend $90 million in calendar year 2006 and about $100 million in the subsequent years to control and prevent the spread of HIV in the country. Even if the increased donor commitments are fully realized the government would need to nearly triple its domestic funding in order to reach spending levels envisaged by UNAIDS. The government would also have to mobilize increased funding from donors and ensure that more of their funding supports the NMSF priorities and more of it is disbursed through the budget in ways to enable the government to avoid unbalanced spending patterns that result from uncoordinated project commitments.

Role of NGOs

Though the MOF accounts include official donor commitments to NGOs, they do not fully capture the spending by NGOs. As Table 2 indicates there is a doubling of donor commitments to NGOs in 2003/04. The flow of funds to NGOs is likely to increase to $20-25 million per year and such increase may require substantial increase in absorptive capacity. Provisions for capacity building are incorporated in the US programs and in the World Bank supported TMAP Community Fund. The Clinton Foundation Care and Treatment Plan is a major initiative for HIV/AIDS management prepared with the support from the Clinton Foundation. The cabinet has approved this program with a total cost of $539 million over the first five years. A large part of the resources under this project is additional although the government is expected to incur significant expenditure in the longer term.

Sectoral analysis of actual and planned HIV/AIDS spending

The sector strategies are compared with the budget available for HIV/AIDS spending in this section. The individual sectors are discussed below:

Health

Health sector has the most fully developed and costed plan for HIV/AIDS. The resources required for the strategy are compared in Table 3 with the available resources based on the MTEF projections and the information on donor commitments. Even with government and donor commitments, Table 3 shows that there are substantial financing gaps through the MTEF period. The figures show that the total commitments will finance only about 40 percent of the requirements. Even with the continued growth in donor inflows the funding strategy will require a substantial increase in government commitments to HIV/AIDS programs in the health sector.

Education

The education sector strategy for HIV/AIDS that is currently being finalized will estimate the cost implications of reaching full coverage of preventive actions for the youth in school. Such a program targeting the youth would result in an estimated per capita cost of $2.3. Assuming that pupil enrolment in these grades to grow to 2.5 million, and also that some interventions are required at the secondary level, the required budget would grow to Tsh. 8.4 billion.

81

Table 3: Financing the Tanzania Health Sector HIV Strategic Plan14 (Shillings, Billions)

2003-4 2004-5 2005-6 2006-7 TotalPreventionCost 11.3 22.4 30.4 31.4 95.3MTEF, from Govt 1.6 2.9 2.4 6.9MTEF, from Donors 3.6 5.7 3.4 12.7MTEF, Total 5.2 8.6 5.8 0 19.6Donor Commitments 6.2 5.9 0.2 0.0 12.4Shortfalls (Surplus) 3.4 13.5 27.7 31.4 76.1Care Cost 16.5 32.6 38.5 51.7 139.3MTEF, from Govt 5.3 5.5 5.9 16.7MTEF, from Donors 0 0.1 0.1 0.2MTEF, Total 5.3 5.6 5.9 16.8Donor Commitments 7.0 9.8 24.9 41.6Shortfalls (Surplus) 11.2 20.1 22.8 26.9 81.0Cross-Cutting Cost 1.2 2.4 2.8 3.2 9.6MTEF, from Govt 2.1 1.7 1.7 5.5MTEF, from Donors 0.7 2 0.1 2.8MTEF, Total 2.9 3.6 1.9 8.4Donor Commitments 3.9 3.3 3.9 0.0 11.2Shortfalls (Surplus) -4.9 -2.6 -2.8 3.2 -7.0

Total Strategy Cost 28.9 57.4 71.7 86.3 244.3MTEF, from Govt 9 9.9 9.9 28.8MTEF, from Donors 4.2 7.8 3.7 15.7MTEF, Total 13.3 17.7 13.6 44.6Donor Commitments 10.1 16.2 13.9 24.9 65.2Shortfalls (Surplus) 9.8 31.2 47.9 61.5 150.3

Sources: Health sector strategy; health sector MTEF; analysis of donor commitments to MOH based on MOF database plus analysis of donor documents. See Annex 2 for assumptions.

TACIDS

The current MTEF envisages TACIDS budget to fall in 2004/05 and 2005/06 (Table 4), which is not consistent with the expanded leadership role assigned to TACIDS. The present budget do not allow for increased growth in non-personal related OCs. TACIDS will also require an increase in spending on advocacy materials, monitoring and supervision and on measures to enable TACIDS to be an effective “one stop shop” for advice and information about HV/AIDS in the country. TACIDS will also manage the World Bank TMAP funding and the Clinton Foundation Care and Treatment Plan.

14 Strategy figures are for calendar years and are US dollar figures at 2002 prices converted at exchange rates used in the MOF external finance department database.

82

Table 4: TACAIDS Budget: 2003-4 to 2006-7, (Tsh. ’000)

2003-4 2004-5 2005-6 2006-7PE & Staff benefits 444 450 450 450OCs, Other 3851 3240 3511 3511Govt Development 287 1515 1515 1515Total existing 4582 4755 5026 5026IDA 1169 2710 5273 5363Budget Ceiling 5751 7465 10299 10389Other donor flows 2026 4340 660Total 7777 11805 10959 10389

Local Government Authorities

There is no accurate data on actual government expenditure on HIV/AIDS at the district level. The 2003/04 budget for subventions to local government authorities for HIV/AIDS was just Tsh. 433 million which is equivalent to Tsh. 3.6 million per LGA. Most LGAs lack planning capacity and resources to develop multisectoral plans for combating HIV/AIDS at the district level. PORALG which is mandated to oversee development of local authorities and therefore indirectly the role of communities in fighting HIV/AIDS has not itself prepared a strategic plan for HIV/AIDS.

The donors through NGOs and CBOs fund most of the HIV/AIDS activities that are taking place in the communities. Expected disbursements by donors for HIV/AIDS are expected to reach $16 million in 2003/04 compared to $5 million in 2002/03. In addition to this, an additional $87 million from the Global Fund and the IDA will also flow to districts and communities. The Global Fund will address requirements for care and support within 45 districts focusing on VCT for both HIV/AIDS and TB and on providing comprehensive care and support to people living with HIV/AIDS. Other programs that are being implemented at the LGA level are the TMAP project, Community AIDS Response Fund that will provide resources mainly through NGOs, CBOs and the CARF that will support a full spectrum of AIDS activities. Assuming a 75 percent disbursement of the commitments, the total resource flows to districts and communities from the Global Fund, IDA and support via NGOs might reach around $26 million per annum. This will work out to roughly $216 thousand per district or $0.75 per head. The PER recommended substantial improvement in the capacity of the local government and community-based organizations to absorb these significantly increased resource flows.

The PER further raised the following two main concerns in the distribution of funds at the LGA levels:

1. The distribution of funds is likely to be highly skewed towards more accessible districts and those districts that have more active NGOs and CBOs.

2. There is also the likelihood of a skewed distribution of funds towards care and treatment at the expense of prevention.

83

In order to ensure better distribution of resources, it is recommended that funds from such sources as IDA be targeted to less served districts and towards prevention rather than treatment and care.

Budget process for HIV/AIDS

Budget guidelines

The PER has identified the following problems while reviewing the current MTEF for HIV/AIDS:

1. The MDAs fear that allocating money to HIV/AIDS will squeeze their priority programs when there are TMAP resources that can be accessed outside the national budget process. This has affected the allocation for HIV/AIDS.

2. The MTEF is not used at present to make strategic choices about future allocation of resources. This problem is manifested in HIV/AIDS programs with no growth in allocation, but rather a repetition of the activities and allocations in the current year.

3. TACAIDS have not yet established for themselves a formal role in supporting the preparation of MDA action plans and budgets, nor in supporting MOF to review sector MTEFs for their HIV/AIDS treatment

Role of TACIDS in budget process

The PER found that TACIDS has neither played a strategic leadership role with respect to the treatment of HIV/AIDS in the national budget nor the internal TACIDS budget has been used to plan a medium-term strategy. The PER recommends two roles for TACIDS in the broader budget process:

1. To provide technical support to MDAs in developing strategies and medium term expenditure frameworks

2. To support the Ministry of Finance in providing guidance on how to mainstream HIV/AIDS within the budget, reviewing the proposed MDA action plans and budgets and briefing MOF on HIV/AIDS aspects for their meetings to discuss the budget proposals coming from MDAs.

In order perform these roles TACAIDS will need to strengthen their relationships with the budget division of MOF, and with the heads of planning and budgeting within individual MDAs. This could be initiated by arranging a series of meetings between TACIDS and their technical advisors together with the budget department of MOF and the budget and HIV/AIDS officers of relevant MDAs.

Accounting and data management on HIV/AIDS

At present it is very difficult to track the planned and actual expenditures on HIV/AIDS. As a result some of the expenditures are not accounted for in the MOF database system. The PER recommends the following actions for better accounting of the expenditures:

TACIDS should ensure that they are copied on the relevant sections of MTEFs and action plans prepared by MDAs and by Districts, so that the data

84

can be readily retrieved. The Policy and Planning Department of TACAIDS should support and monitor HIV/AIDS aspects of the MTEF and budget process across Government.

There should be routine data reporting on all aspects of HIV/AIDS and TACIDS may coordinate this.

At present it is very difficult to track planned and actual expenditures on HIV/AIDS. Although Tanzania has an integrated financial management system the way in which some of the codes are operated makes it difficult to consolidate reports on cross cutting issues like HIV/AIDS. The PER has recommended an alternative coding structure that is more objective in tracking funding and expenditures on various HIV/AIDS activities.

85

ANNEX 2CONSISTENCY BETWEEN POLICY, PLANNING, AND BUDGETING

A.1 INTRODUCTION

Good and pragmatic expenditure projections with reliable budget execution are essential for successful expenditure management. Leaving expenditure policy aside, the level of consistency between the projected resource envelope, the detailed sectoral programmed budget, and the expenditure out-turn, is a major indicator of the maturity and the precision of fiscal targeting and programming, policy coordination and coherence between various actors in the budget process, and the capacities of the executing agencies/ministries. This note explores the extent this consistency was achieved in Tanzania as manifested by the available expenditure data with more focus on the programming side.

This note is organized as follows: Section 2 provides some background of the recent evolution of budget planning, programming and execution in Tanzania. Section 3 introduces the budget statistics series on which the analysis is based. Section 4, discusses the consistency in major statistical aggregates in public expenditure programming and execution. Section 5 repeats the same discussion for priority sectors. Finally Section 6 provides concluding statements.

A.2 THE EVOLUTION OF PLANNING, PROGRAMMING, AND EXECUTION

Public expenditure management is a delicate process both from the programming as well as from the execution part of it. If predictions and programming are weak, one can expect major uncertainties, unanticipated events and disruptions during execution that will necessitate sizeable adjustments throughout the fiscal year, which could undermine delivery of services, project implementation, and fiscal objectives and balance. On the other hand, if the projections and the construction of the programming are reasonable but the out-turn has little resemblance with the approved budget, it indicates serious problems at the execution stage. In both cases, budget predictability, transparency, and effectiveness, suffer and stake holders, including creditors, and particularly important in the case of Tanzania, donors, may loose confidence in the expenditure management altogether. The consequences could be far reaching and negative to credibility of public institutions, fiscal and macroeconomic stability, and growth.

According to FY97 PER and other Bank documents, in the mid 1990s Tanzania demonstrated symptoms of poor expenditure planning, programming and execution. Those symptoms included:

Over optimistic expenditure projections;

Adherence to the resource envelope was poor and the resource allocations were frequently overridden during subsequent preparation of the Budget.

86

Government commitment to the budget was weak as there were frequent reallocations early during the execution;

Budget allocations lacked a policy framework to prioritize expenditure,

There was lack of a sense of ownership in the budget preparation by the line agencies;

Most of the programming was focused on development expenditure which constituted a small proportion of the total expenditure;

At the execution end, weak monitoring and accounting added to the uncertainties; and Donors were also uncertain of the formal budget process and a high proportion of their assistance bypassed the formal budget.

In the mid 1990s, the Government of Tanzania committed itself to improving expenditure management. Since then, public expenditure management has been in a state of transition and improvement. Many of these weaknesses were successfully addressed. Some of the achievements included:

Integrating the institutional responsibility for both recurrent and development expenditure programming under the Ministry of Finance.

Improving the accuracy of forecasting, particularly of revenues and external financing. The capacity of the government officials for macroeconomic and fiscal forecasting is better, particularly with the use of a macro model (MACMOD).

The introduction of MTEF improved the linkages between planning, budgeting, and execution.

The PER process allowed sector policies and priorities and their detailed program costing to be fed into the MTEF. It also allowed stakeholder involvement and participation. Because line ministries are represented in various committees in the budget process and responsible for forwarding their forward budgets based on the ceilings in the Budget Guidelines, the line agencies and ministries’ inputs and ownership of the process have significantly increased.

The PRSP became the basis for prioritizing the allocation of public resources.

The introduction of the IFMS has permitted detailed target-setting, programming, and monitoring execution; brought discipline, transparency, and accountability, to the budget process; simplified controls over releases to all spending units (use of IFMS in districts is still in a pilot stage); made easier to monitor progress by having reliable and timely information and quarterly reports; and allowed financial performance to be fed back into the next budget cycle and the MTEF.

Though some problems remain or persist, the accomplishments made so far fostered a greater certainty, participation, and confidence in the budget process. They allowed greater integration between sectors, national resource envelope, and the macro policies. The budget process has earned its credibility.

87

A.3 DEFINING THE BUDGET DATA

To explore the extent to which the improvements in programming and execution are reflected in the budget data, we analyze four sets of budget data in the subsequent sections. They are the Budget Guidelines, cross-sector MTEF, Annual Budget, and Actual Expenditures. The comparability of the data sets somewhat relate to the stage in the budget cycle they were generated. For the budget preparation, Table 20 below illustrates the stages in which the first three data sets are generated. The Budget cycle starts annually in September/October, programming ends in June when the budget is submitted to the parliament, and implementation ends in June of the following year.

Table 20: The Budget Preparation Process

Sept-Oct Nov-Dec Jan-Feb Mar-April May-June

(a) Budget Guidelines Preparation

(b) The MTEF processes

(c) The Approval of the Annual Budget

Briefly, the four sets of data can be described as follows:

(a) The Budget Guidelines (Issued in January). It is a 3-year rolling medium-term expenditure plan put together by the Budget Guidelines Committee. The overall resource envelope of the three-year projection is developed with inputs from the Ministry of Finance, the Central Bank, Civil Service Department, line ministries, and Office of the President (Planning and Privatization). The PER working group assists in the process by providing projections of donor finance as an input to the budget guidelines. The most important determinants of the level of the overall resource envelope are the domestic revenue and the foreign financing projections. The Budget Guidelines Committee simply agrees upon budgetary ceilings by sector and vote after consultations with concerned parties, and submits the proposed budget ceilings to the cabinet for approval by November/December. After its subsequent approval, the sectors and agencies prepare their detailed forward budget proposals and MTEF projections. The sector ceilings are generally binding and respected except for the updates of the resource envelope and for possible changes in external commitments.

(b) The MTEF (Concludes in April/May). In Tanzania, the MTEF comprises the macro-fiscal framework, the cross-sectoral MTEF, as well as detailed sectoral MTEFs for the priority sectors which incorporate sector policies and PRS priorities into the detailed sector programming and costing. In principle, the allocations and the resource envelope, are those specified in the 3-year rolling Budget Guidelines. Differences to the budget guidelines ceilings arise primarily from updates with regard to foreign funded projects/programs.

The first three MTEFs (FY99, FY00, and FY01) focused only on recurrent expenditure, reflecting the fact that the development budget is almost entirely financed by donors in

88

processes that are largely outside the regular budget process. Following the introduction of the PRS and the provision of direct budget support by a large number of donors, the last two MTEFs (FY02 and FY03) were comprehensive and included development expenditure. It is expected that, eventually, all sectors and agencies of government will join the MTEF exercise.

(c) The Annual Budget (June). The National Budget Committee reviews all submissions and submits to the Cabinet. Ministry of Finance incorporates changes, on i.e., outstanding requests from sectors and agencies on payroll, in to the final budget and submits it to the parliament for debate and approval by mid-June.

(d) Actual Expenditure. Actual expenditure refers to the out-turn following the execution of the annual budget as reported in the appropriation accounts and audited by the National Audit Office. Tanzania continues to apply strict cash budgeting. Deviations between budget estimates and actual expenditures can come from: (i) unanticipated shortfalls in resources; (ii) reallocation of funds among the sectors occurs mainly to distribute new funds and/or the retained funds at the MoF for contingencies; (iii) under spending by vote-holders as a consequence of low rate of budget releases or implementation; and (iv) unanticipated changes in the flow of external assistance. Most of the deviations are in the development budget.

The budget data reflects the accounts of the Central Government. It excludes local government revenues and the expenditures these revenues finance. Unfortunately, it also excludes an undetermined number of foreign financed projects that bypass the formal budget system. In the latter, though in principle, they should be included in the budget, accounting for all of them remains to be a challenge.

A.4 OVERALL BUDGET TRENDS

Data Sources and Definitions

The Budget Guidelines, Annual Budget, and Actual Out-turn data presented in this section are obtained solely from the annual documents of the Budget Guidelines. There were obtained from the ‘Budget Frame (Accounting)’ table. The reason that the ‘Budget Frame (Accounting)’ is that it was consistently provided in all Budget Guideline documents. Several characteristics of the data in the Table are worth noting. First, because the classification is not the standard GFS/IMF format, and the total recurrent expenditure includes amortization and payment of arrears. Second, besides PE and OC, there are other categories in the composition of recurrent expenditure, including: (i) ‘designated items’ which include items such as the Road Fund, Parastatal Wages, TRA and Retention Scheme; (ii) CRS which includes debt service; and (iii) payment of arrears. Third, in some years, total expenditure is higher than the sum of recurrent and development expenditures. The difference is ‘contingency funds.’

89

Within-Year Consistency of Aggregate Expenditure Estimates (Budget Guidelines – MTEF – Budget – Actual Expenditures)

The primary role of the budget guidelines is to define expenditure priorities for the next fiscal year, project the overall resource envelope, and set expenditure ceilings for key expenditure categories such as wages and salaries or debt service as well as sectoral and ministerial expenditure ceilings. Based on this information, spending units prepare their detailed budget submissions and MTEFs.

Aggregate Resource and Expenditure Estimates

Projections of resource projections play a key role in Tanzania’s budget process, as they determine the level of expenditure. The total resource envelope includes domestic revenue as well as foreign program and project finance. Other resource flows such as domestic borrowing, or privatization proceeds are also included in the resource envelope, and are typically small compared to the other resource types. During the past five years, resource estimates at the time of the budget guidelines were typically below those in the budget. In recent years the gap has narrowed from around 12 percent difference between the budget guidelines and the budget in FY02 to around six percent in FY04. Given the inherent uncertainties in making projections, a difference of five percent can well be interpreted as a sign of cautious budgeting rather than a symptom of major problems in making macro-fiscal projections.

Figure 9: Differences in Total Resource Estimates

Difference in Total Resource Estimates

-10.0%

-5.0%

0.0%

5.0%

10.0%

15.0%

FY00 FY01 FY02 FY03 FY04

Budget - BG Actual - Budget

The analysis of the components of the resource envelope indicates that projections of program support undergo the largest change between the budget guidelines and the budget. At the same time, the narrowing of the difference between resource estimates in the budget guidelines and the budget is also mainly attributable to improved program support projections in the budget guidelines. While in FY00, program support shown in the budget was about 60 percent higher than estimates in the budget guidelines

90

(accounting for about six percentage points of the difference in estimates of the resource envelope), in FY03 this figure has declined to below 20 percent (or about 3 percentage points of the total difference). This improvement is due to efforts by government and the donor community to enhance predictability in program support through indicative medium term commitments of budget support through the PRBS facility.

Figure 10: Domestic Revenue, Program Grants and Loans, and Project Grants and Loans, Difference between Budget and Budget Guidelines, FY00 – FY04

Dif ference betw een Budget and BG(% of component)

-20.0%-10.0%

0.0%10.0%20.0%30.0%40.0%50.0%60.0%70.0%


Domestic Revenue Program Grants and Loans

Project Grants and Loans

Reflecting the observations made on the resource side, estimates of aggregate expenditure levels at the time of the budget guidelines are consistently below the estimates that are presented in the budget six months later. However, the degree of underestimation has declined in recent years. While in FY02, the difference between the estimates of total expenditure was twelve percent, it has fallen to only six percent in FY04.

Figure 11: Total, Recurrent and Development Expenditure, Difference Between Budget and Budget Guidelines Estimates, FY00 – FY04

Diff erence betw een budget an budget guidelines estimates (% of total expenditure)

-4.0%-2.0%0.0%2.0%4.0%6.0%8.0%

10.0%12.0%14.0%


Total Recurrent Development

91

Dif ference betw een Budget and BG (% of total resources)

-2.0%

0.0%

2.0%

4.0%

6.0%

8.0%


Domestic Revenue Program Grants and Loans

Project Grants and Loans

It is interesting to note that that the differences between the budget guidelines and budget estimates are mainly on account of underestimation of recurrent expenditures, while estimates of development expenditure seem to undergo less change between the budget guidelines and the budget.

Figure 12: Wages and Salaries, Operations and Maintenance, and Debt Service, Difference between Budget and Budget Guidelines, FY00 – FY04

Difference betw een budget and budget guidelines estimates (% of BG recurrent expenditures)

-2.0%

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%


Wages and Salaries Operations and Maintenance

Debt Service and Others

Looking at the main components of recurrent expenditure, i.e., wages and salaries, operations and maintenance, and debt service reveals interesting patterns for these sub-categories. The estimates of wages and salaries show virtually no difference between the budget guidelines and the budget. This reflects the Tanzanian practice of setting wage bill targets as a percentage of GDP during the budgeting stage. In the budget, resources for salary increases are retained under the category of other expenditures and are typically only transferred to specific votes during budget execution.

Estimates of debt service are typically under-estimated at the budget guidelines stage, with the magnitude of change between the budget guidelines and the budget ranging between five and 20 percent.

Finally, estimates of expenditures for operations and maintenance also show a consistent pattern of budget guidelines estimates being less than the final budget estimates. Among the various expenditure categories in the budget guidelines, this is certainly the most important one for the budget process since the sectoral ceilings for operations and maintenance are based on this figure. Overall, it appears that changes between the budget guidelines and the budget have become smaller, declining from around 23 percent in FY00 to around 10 percent in FY04.

In summary, our analysis indicates that in recent years the projections of broad resource and expenditure categories have improved and changes between the budget guidelines and the budget have become smaller. A driving force behind this development is apparently greater predictability of program support as a result of efforts by government

92

Difference Budget and BG

-5.0%

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%


Wages and Salaries Operations and Maintenance

Debt Service and Others

and the donor community to provide clear indications of likely budget support already prior to the preparation of the budget guidelines.

Assessment of Outer Year Estimates

The budget guidelines also contain projections of revenue, aggregate expenditure, as well as sectoral and ministerial expenditure levels over a three-year horizon. In the Tanzanian context, these expenditure projections are intended to fulfill a dual purpose. The first is related to expenditure planning where the adoption of a medium term outlook on expenditure planning is intended to increase the strategic focus of decision and allow for a better link between policy, planning, and budgets. The second objective relates to the dialogue with the donor community. As donors are providing a significant share of their assistance in the form of general budget support, the government budget has moved to the center of the dialogue between government and the donor community. A credible MTEF allows this dialogue to focus on the overall public expenditure strategy of the authorities rather than the discussion of specific budget items.

Overall Trends

Generally, since FY00, total expenditure has been increasing at higher rate, apparently, than the planning officers have expected. On average, the Annual Budget (in nominal terms) was increasing annually by 23% during the FY00 to FY03 period. As illustrated in Figure 13, the first-year of each of the rolling three-year Budget Guidelines is significantly higher than the second-year of the previous rolling Budget Guideline. It is also modestly higher than the Annual Budget being executed that fiscal year. Yet, despite the upward adjustments, the Annual Budget and the Actual Out-turn always turns higher than the 1st-Year Budget Guideline. In Figure 13, 1st-year Budget Guidelines are connected by a solid bar.

Figure 13: Total Expenditure, Budget Guidelines, Annual Budget, and Actual Expenditure, FY00 – FY07

Figure 1: Total ExpenditureBudget Guidelines (BG), Annual Budget, and Actual Expenditure

BG FY00-FY02

BG FY01-FY03

BG FY02-FY04

BG FY03-FY05

BG FY04-FY06

Annual Budget

Actual Out-turn

BG FY05-FY07

1,000,000

1,500,000

2,000,000

2,500,000

3,000,000

3,500,000

FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07

Ts

h M

illi

on

s

93

As indicated above, though the 1st-year Budget Guideline was significantly higher than the 2nd-year of the previous plan, it seems that the outer years of each Budget Guidelines (BG) are rolled over to the following year’s plan. Each retiring plan serves as a starting point in the preparation of the first and second year of the following plan.

Figure 14: Recurrent Expenditure, Budget Guidelines, Annual Budget, and Actual Outturn, FY00 – FY07

Recurrent ExpenditureBudget Guidelines (BG), Annual Budget, and Actual Expenditure

BG FY00-FY02

BG FY010-FY03

BG FY02-FY04

BG FY03-FY05

BG FY04-FY06

Annual Budget

Actual Expenditure

BG FY05-FY07

500,000

700,000

900,000

1,100,000

1,300,000

1,500,000

1,700,000

1,900,000

2,100,000

2,300,000

2,500,000


Tsh

Mil

lio

ns

Recurrent Expenditure: The programming of recurrent expenditure demonstrates a clear and pattern, with each of the rolling 3-yrar BG showing steep and similar slopes. As indicated before, though planners were optimistic and programmed significant increases, it turns out that they had understated the resource envelope. Roughly, in every plan, planning officers allow T.Shs 30-70 billion annual increases. Yet, every 1st-year Budget Guideline is higher T.Shs 200 billion than the 1st-year of its predecessor and the Annual Budget and Actual Expenditure turn out even higher.

94

Figure 15: Development Expenditure, Budget Guidelines, Annual Budget, and Actual Outturn, FY00 – FY07

Development Expenditure: Planning and budgeting for development expenditure over the medium-term have apparently been less predictable than the recurrent expenditure. The successive 3-year Budget Guidelines illustrated in Figure15 reflect uncertainties on the part of planners on the prospects of development programs over the medium-term. The 1st-year Budget Guideline for development expenditure is very different from the 2nd-year of the previous plan.

Recurrent Expenditure Sub-Trends

Based on the Budget Frame (accounting) classification, recurrent expenditure is divided in to three main categories: Other Charges (OC), Wages and Salaries (PE), and CFS plus other small items (DI&CFS). The percent composition of the recurrent expenditure is provided in Table 21. Composition of Recurrent Expenditure. The Table illustrates that the share of OC increased in relation to PE and DI&CFS. The trend is true for Budget Guidelines, Annual Budget, and Actual Expenditures.

95

Development ExpenditureBudget Guidelines (BG), Annual Budget, and Actual Out-turn

BG FY00-FY02

BG FY01-FY03

BG FY02-FY04

BG FY03-FY05

BG FY04-FY06

Annual Budget Actual Out-turn

BG FY05-FY07

-

200,000

400,000

600,000

800,000

1,000,000

1,200,000

1,400,000


Tsh

Mil

lio

ns

Table 21: Composition of Recurrent Expenditure

FY00 FY00 FY03 FY03 Budget Actual Budget Actual Salaries and Wages 29.7 31.6 27.4 26.7 Other Charges 19.9 24.0 33.8 59.7 Designated Items, CFS, and Other 50.3 44.4 38.8 13.6 Total 100.0 100.0 100.0 100.0

Figure 16: OC Expenditure, Budget Guidelines (BG), Annual Budget, and Actual Expenditure, FY00 – FY07

Figure 5: OC ExpenditureBudget Guidelines (BG), Annual Budget, and Actual Expenditure

BG FY00-FY02

BG FY01-FY03

BG FY02-FY04

BG FY03-FY05

BG FY04-FY06

Annual Budget

Actual Out-turn

BG FY05-FY07

-

200,000

400,000

600,000

800,000

1,000,000

1,200,000


Tsh

Mil

lio

ns

Though the increase in OC expenditures relative to other categories were policy-driven, the OC often benefits also from updates in the resource envelope Figure 16 indicates that the ‘Other Charges’ (OC) category in the Annual Budget or Actual Expenditure figures represent upward adjustment compared to that specified in the Budget Guidelines. Some of the upward shift can also be a reallocation within the resource envelope, particularly from DI&CFS.

96

Figure 17: PE Expenditure, Budget Guidelines (BG), Annual Budget, and Actual Expenditure, FY00 – FY07

Figure 6: PE ExpenditureBudget Guidelines (BG), Annual Budget, and Actual Expenditure

BG FY00-FY02

BG FY01-FY03

BG FY03-FY05

BG FY04-FY06

BG FY02-FY04

BG FY05-FY07

250,000

350,000

450,000

550,000

650,000

750,000

850,000


Tsh

Mil

lio

ns

In contrast to OCs, PEs are often not subject to significant change since they are centrally allocated. Compared to the Budget Guidelines, Figure 17 illustrates that there are far less changes in PE in the Annual Budget or in the Actual Out-turn for the FY00 to FY03 period. In Figure 18, the Annual Budget and the Actual Out-turn are identical and the two representative lines are on the top of each other.

Figure 18: CFS Expenditures, Budget Guidelines (BG), Annual Budget, and Actual Expenditure, FY00 – FY07

Figure 7: Designated Items and CFS ExpenditureBudget Guidelines (BG), Annual Budget, and Actual Expenditure

BG FY00-FY02

BG FY01-FY03

FY02-FY04

BG FY03-FY05

BG FY04-FY06

Annual Budget

Actual Out-turn

BG FY05-FY07

100,000

200,000

300,000

400,000

500,000

600,000

700,000

800,000

900,000


Tsh

Mil

lio

ns

97

Of the three components of the recurrent expenditure, CFS expenditure is the least predictable. Figure 18 illustrates the pattern of the rolling three-year Budget Guidelines, the Annual Expenditure, and the Actual.

Analysis of budget deviation at sub vote level

In looking at the degree of consistency between approved budgets and actual expenditure, it is important to note that the level deviation observed varies greatly with the level of aggregation of the analysis. Many of the sharpest concerns on effective implementation of the budget concern expenditure at the level of individual programs rather than broad sector aggregates. Performance on implementation of individual programs can be examined, but it is valuable to draw these results into a more general picture of how well the budget predicts actual expenditure at this level of detail. What is needed is analysis which draws on a lower level of aggregation and yet provides a picture of what is happening in terms of overall budget implementation, rather than a commentary on individual programs or sub votes.

For a preliminary analysis of budget execution at lower levels of aggregation, it seemed most useful to focus on expenditure data by sub vote. Data are available for three years from FY01 to FY03. An index of budget deviation was calculated as the sum of absolute differences between approved budget and actual expenditure at the sub vote level, expressed as a percentage of the total budget. For each subvote the budget figure under reference is from the budget, book as approved by parliament, taking no account of any reallocations subsequent to parliamentary approval. The table below shows preliminary summary results for these years, for central government recurrent expenditure.

Table 22: Budget Deviation at Sub-Vote Level (average as % original budget)Percentage deviation

FY01 22.6FY02 26.3FY03 12.3

Some observations can be made on these preliminary results. First, it is important to consider the implications of the degree of deviation observed. Deviation at subvote level of the order reported is clearly a significant problem for effective management of resources. Given that the level of analysis is still fairly aggregate, it must be assumed that deviations at the level of individual spending units are even higher, implying a serious problem for spending managers in planning activities. Variances can arise from a wide range of possible sources and it should not be assumed that deviations should optimally be reduced to zero. Nevertheless, deviations of the order observed imply reallocation at a level, which must be expected to have adverse effects on the efficiency of expenditure.

Secondly, there appears to be a significant improvement in performance during FY03. Tracing through the potential causes of the pattern in observed deviation will require further analysis, but there seem to be good grounds for associating the improvement with a tightening of the approval process for reallocations within year, starting FY03. Performance may also have been assisted by the reform of budget classification in the

98

previous year, allowing better alignment between expenditure classification and the organizational structure.

A.5 TRENDS IN THE PRIORITY SECTORS

The priority sectors (Education, Health, Water, Agriculture, Roads, HIV/AIDS, and Judiciary) in comparison to non-priority sectors, enjoy a greater attention and are subject to an intense and detailed scrutiny in programming and execution. The seven priority sectors absorb an increasing large proportion of the central government expenditures. Between FY00 to FY03, the proportion of the allocated Annual Budget to the priority sectors has increased from 33% to 43% of the total, including CFS.

Data Sources and Definitions

The data presented in the following analysis are extracted from the ‘Sectoral Allocation’ Tables in both the Budget Guidelines and Cross-Sector MTEF documents. Given that the available data in these documents provided same format and coverage only for some years/sectors, the coverage is limited. For sectors, they fortunately cover four key priority sectors: Education, Health, Water, and Roads. These sectors account for over 90% of total expenditure to priority sectors, and therefore are fairly representative.

A key difference (from Section 4) in the classification of this data is that recurrent expenditure is the sum of PE and OC. The total recurrent expenditure for the priority sectors in the following discussion represents the sum of the four selected sectors. Since PE is largely not subject to change (as illustrated in Figure 17) and development expenditures were not treated in earlier years/documents, the discussion of MTEF in the following paragraphs is solely on OC. All differences with respect to recurrent expenditure are OC differences.

Overall Trends

The recurrent expenditure allocation to priority sectors was, in general, increasing at a faster rate than the average overall increase. Figure 19 illustrates that, in the short period between FY00 to FY03, the recurrent expenditure to priority sectors more than doubled. There is a slight upward adjustment from BG/MTEF to Annual Budget, and then to the Actual. As indicated earlier, all adjustments primarily occur in the ‘Other Charges’ category.

The consistency between Budget Guidelines, MTEF, the Annual Budget, and the Actual, is sharper than average for the sectors discussed in the previous section. The difference between the data sets is very small indeed (Figure 20). The difference between the 1st years of BG and MTEF I and II (FY00-02 and FY01-03) is nearly zero. According to the statistics, the discrepancy in FY02-04 has to do with an upward adjustment for the Health sector totaling T.Shs 25 billion thereby causing a difference of 6% in that fiscal year. The percent discrepancy between MTEF and the Annual Budget and between the Annual Budget and the Actual ranged from 1% to 6% and 2 to 4%, respectively.

99

Figure 19: Recurrent Expenditure of Selected Priority Sectors (Education, Health, Water, and Roads), Budet Guidelines, MTEF, Annual Budget, and Actual Expenditures, FY00 –

FY04

BG FY00-02

BG FY01-03

BG FY02-04

MTEF FY00-02

MTEF FY01-03

MTEF FY02-04

Annual Budget

Actual Out-turn

250

300

350

400

450

500

550

600

650


Figure 20: Selected Priority Sector Expenditures, Difference between 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04

-3.0

-2.0

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

FY00 FY01 FY02

Per

cen

t

1st-Year BG vs 1st-Year MTEF

1st-Year MTEF vs Annual Budget

Annual Budget vs Actual

Just like the Budget Guidelines, the outer two years of MTEF are rolled over to the subsequent MTEF plan. Just like the Budget Guidelines, the outer years of MTEF (year 2 and 3) allowed modest increases but turned out to be conservative when compared to

100

the Annual Budget and Actual Out-turn. The allocations for outer two years have been conservative compare to the following MTEF plan.

Sector Trends

The priority sector trends reflect the general trend described above for the priority sectors. Education and Health, in particular, receive the highest priority within the priority sectors and, in the past, received most of their expenditure requirements. Trends in Education, which absorbs about half of the allocation to priority sectors, is illustrated in Figure 21. The 1-st Year Budget Guidelines and that of MTEF exactly coincide, and the graph lines are on top of each other. There is some of adjustment upward with regard to the Annual Budget notably in FY02 and another upward adjustment during the execution in the same fiscal year.

Figure 21: Recurrent Expenditure on Education, 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04

100

150

200

250

300

350

400

450

FY00 FY01 FY02 FY03

Tsh

Billio

ns

1st-Year Budget Guidelines

Annual Budget

Actual Out-turn

1st-Year MTEF

Recurrent expenditure in Health, the second most important priority sector, is illustrated in Figure 22. With regard to variance, the most notable one is the T.Shs 25 billion adjustment (discussed before) in FY02 but which was scaled down in the Annual Budget and the Out-turn.

101

Figure 22: Recurrent Expenditure on Health, 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04

Figure 23: Recurrent Expenditure on Water, 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04

The recurrent expenditure on Water and Roads sectors are presented in Figure 23 and Figure 24. In water, the most significant difference between the budget data sets is in FY02. In the latter, BG and MTEF figures are identical. However, both the Annual

102

50

60

70

80

90

100

110

120

130

FY00 FY01 FY02 FY03

Tsh

Bil

lio

ns


1st-Year MTEF

Annual Budget

Actual Out-turn

50

60

70

80

90

100

110

120

130

FY00 FY01 FY02 FY03

Tsh

Bil

lio

ns


1st-Year MTEF

Annual Budget

Actual Out-turn

50

60

70

80

90

100

110

120

130

FY00 FY01 FY02 FY03

Tsh

Bil

lio

ns


1st-Year MTEF

Annual Budget

Actual Out-turn

4

6

8

10

12

14

16

18

20

FY00 FY01 FY02 FY03

Tsh

Bill

ion

s


Annual Budget

Actual Out-turn

1st-Year MTEF

Budget and Actual Out-turn are relatively higher. Similarly, in the Roads sector, the BG and MTEF data are identical. The Annual Budget is higher, but the out-turn is lower than the MTEF program.

Figure 24: Recurrent Expenditure on Roads, 1st Year Budget Guidelines, 1st Year MTEF, Annual Budget, and Actual Outturn, FY00 – FY04

40

50

60

70

80

90

100

110

FY00 FY01 FY02 FY03

Tsh

Bill

ion

s


Annual Budget

Actual Out-turn

1st-Year MTEF

A.6 CONCLUSION

The preceding analysis of the budget data illustrates and confirms that some consistency, in aggregate terms, has been achieved with regard to planning, programming, and execution of the budget. The consistency is particularly high in the priority sectors where there is less deviation between the four sets of budget data. Compared to the situation of planning, budgeting, and execution problems a decade ago, this is an impressive achievement. It signifies that budget projections/programming are/is realistic, the policy and activities’ coordination between various actors inside and outside government is very good, and the institutional capacity to program and execute the budget with confidence and precision (particularly in the priority sectors) has improved.

Overall, the budget cycle has earned respectability and increased predictability. The approach is pragmatic but conservative with norms that: (i) stress realism, participation, and local ownership; (ii) protect against over-estimation and over-spending; (iii) enforce fiscal discipline and accountability; and (iv) give precedence to priorities.

A key cause of the difference between the Budget Guidelines and the Annual Budget is the conservative attitude of the planning officers toward setting an overall resource envelope plus the mere fact that external funding for recurrent expenditure has been increasing at a rate faster than any one has expected. In the priority sectors, there is little deviation between Budget Guidelines and MTEF figures, MTEF and the Annual Budget, and the Annual Budget vs. Actual Out-turn. Updates in the budget cycle are

103

predominantly an upward adjustment of resources that benefited OCs most. PEs are hardly subject to change from planning to execution. This is, overall, an remarkable achievement which clearly demonstrates the strong commitment on the part of the government, and the degree of compliance and manifestation of ownership by all agencies/sectors. Whether this manifested consistency between the aggregates of data sets, however, trickles down to line items and sub-votes, is an open question and the available data does not allow us to explore it further.

There are two issues worth noting. First, since the consistency is higher in priority sectors than in the overall average, the margin of adjustment in non-priority sectors during the budget cycle must be higher and, therefore, should be reduced. Second, programming development expenditure over the medium-term has been less effective than for recurrent expenditure. Past budget programming (as reflected by the rolling plans of the Budget Guidelines since FY00) did not convey or indicate a clear path for investment resources to flow over the medium-term. This is an issue, which may require both strengthening of institutional capacity, but may also reflect uncertainties in the funding and external commitments.

104