United Republic of Tanzania Public Expenditure and Financial Accountability Review 2008

Transcription

1 Public Disclosure Authorized Report No TZ United Republic of Tanzania Public Expenditure and Financial Accountability Review 2008 Public Disclosure Authorized June 2009 Prepared by the Members of Tanzania PER Working Group Public Disclosure Authorized United Republic of Tanzania, Ministry of Finance and Economic Affairs Development Partners, PER Macro & PFM Group Public Disclosure Authorized Document of the World Bank

2

3 Table of Contents ACKNOWLEDGEMENTS... i EXECUTIVE SUMMARY PART I: BUDGET ANALYSIS INTRODUCTION MACRO-FISCAL POLICY AND MACRO-ECONOMIC BUDGET FRAMEWORK. 3 MACRO-FISC AL. POLICY: THE IMPACT OF REVENUE AND FINANCING ON THE MACRO CONTEXT... 4 MACRO-FISCAL POLICY: IMPACT OF EXPENDITURES ON THE MACRO CONTEXT... 6 THE MACRO ECONOMIC FRAMEWORK OF THE BUDGET: OVERVIEW... 9 RECOMMENDATIONS BUDGET ALLOCATION ALLOCATION BETWEEN MKUKUTA AND NON-MKUKUTA SECTORS ALLOCATIONS BETWEEN MKUKUTA CLUSTERS ALLOCATION BETWEEN MAIN SECTORS ALLOCATION WITHIN SECTORS ALLOCATION BETWEEN CENTRAL AND LOCAL GOVERNMENT RECOMMENDATIONS IMPACT/ RESULTS IN PUBLIC SERVICE DELIVERY BUDGET EXECUTION PERFORMANCE EFFECTIVENESS OF PUBLIC SERVICE DELIVERY EFFICIENCY PART 11: PFM ASSESSMENT INTRODUCTION COUNTRY BACKGROUND INFORMATION ECONOMIC SITUATION BUDGETARY OUTCOMES THE LEGAL AND INSTITUTIONAL FRAMEWORK FOR PFM ASSESSMENT OF THE PFM SYSTEMS. PROCESSES AND INSTITUTIONS BUDGET CREDIBILITY COMPREHENSIVENESS AND TRANSPARENCY POLICY-BASED BUDGETING PREDICTABILITY AND CONTROL IN BUDGET EXECUTION ACCOUNTING, RECORDING AND REPORTING EXTERNAL SCRUTINY AND AUDIT DONOR PRACTICES PFM REFORM PROCESS... 72

4 RECENT AND ONGOING REFORMS INSTITUTIONAL FACTORS SUPPORTING REFORM PLANNING AND IMPLEMENTATION PART 111: PEFAR CAPACITY BUILDING PROGRAM BACKGROUND AND OBJECTIVES OF THE MISSION MAIN ACCOMPLISHMENTS OF THE MISSION STRENGTHENING THE MACROECONOMIC MODEL - MACMOD PREPARING THE MACRO FRAMEWORK 2009/10-11/12 FOR THE PLAN BUDGET GUIDELINES PREPARING THE BUDGET BACKGROUND AND MEDIUM TERM FRAMEWORK STRENGTHENING THE MTEF BUILDING A PROGRAMMATIC DIMENSION IN THE SP AND THE MTEF ACTION PLAN FOR THE REMAINDER OF THE YEAR (2009) STRENGTHENING MACMOD-TZ AND THE MACRO POLICY FRAMEWORK PAPER 2009/10.11/ MTEF UPGRADES AND SECTOR PROGRAM BUDGETING... 84

5 List of Annexes ANNEX 1 : PERFORMANCE INDICATORS SUMMARY TABLE ANNEX 2 : ASSESSMENT OF PEFAR 08 CAPACITY BUILDING WORK ANNEX 3 : STATISTICAL TABLES List of Figures FIGURE 1 : YEAR-ON-YEAR MONTHLY CPI INFLATION RATE. NEW SERIES... 3 FIGURE 2: CURRENT ACCOUNT AND FOREIGN AID FIGURE 3: TOTAL AID INFLOWS RELATIVE TO DOMESTIC REVENUES 2005/ / FIGURE 4.. ECONOMIC CLASSIFICATION OF THE 2008/09 BUDGET... 7 FIGURE 5: CHANGE IN BUDGET SHARES BETWEEN 2007/08 AND 2008/ FIGURE 6: LOCAL AND FOREIGN FUNDING IN INFRASTRUCTURE AND WATER BUDGETS FIGURE 7: SECTOR SHARE FOR SELECTED MKUKUTA SECTOR FIGURE 8: FUNCTIONAL COMPOSITION OF THE 2008/09 APPROVED AGRICULTURE SECTOR BUDGET FIGURE 9: COMPOSITION OF TOTAL EDUCATION EXPENDITURE BY LEVELS FIGURE 10: GOVERNMENT ALLOCATION TO RURAL & URBAN WATER SUB-SECTORS (TSH. MILLIONS) FIGURE 11: MAINTENANCE AND INFRASTRUCTURE SPENDING IN SELECTED SECTORS (IN BIL. TSH.) FIGURE 12: ROAD MAINTENANCE NEEDS AND BUDGET (BILL. TSH) FIGURE 13: DEVELOPMENT BUDGET EXECUTION /07 AND 2007/ FIGURE 14: NET ENROLLMENT RATIOS - PRIMARY, LOWER SECONDARY, , HBS FIGURE 15: ACCESS TO SAFE DRINKING WATER SUPPLY PER PERCENT OF POPULATION FIGURE 16: UNIT COST PER KM IN ROUTINE AND RECURRENT MAINTENANCE FIGURE 17: EDUCATION PERSONNEL EMOLUMENT EXPENDITURE PER CHILD AGED 7-13 ACROSS LGAS. 33 FIGURE 18: HEALTH PERSONNEL EMOLUMENT EXPENDITURE PER CAPITA ACROSS LGAs FIGURE 19: LGA RANKED IN TERMS OF PER CAPITA EXPENDITURE BUDGETED FOR 2008/ FIGURE 20: DISTRICT RECURRENT BUDGET AND ROAD LENGTH List of Tables TABLE 1 : DOMESTIC REVENUE: TREND OVER 2004/ / TABLE 2: CONTRIBUTION OF T U DEPARTMENTS TO OVERALL TAX COLLECTION... 5 TABLE 3. SELECTED MDAS BUDGET EXECUTION GAP (IN PERCENT)... 8 TABLE 4: DECOMPOSITION OF INCREASE IN BASIC WAGES IN 2008/ TABLE 5: FISCAL FRAMEWORK, 2005/ / TABLE 6: MEDIUM-TERM PROJECTION ANALYSIS TABLE 7: MKUKUTA AND NON-MKUKUTA EXPENDITURES (SHARES OF TOTAL) TABLE 8: SELECTED NON-MKUKUTA BUDGET ALLOCATIONS (SHARES OF TOTAL) TABLE 9: BUDGET GUIDELINES: CLUSTER ALLOCATION TABLE 10: MKUKUTA ALLOCATIONS LGA TRANSFERS, EXCL. MDA WAGES) TABLE 11: MKUKUTA ALLOCATIONS (INCL. LGA TRANSFERS AND MDA WAGES) TABLE 12: 2008/09 HEALTH BUDGET BY LEVEL OF CARE AND TYPE OF EXPENDITURE (SHARE) TABLE 13. OVERALL BUDGET EXECUTION GAPS TABLE 14. MDAs DEVELOPMENT BUDGET EXECUTION / TABLE 15: CENTRAL GOVERNMENT OPERATIONS (% OF GDP) TABLE 16: ECONOMIC CLASSIFICATION OF EXPENDITURES (% OF GDP) TABLE 17: COMPARISON OF ORIGINALLY BUDGETED AND ACTUAL EXPENDITURES TABLE 18: VARIANCE OF EXPENDITURE COMPOSITION TABLE 19: DOMESTIC REVENUE PERFORMANCE TABLE 20: EXPENDITURE ARREARS COMPARED TO TOTAL EXPENDITURES... 47

6

7 ACKNOWLEDGEMENTS Context Tanzania has carried out an annual Public Expenditure Review (PER) process since 1997/98 with twin objectives of supporting the budget process and undertaking an external review of fiscal developments. The Tanzania PER Working Group, comprising representatives from the Government of Tanzania (GOT), the World Bank, United Nations (UN) agencies, other bilateral and multilateral donors, research and academic institutions, and NGOs, determined the agenda for the annual PER process, guided and financed the implementation of the agreed work program, and reviewed all outputs. In 2004/05 all stakeholders agreed as part of improved harmonization and coordination that assessment of Tanzanian public expenditure issues and financial management (including procurement) issues should be carried out as part of a broader Public Expenditure and Financial Accountability Review (PEFAR). The PEFAR aims at integrating two diagnostic exercises, PER and Country Financial and Accountability Assessment (CFAA), to facilitate a comprehensive assessment of Public Finance Management (PFM). The overall purpose of the PEFAR is to provide the Government of Tanzania and Development Partners with a comprehensive, integrated and candid assessment of Tanzania s key fiduciary risks as reflected in GOT S resource allocation, resource management and control, resource utilization, and accountability processes, and to make recommendations for improving the PFM framework, institutional performance, and capacity building. The Macro Subgroup of the PER Working Group is charged with responsibility for coordinating and guiding the work of reviewing public expenditure and assessing financial accountability under the broader PEFAR. The Annual PEFAMUKUTA (Mkakati wa Kukuza Uchumi na Kupunguza Umaskini Tanzania) Consultative Meeting is an important forum for raising awareness and promoting discussion of poverty reduction, economic growth, public expenditure and financial accountability issues among government and a wide array of interested stakeholders in Tanzania. Team composition The report is primarily based on the findings of the PEFAR 2008 team, which comprised members from the World Bank, the International Monetary Fund (IMF), IMF AFRITAC, European Union, Kreditanstalt fuer Wiederaufbau (KfW), Department for International Development (DFID), JICA, Canadian High Commission, Embassy of Finland, the Netherlands Embassy, and Ireland Embassy. The team was also joined by a number of independent consultants (from University of San Francisco, University of Cape-Town and University of Dar es Salaam) and members from Civil Society Organizations (CSOs), including Hakielimu and Policy Forum. The task manager and principal author of the report is Emmanuel Mungunasi Substantive inputs and background papers were prepared by Paolo Zacchia (rapid budget analysis synoptic note); Florence Charlier, Denis Biseko, Emmanuel Mungunasi, David Robinson, Chelaus Rutachurunva, Florence Kutesa, Vera Mshana, and Gregory Smith (aggregate/macro analysis, wage bill analysis, and budget execution analysis); Shireen Mahdi (budget execution analysis); Gregory Smith, Angela Tormin, Tomi Sarkioja and Goodluck Mosha (local government); Josaphat Kweka, Stevan Lee and Samuel Saigurani (education); Cam Do and Goodluck Mosha (health); Caroline van den Berg (water); Asuka Tushoboike, Retsu Hagiwara, and Jonathan Wolsey (road transport); Sergiy Zorya (agriculture); Rene van Nes, Parminder Brarr and Feridoun Saraff (public financial management assessment); Charles Ncho, Florence 1

8 Charlier, Fulbert Tchana Tchana, Leonidas Luteganya and Emmanuel Mungunasi (PEFAR capacity building work program progress report) and Claudio Paul (data). The financial support and overall leadership of the European Commission on the PFM assessment is highly acknowledged. Also, financial support by the CIDA, the Swiss State Secretariat for Economic Affairs (SECO), and the World Bank on the PEFAR capacity building work program is highly appreciated. The report presents the findings of the budget analysis and execution (core analysis), public financial management assessment as well progress in PEFAR capacity building work program in Tanzania. The initial drafts were shared with government and members of the PER Macro sub-group as well as with a wider audience at the Annual PEFAWKUKUTA Consultative Meeting held in November 2008, where they provided useful comments that are included in this report. The task was carried under the general guidance of the Country Director, John McIntire. Technical supervision and quality assurance was provided by Kathie Krumm, Sector Manager, AFTP2 and Mr. Ramadhan Khijjah, Permanent Secretary, Ministry of Finance and Economic Affairs. Allister Moon (AFTP2) served as peer reviewer. Mary-Anne Mwakangale and Mwanaisha Kassanga were responsible for the word processing and actual production of the report. Finally, the PEFAR 2008 team would like to express its sincere gratitude for the assistance and courtesies extended by all parties that participated in the exercise. In particular, the team wants to mention the names of Mr. Mugisha Kamugisha (Commissioner for Policy Analysis) and Mrs. Monica Mwamunyange (Commissioner for Budget) for the coordination of the whole exercise as the cochairs of the PER Macro sub-group. 11

9 EXECUTIVE SUMMARY (FINDINGS AND RECOMMENDATIONS) Macro-fiscal Policy and Macro-economic Budget Framework 1. Rising inflation represents a serious challenge for the government, including fiscal policy. By December 2008, inflation has risen to 13.5 percent, far above the government target of 5 percent since It is important the government continue its effort to reduce inflation through monetary and fiscal policies, including reducing inflationary pressures on the budget by controlling wage bill growth in line with medium-term pay policy. 2. The 2008/09 Medium-Term Expenditure Framework (MTEF) projected optimistic targets for domestic revenue and, at a same time, pessimistic targets for foreign aid. Although domestic revenue has been rising over the recent past, the targeted 18.5 percent of Gross Domestic Product (GDP) in revenue effort is high given the fact that revenue has increased by only 1 percent of GDP annually in the past 5 years. It is important to improve reliability of the macro-framework by sustaining efforts to develop capacity in macro-fiscal policy and macro-modeling, which is an important tool for the government to properly assess expected levels of domestic revenue collection over the medium term. 3. The government has continued to reduce aid dependency, with share of aid in the total budget coming down to 35 percent in 2008/09. Aid dependency is expected to decline further over the medium term as government increases its domestic revenue effort. However, the projected trend of decline in foreign aid raises some concern, since aid level has not declined over the past 5 years. It is important that the government engage Development Partners (DPs) over the medium-term projected level of financial assistance, with a view to avoiding abrupt reduction. 4. The pattern of aid has remained mixed over the past 4 years, with years of significant overbudgeting and under-budgeting of aid. Failure to properly assess aid flows has hampered budget management and execution. In order to avoid this problem, it is important that the government develop a Priority Expenditure Plan (especially a priority investment plan) in line with MKUKUTA objectives, with clear indications of (i) programs that would be protected and financed through domestic borrowing in case of aid shortfall and (ii) a pipeline of high return projects that could absorb any aid surplus. 5. The government has cautiously continued with its policy of zero net domestic financing. However, the government may consider using domestic financing for a limited amount of expenditures smoothing, as long as real domestic interest rates remain low. This will be important in order to protect key programs given unexpected reductions in foreign aid. 6. Development budget share in total budget has increased significantly over the recent past, reaching 35 percent and equivalent to 10 percent of GDP in 2008/09. Local funding component of development budget has also increased, especially in infrastructure sectors. Despite increased share of development budget, the share of the public investment in the budget has remained the same at 24 percent. With the government objective of sustained annual GDP growth of 8-10 percent, increased contribution of public investment in total investment is important. This can be achieved only by improving growth orientation of the budget by increasing the share of public investment. 7. The government has also increased the share of local funding in development budgets of key infrastructure sectors like water, roads, and energy. The increased local funding has given the government some flexibility to accommodate fluctuations in project funding as well as reducing the possibility of

10 donors imposing their budget priorities though earmarked project funding. It is important to reduce this vulnerability by continuing to provide sizeable local funding to the development budget in key sectors. Budget Allocation 8. The government has continued to allocate a significant share of its budgetary resources to implement MKUKUTA. The share of MKUKUTA allocations in the 2007/08 and 2008/09 budgets are 70.6 percent and 70.8 percent, respectively. The slight increase in share of MKUKUTA in 2008/09 is due to an increase in ministries, departments and agencies (MDAs) wages, since MKUKUTA shares without MDAs decline from 64.5 percent in 2007/08 to 62.0 percent in 2008/09. It is therefore important that the government have a clear commitment to protect MKUKUTA allocations. 9. The applicable definition of MKUKUTA allocations applies only to other charges and the development budget. The MDA wages and transfers to Local Government Authorities (LGAs). are not allocated according to MKUKUTA. However, MDA wages and transfers to LGAs are important elements for implementing MKUKUTA, especially public service delivery in health and education. It is important that MKUKUTA definitions be redefined to include MDA wages and transfers to LGAs in order to be strategically relevant. 10. The Planning and Budget Guidelines that provide strategic direction in preparation of the budget and the MTEF do not seem to show a clear medium-term strategic direction in terms of allocation among three clusters. Allocations have moved more toward 40:40:20 shares despite the government intentions of boosting economic growth. It is important, therefore, that the government express a clear medium-term strategy, which is economic growth, through MKUKUTA cluster allocation and strives to implement it The Transport Sector Investment Plan has proved a useful anchor for medium-term budgeting in the road sectors by providing a costing of the sector strategy and some form of prioritization. This experience could be extended to other sectors, such as agriculture. 12. Although the share of allowances in the overall wage bill has decreased over the past two years, for some sectors, the share has increased. For instance, the education and health ministries have entered the group of the top five allowance receiving ministries. Most services provided by these sectors are at the LGA level, but most duties facilitating allowances are paid to staff at the central level (ministry). Hence there is a need to control the proliferation of allowances in the social sector ministries (especially the health and education ministries). 13. Slightly less than 1 percent of the allocations for social sectors are allocated to infrastructure maintenance. With the current government effort toward construction of social infrastructure, such as classrooms, teacher houses, and health centers, maintenance cost should go up in the future. This implies that the government should also increase the share of maintenance in health and education budgets in order to ensure these infrastructures good status. 14. The government has continued to make progress toward the 25 percent target of LGA transfers in However, there has been a stagnation of these transfers at around 20 percent over the past two years. This poses a serious challenge to the government in terms of achieving its target, which implies further effort would be required in order to reach the 25 percent objective of LGA transfers. iv

11 Impact / results 15. The government should continue to improve the results/ outcome orientation of the budget and budget execution. Steps include: 16. Streamline basket funding modalities. Despite basket funds operating under the same financial management system, there are a lot of inconsistencies in frequency of approval and the conditions for approval of disbursement of funds from the holding accounts. While all basket arrangements require effective implementation, monitoring and auditing processes, not all these processes are linked to disbursement. Developing a general standard procedural requirement for a well functioning basket fund arrangement is a key to effective execution of a development budget. 17. Streamline internal MDA procedures: While administrative procedures for triggering payment to suppliers are necessary, such measures need to be streamlined. It is understood there are some standard administrative procedures set by the Ministry of Finance and Economic Affairs (MoFEA) for effecting payment to suppliers, but some MDAs have added more procedures, which slows down the payment process and execution of the budget. Ensuring that MDAs adhere to standard streamlined procedures for effecting payments to suppliers is important for effective execution of the budget. 18. Improve the cash management system: It is important that spending agencies produce realistic cash flow plans at the beginning of the fiscal year and that MoFEA tracks what happens to these cash flows within the year. This is a key step in order to improve predictability of funds to spending units and also to improve the cash management process. 19. Improve capacity to implement capital investment projects: It is important to improve the capacity of the key sectors to design and implement quality projects, from feasibility and economic evaluation to procurement and physical implementation as well as monitoring, especially for sectors where development projects have a large component of capital spending, as in roads and energy. 20. Improve quality of expenditure flash reports (including data reliability) given that they are key instruments for tracking on a monthly basis the intra-year budget execution. This report is prepared by the Accountant General (ACGEN's) office but compiled manually from reports generated automatically from the Integrated Financial Management System (IFMS). It is recommended that this report be automated so that it is generated from IFMS directly. 21. Make more regular use of Public Expenditure Tracking Surveys to obtain better information about effective budget execution at the service delivery level. It is suggested that Public Expenditure Tracking System (PETS) be regularly carried out in the education, health, water, and agriculture sectors. The PEFAR 2009 work program has already included water and education as sectors where PETS will be executed. 22. Continue with plan to develop significant results indicators for the budget and restructure budget programs to be aligned to such indicators. It is important that MDA and LGA budgets and MTEFs are organized by programmatic approach, with programs drawn from the Strategic Plans (SP) and result indicators linked to programs. This will make it easy to link amount of resources and outcome/results achieved. 23. Make the discussion of results indicators a central element of the sector budget scrutiny. This is a key step to ensure that sectors link resources to results or outcomes, since sectors will have to justify use V

12 of the previous year s resources in terms of results or outcomes achieved before making a request for the following year. 24. Introduce key efficiency indicators along with results and quality indicators in budget management. It is important to introduce efficiency indicators in order to ensure that scarce budgetary resources are spent where more results can be achieved. 25. Inequity among LGAs in terms of resource allocations, especially for education and health, remains wide. Although the government has a formula for allocating recurrent block grants, which are largely for PE, some LGAs receive significantly less or more than their fair share. These inequities in resource allocation have also translated into significant differences in outcome indicators among LGAs. In order to resolve inequity in public service delivery, like education and health, the government would need to attract and retain staff in underserved LGAs. Integrated Assessment of PFM Performance 26. Credibility of the budget: The budget has remained credible over the recent past, with improved revenue mobilization and expenditures contained. Significant improvement in revenue collection is the result of improved economic performance, continued tax policy and administration reforms, and a conservative revenue projections policy. Expenditures have been contained, although at substantially less than original estimates, especially in the categories of goods and other services, and domesticallyfinanced development expenditures. The variations between actual expenditures and originally approved budgets, depending on the year, have been greater when looking at the expenditure composition within MDAs. In a relevant matter, payment arrears that in 2005/06 were minimal have increased in 2006/07 and 2007/08. Although there are reasons for variations between budgets and actual expenditure, it is important that the government stick to its expenditure plans for the budget to remain credible and achieve intended results. 27. Comprehensiveness and transparency: The budget preparation and documentation process is extensive and is supported by Planning and Budget Guidelines (PBGs) issued by the MoFEA to the MDAs and separately to the LGAs. Transfers to the Autonomous Government Agencies (AGAs) are recorded in the budgets of the MDAs, but the remaining part of the AGA budgets is not part of the budget documentation, as the AGA budgets are approved by their own authorities established by individual laws. Some AGAs might pose fiscal risks for future budgets in the form of a need for increased transfers resulting from commitments that are not made known to the MoFEA. Furthermore, it is not known to what extent potential fiscal liabilities created by the Public Enterprises (PES) are taken into account in fiscal planning, as these are not highlighted in the government budget documentation. 28. Public access to fiscal information, in both the budget presentation and execution phases, is good. Fiscal information in both budget documentation and execution reports is provided through several means, including the media and some government websites. However, the coverage and details of information is limited in the budget execution phase. For example, the actual composition of the MDA budgets is not published in the course of the year; these details are even excluded from the government final accounts, as the final accounts and government financial statements are very brief summations of government transactions. The lack of an internationally-accepted, functional classification in the budget, as well as the presentation of recurrent and development expenditures in different formats, reduces the value of the budget documentation. 29. The allocation of all types of transfers to LGAs is guided by certain measures, but because the transfers are conditional, sectoral policies are also taken into account. Budget ceilings, along with detailed instructions and technical manuals, are issued to the LGAs three months before the beginning of the fiscal vi

13 year. However, the budget ceilings are subject to change as the government budget is finalized. Due to the conditional nature of the transfers, such changes are unavoidable Policy-based budgeting: Several policy papers and technical documents are prepared to support the budget preparation, including macro-fiscal analysis papers, the National Strategy for Growth and Reduction of Poverty (NSGRF'), MTEF, and more recently, a document called Strategic Budget Allocation System (SBAS), to help establish MDA budget ceilings. However, there is weak linkage among medium-term sectoral policies and annual budgets. Although the MoFEA is responsible for the preparation of both recurrent and development budgets, these two budgets are technically not integrated, and the absence of a functional classification in the budget does not help improve the policy base of the budget. A clear link between long-term sectoral planning and strategies and the budget formulation process is missing. There is a clear budget calendar, but the budget is normally submitted to the National Assembly just a few days before the beginning of the fiscal year. 31. Predictability and control in budget execution: Tanzania continues to improve tax policy and administration as well as procurement. In spite of good tax collection performance, as well as increased and timely budget support from development partners, budget execution and implementation of government operations face continuous uncertainty. It is widely believed that this stems from the monthly cash rationing system that limits timely purchase of goods and services in the MDAs. The weak cash management procedures employed in the process of monthly budget allocations hampers a meaningful budget execution. Some unbudgeted operations may be initiated in the course of the fiscal year in an MDA budget, thereby crowding out other MDAs' spending plans. The bulk of recurrent MDA expenditures are salaries and wages, but payroll controls are weak and difficult to manage. Also, internal controls and internal audit functions in the MDAs remained weak and challenging. 32. Accounting, recording and reporting: Although the IFMS has been very useful, as a central payment and recording system and, by itself cannot improve the back-end accounting processes, and cannot deliver a full accounting reforms. For example, reconciliation of accounting books with their corresponding bank transactions seems to be very weak and faces long delays. Accordingly, the quality of accounting in the MDAs is questionable, and external audit reports continue to refer to this problem. The government changed its accounting standards in 2007/08, but its financial statements are yet to include a full set of asset and liability data. 33. Although MoFEA has initiated and implemented some expenditure tracking surveys on the final service delivery units, its methodology and reports have not been made public. The lack of publication of in-year budget execution reports by the MDAs, as well as the summary style coverage of the annual financial statements, remain problem areas. Fiscal reports are helpful and are published regularly, but they do not refer, by their nature, to any detailed revenue or expenditure data. Both the in-year and year-end reporting need to be improved and the corresponding data need to be published. 34. External scrutiny and audit: Significant improvement has been made in terms submitting external audit reports to legislature in timely manner, within nine months of the end of fiscal year. However, there is little time available for the National Assembly to scrutinize these reports, and even with the little time available, scrutiny is much delayed. A new external audit law, which was passed by the National Assembly and made public in a September 2008 gazette, is expected to bring about some future improvements in the external auditing task. Delays in the reporting of the Public Accounts Committee (PAC) have prevented the follow-up of audit concerns, and the MoFEA does not respond to the PAC findings. The work of the National Audit Office (NAO) has significantly improved, but there is need for further improvements, especially value for money auditing. vii

14 Assessment of the Impact of PFM Weaknesses 35. Tanzania s overall budget performance and fiscal discipline in the context of economic growth and macroeconomic stability remain good. The legal aspects of PFM have also been well addressed in recent years. However, the PFM system processes face a number of challenges. Although the weaknesses have already been identified and the DP has been providing technical assistance in a number of areas, further attention is needed by the authorities to overcome the PFM related problems in a more systematic manner. 36. There are still some serious challenges with regard to the PFM. The challenges includes engaging the legislature in the budget process, the low quality of budget classifications, the lack of a realistic resource-supported MTEF, wider goaldtargets without adequate financing scenarios, and the lack of full integration of recurrent and development budgets. There is an urgent need to improve quality of budgeting and bring back credibility to the budget as a firm government financial and operational plan. 37. Predictability and control in the budget execution is still weak. The uncertainty in availability of funds for the MDAs is an example of the lack of predictability, especially for the development budget. Due to the persistence of modified cash rationing, the MDA request for cash releases cannot always be met, resulting in difficulties in implementing their policies as planned. On the other side, areas of concern include the ineffectiveness of payroll controls and insufficiency of internal controls and audit in nonsalary expenditures in the MDAs. Internal audit is still weak, although improvement efforts are underway. There is little available information on the delivery of resources to service delivery units. Prospects for Reform Planning and Implementation 38. Further reforms need to be undertaken in a number of areas, including strengthening the budget preparation process, improvement in budget execution cash management, and accounting and reporting systems. The MoFEA has recognized the weaknesses and redesigned its PFM reform strategy (PFMRP 11), which was approved in July The new PFM reform strategy demonstrates how the MoFEA intends to address the observations of key diagnostic reviews in the wide-ranging PFM cycle. However, a carefully phased implementation with interval outcomes of the plan will be critical to measure the impact of reforms in the new round. Phasing the actions and harmonizing the reforms with the government capacity and commitment to reform will remain key factors in the success of the new PFM reform strategy. Status of PEFAR Capacity Building Work Program 39. The PEFAR capacity building is part of the collaborative effort between the government and development partners to strengthen public financial management in Tanzania. Initiated as one of the major outputs of the PEFAR work program in 2008, it aims to strengthen budget preparation in order to better link resource allocations and MKUKUTA priorities. The major aim is to enhance expenditure efficiency in order to realize growth and social impacts of public expenditure. Three major priority areas for intervention in capacity building were identified as strengthening (i) macroeconomic model (MACMOD) in order to improve its capacity to simulate growth and poverty reduction policies as well as to strengthen macro policy framework for planning and budgeting; (ii) the MTEF in order to facilitate and optimize the resource allocation system; and (iii) sector strategic plans (SPs) and sector MTEFs with a programmatic approach in order to enhance the linkage between SPs, MKUKUTA, MTEF and results/outcomes. viii

15 40. Considerable achievements have been obtained in most of the priority areas of intervention since the launch.of capacity building activities in September Among the achievements are (i) a vastly improved and user-friendly macro simulation tool (MACOD-TZ); (ii) enhanced human capacities (about 25 officials) for macro policy simulation and analysis using MACMOD-TZ; (iii) enhanced macro framework paper for the 2009/10 Budget Guidelines; (iv) first edition of Budget Background and Medium Term Framework document; and (v) an organized database for annual budget analysis as well as a stronger budget analysis department. Achievements were also realized in other priority areas of intervention, like experimental exercise of upgrading the MTEF. These include development of a simple Excel-based MTEF optimization model and the revised SBAS Macro, which provides a stepping stone for a full upgrade of the MTEF. In terms of building a programmatic dimension in the SP and the MTEF, an experimental framework developed for the health sector proved very successful. However, the framework needs to be revised in order to align it with the new health sector SP (HSSP 111) and revise the sector MTEF accordingly Although some major achievements have been obtained in capacity building program, priority areas for intervention have also been identified for These areas are (i) strengthening of MACMOD- TZ and the Macro Policy Framework Paper, taking into account the world economic crisis as well as new datddevelopments in Tanzania; (ii) initiating MTEF upgrades (both central and sector) as well as program budgeting starting with the Ministry of Health and Social Welfare; and (iii) preparing the second edition of the Budget Background and Medium Term Framework for 2009/ /12. ix

16

17 PART I: BUDGET ANALYSIS 1

18 1. INTRODUCTION 1.1 This part of the PEFAR presents an overview of the main findings and key messages of the various sector and thematic reports carried out as part of the Budget Analysis of the FY07/08 PER cycle. The analysis for FY07/08 uses both original expenditure estimates approved by the parliament and actual expenditures data, while the analysis for FY 08/09 uses the expenditures estimates approved by parliament (original estimates) only. As much as possible, the analysis is cast in a medium-term perspective. 1.2 The key focus of this part is on the alignment of the budget to the MKUKUTA, the National Growth and Poverty Reduction Strategy of Tanzania. The Rapid Budget Analysis addresses to what extent the national budget is an effective financial instrument to implement the policies and achieve the objectives of the MKUKUTA. The analysis of alignment is carried out along the three following directions: 1.3 The macro-economic dimension is relevant to the MKUKUTA inasmuch as a positive macroeconomic environment, which is supportive of sustainable growth and low inflation, is by itself an important ingredient in achieving the MKUKUTA objectives. In this respect, the report discusses the impact of the budget on key macroeconomic variables, such as the inflation rate, the GDP growth rate, and the interest rate. The macro-economic environment in turn is relevant to MKUKUTA because it is a major determinant of the resources envelope available to the budget and influence expenditure allocations. 1.4 The analysis of budget allocation looks at how the budget is distributed among different MKUKUTA sectors and clusters as well as the composition of the budget in terms of wages, current expenditures, maintenance, and public investment. It discusses whether the sectoral and economic composition is consistent with the policies and objectives of MKUKUTA and other sector policies. This analysis is carried out with respect to both the execution of the FY07/08 budget and the estimates of the FY08/09 budget. 1.5 The last stage of the analysis addresses the question of whether the budget achieves its stated impact on the ground, in terms of providing effective public services to the population in a way that is both equitable and efficient. This part of the analysis is still sketchy, as the evidence is scarce. It builds on whatever performance reports have been produced by MDAs and on other sources, such as the Household Budget Survey. It will be developed further in the years to come, starting with the PER FY08/09 cycle, which has a strong focus on results in health, education and water. 2

19 2. MACRO-FISCAL POLICY AND MACRO-ECONOMIC BUDGET FRAME WORK 2.1 This section analyzes two issues: (i) the impact of the budget on the macro-economic situation, which is labeled as the macro-fiscal policy issue; and (ii) the constraints imposed by the macroeconomic situation on budgetary choices, which is labeled as the macro-economic framework of the budget. 2.2 Within the context of a broadly stable macro-economic framework, emerging vulnerabilities faced by Tanzania involve rising inflation and some pressures on the external balance of payments, making the conduct of macro-fiscal policy particularly sensitive. 2.3 Inflation has been on a rising trend above the government target of 5 percent since 2004 and significantly accelerated in 2007 and 2008 as internaiional cornmod& inflation added to domestic pressures (Figure 1). Figure 1: Year-on-year monthly CPI inflation rate, new series 9.0% 7.0% 5.0% 3.0% Source: NBS 2.4 The inflation experienced by families, measured in terms of a basket of food and non-food basic goods by the household budget survey (HBS), appears to have been also significantly higher than market prices revealed by the Consumer Price Index (CPI). The cumulative consumer inflation measured by the HSB between 2001 and 2007 is around 98 percent (an almost doubling of prices), while the CPI cumulative inflation for the same period measures at 41 percent. 2.5 All together, it is clear that rising inflation represents a serious challenge for the government, including for fiscal policy, as it tries to preserve hard-won macro-stability and protect the purchasing power of the population. 2.6 The Balance of Payments (BOP) of Tanzania is currently strong, with a high level of foreign reserves, while a more resilient outlook for gold exports compared with other commodities, as well as tumbling oil prices, will be helpful going forward. Over the last five years, however, the composition of the BOP has become a bit more vulnerable, with the current account deficit exceeding the levels of foreign 3

20 aid, making the BOP dependent on private inflows, in particular Foreign Direct Investment (FDI), as shown in Figure 2. The current international economic crisis could have an impact on the level of FDI inflows, as well as on tourism receipts. Figure 2: Current acccunt -~ and foreign aid B)oom ?ao4m5 2006M7 Source: IMF staff report MACRO-FISCAL POLICY: THE IMPACT OF REVENUE AND FINANCING ON THE MACRO CONTEXT Domestic revenue 2.7 Domestic revenue has witnessed a remarkable sustained upward trend since the early 2000s. This development is particularly important as it provides the foundation to reduce aid dependence and improve domestic accountability. Domestic revenue is also a source of financing that does not have an inflationary impact through monetary aggregates, although potential tax pass-through to the consumers is inflationary. Table 1: Domestic revenue: trend over (in percent of GDP) 2004l Budget Actual Budget Actual Budget Actual Budget Actual Total Revenue Tax Revenue Import Duties VAT Excise Duty Income Tax olw PAYE olw Corporate o 1.o Other taxes 1.o o Nontax Revenue o o Source. Tanzania Authorities (TU) 2.8 At around 16% of GDP in , and projected at around 18.5% of GDP in , domestic revenue in Tanzania is still below the average for Sub-Saharan Africa, and not high by international standards (Table 1). This indicates that the level of taxation in Tanzania is not excessive, as far as its macro-aggregate impact is concerned, and that there is scope for further increase. A 2006 FIAS report assessed the impact of the taxation structure on economic growth as follows: the income tax code and Value Added Tax (VAT) legislation are broadly appropriate and conducive to growth. Problems remain with implementation, the extent of exemptions, and local government taxation. The marginal effective tax 4

21 rate in Tanzania is comparable with those in South Africa, and Zambia... it is excessive only for non- VAT registered medium and small scale enterprises, for which it can be as high as 50.5%. 2.9 The main risk is that the tax base is relatively narrow. The increased performance has been reliant on higher contribution by the large taxpayers (Table 2). Expanding the tax base is important in order to avoid excessive concentration on the narrow base - large tax payers. The government s plans to review tax and non-tax revenue in mining and natural resources are welcome, as they allow the tax base to broaden. Table 2: Contribution of TRA Departments to overall tax collection (in percent share) DeDartment Actual Actual Actual Actual Domestic revenue Customs and Excise O 44.0 Large Taxpayers Gross Collection Source: Tanzania Authorities (TRA) Foreign financing 2.10 In 2007/08, the share of foreign aid in GDP declined for the first time in the last five years. (Figure 3). The Government has projected a further decline in FY08/09, both as a share of GDP and in absolute value. Figure 3: Total aid inflows relative to domestic revenue m Domstio Revenue (% GDP) I Forllbn Aid (% QDP) Source: Budget Frame 12/10/2008, MOEA Foreign aid in FY07/08 was composed of 66 percent in grants and 34 percent in concessional loans. The FY08/09 budget projects an increase in concessional loans to 40 percent; given the nature of concessional lending, this should not significantly affect Tanzania s debt sustainability profile The slight reduction in foreign aid potentially helps to control inflation, as foreign currency funding of the budget increases the monetary base for an equivalent amount net of the import content of expenditures and needs to be sterilized The reduction in foreign funding is, however, problematic with respect to the potential vulnerabilities in the balance of payment. These could be exacerbated in case of a potential negative impact of the current global economic crisis on flows of foreign direct investment, as well as lower export prices (which are likely to be offset by higher gold prices and/or lower oil prices). Sector Study of the Effective Tax Burden in Tanzania, May 2006, FIAS, the World Bank Group. 5

22 Domestic borrowing 2.14 Domestic borrowing (excluding Central Bank financing) is a non-inflationary financing source, as it does not increase the monetary base and tends to raise interest rates. A drawback of domestic financing compared with foreign financing is its cost. Foreign financing, when it is fully sterilized, has a cost equivalent to the T-bill interest rate on less than the full amount (since part of the expenditure flows back in foreign currency through imports), while the interest cost of domestic financing is borne on the full amount. The cost advantage of (sterilized) foreign financing is lower, the lower the import contents of expenditures and the lower the domestic interest rate. Thanks to vastly improved T-bill auctions and monetary management by the Bank of Tanzania (BOT), over the last year, in parallel to rising inflation, real interest rates on T-bills have become negative. Therefore, use of domestic borrowing to smooth expenditure pattern' in the current conditions would be an option for the authorities. MACRO-FISCAL POLICY: IMPACT OF EXPENDITURES ON THE MACRO CONTEXT 2.15 In a context of rising inflation, a large increase in public expenditure can put additional pressure on aggregate demand, depending also on the funding modality and the composition of expenditures. In such an environment, an expenditure profile tilted toward investment, and less toward current outlays high in domestic content, would be preferable, as it would tend to increase productivity and reduce demand pressure on the non-tradable sector. Higher public investment is also necessary to reach the government's objective of sustained annual GDP growth between 8-10 percent, as few countries have achieved such a feat without an aggregate level of investment reaching over 30 pecent of GDP, which requires a significant contribution from public investment The approved budget framework for 2008/09 aims at increasing recurrent spending by 1.3 percentage points of GDP, from 17.2 percent to 18.5 percent of GDP, while development expenditure remains stable at 9.8 percent of GDP. Over the medium term (2009/11), as aid flows are projected to abate, development expenditures are expected to further decrease to reach 7.7 percent of GDP by 2010/11, while recurrent expenditures will stabilize around 17.5 percent of GDP. The trend in projected development expenditure is surprising given the ambition of the Government to sustain the projected higher level of growth The current classification of expenditure does not give a fully accurate deconstruction of Government spending. A more accurate classification shows that percent of the development budget is composed of non-capital spending, mostly on goods and services (accounting for 22.8 percent, of which the purchase of medical supplies accounts for 9.3 percent). Similarly, 7 percent of the recurrent budget includes expenses linked to infrastructure or equipment Based on a more rigorous classification, in FY08/09, 75 percent of expenditures are recurrent against 73 percent in FY07/08 (Figure 4). This relatively low level of investment and the declining trends are worrying signals. 6

23 Figure 4: Economic classification of the 2008/09 Budget Other, es Smd i -, T, Capital, 4.8"/0 Household Furtuture, 1.O%\ Equrpment, 3 7% Wages, 35 9% Current Transfe Goods and Services, 27.8% 2.19 In agriculture, for instance, the share of capital expenditure is much smaller than that of development expenditure. It accounts for only 34.8 percent of total sector budget compared with 46.9 percent of development expenditure. Fifty seven percent of the development expenditure is eventually recurrent expenditure, spent on the provision of goods and services. This indicates that little capital investment is going into the agriculture sector. Moreover, the actual or spent capital expenditures are likely to be much smaller than approved, partially due to the delays in feasibility studies and project implementation problems, as was the case in 2006/07, which further hampers the infrastructure development for agricultural growth in Tanzania In health, an accurate classification shows that recurrent expenditure accounts for 89.8 percent of the total health sector budget, of which wages and salaries account for almost half of all recurrent spending. Goods and services account for 45.2 percent of the total health budget, of which 83.0 percent comprises medical supplies. Therefore, capital expenditure accounts for only 10.2 percent of the total health sector budget, of which 8.1 percent is for investment (see Annex 1). While Government has stated that this year will be focused on addressing the human resources shortage in health, there will need to be a substantial increase in the capital expenditure in order to fulfill the other priority that focuses on rehabilitation, upgrading and establishment of facilities at the primary level The situation is even worse if one looks at actual expenditures. A look at execution rates reveals not only that the development budget shows execution gaps that are much higher than recurrent expenditures, but that some very high execution gaps (very low execution rates) are concentrated in the three key infrastructure ministries (Table 3). This is true even if some of the execution gaps are due to under-reporting of donor-funded projects (direct to project funding modality) through the Dummy Exchequer System. 7

24 ~ 5.9 Table 3. Selected MDAs budget execution gap (in percent) I Vote Ministry 1 Recurrent Develop. Total Recurrent Develop. Total I 58 Min Energy & Minerals I i Min Water & Irrigation Min Infrastructure Dev. I ~ Min Health & Social We1 Min Education & Voc Tr Min Agriculture & Food Sec Source: MoFEA, Expenditure Flash Reports 2.22 In this context of low capital expenditures, the large increase in the wage bill in the new budget, on the heels of the minimum wage increase, is particularly worrying (Figure 5). While the need to increase the minimum wage was widely recognized, such a big increase raises the risk of putting Tanzania in a vicious circle of increasing wages, rising inflation and lowering economic growth. The overall wage bill, including allowances and hidden wage costs, stands at 9.5 percent of GDP in the , against 9.1 percent in Figure 5: Change in budget shares between and O% -2.0% % 2.0% 4.0% 6.0% 8.0% 2.23 The increase in the wage bill appears inconsistent with Medium Term Pay Policy (MTPP). The optimistic scenario of the MTPP projected an increase in the wage bill in the range of 5 percent of GDP by 2007/08. The actual progression to about 7.5 percent of GDP in is way out of medium-term pay targets. The picture looks even worse when allowances are added to basic wages, which increase the wage bill to 9.1 percent of GDP in 2007/08 and 9.5 percent of GDP in Basic salaries represent almost 84 percent of the total wage bill in 2008/09, which is 3 percentage points higher than the share. This implies slight decrease in allowances and hidden wage assimilated costs. This trend can be explained by the deliberate move to consolidate wage related items into basic wage. However, allowances and hidden wage assimilated costs remain high at 16.4 percent of the total wage bill and equivalent to 5.5 percent of the total government budget, also equivalent to 1.6 percent of GDP The increase in number of civil servant employees (new hires) has little impact on the increase in total basic wages in the 2008/09 budget. About 70 percent of the increase total basic wage is due to salary increases, while less than 10 percent of the increase in the total basic wage bill is due to new hires. Note: Execution gap is sum of release and spending gap, where release gap is the difference between original budget estimates and exchequer release: and spending gap is the difference between exchequer release and expenditure. 8

25 Table 4: Deconstruction of increase in basic wages in 2008/09 I BillionTanzanian 1 ~ Item i Shillings (TSh) % of Total % of GDP 2007/08 wage arrears i ! I 2008/09 wage arrears 2 i I new hires i i i 2008/09 normal salary adjustment I 0.9 j I Other (incl. Social benefits, Parastatal wages, I i i etc) I j i Total j i 10 I 2.3 Source: MoFEA, PO-PSM, & Author s Calculation THE MACRO ECONOMIC FRAMEWORK OF THE BUDGET: OVERVIEW 2.25 The following paragraphs review the macro-framework of the 2008/09 budget from a mediumterm perspective, with a focus on two kinds of issues: the influence of the macro-economic situation on the level of the various budget aggregates, and the reliability of the projection of the budget aggregates. It is important to note that the world is in major economic and financial crisis, which may have an impact on the macroeconomic framework for Tanzania. Box 1 Economic and Financial Crisis and Fiscal Stimulus3 The world economy is entering a major downturn, caused by the most dangerous shock in mature financial markets since the 1930s. Developing country growth will fall sharply in 2009 through several channels. Exports will decelerate as a result of falling demand in the advanced economies. A fall in remittances to developing countries has already begun. Primary commodity prices have plunged at the prospect of falling world demand. This reversal, while providing a ray of light for commodity importing nations in the wake of the huge price increases of the past 1-2 years, also poses the need for adjustment in commodity exporters, which until recently were enjoying boom conditions. Private portfolio and bank lending flows to developing countries have fallen sharply, combined with extraordinary declines in stock prices, significant currency depreciation and sharp increases in external borrowing costs. Trade finance has been disrupted. Over a longer horizon, foreign direct investment may also weaken because of weaker growth prospects. There are also concerns that official aid flows could be affected by tighter budgets in advanced countries and those political pressures for more protectionists could gather force. Is Tanzania being affected by the crisis? Although Tanzania s economy is not fully integrated into the global financial markets, it is likely be affected by the crisis, like many other developing countries. Some of the ways in which it will affect Tanzania include: decline in demand and prices of exports (such as cotton, coffee, cloves, horticulture, and tanzanite); decline in FDI; decline in employment in local manufacturing industries; decline in tourism earnings; and decline in remittances. In the long run, there could also be a decline in official aid. Some evidence of the crisis has been reported, including decline in volume and price of tradition exports (tea, tobacco and cloves).the volume of three export crops declined by 7-61 percent and prices declined by 1-36 percent by end December Collection of import taxes and excise duty were short oftarget by percent by December How can Tanzania deal with the crisis? The following fiscal and monetary policy stance can be helpful to the Government of Tanzania in dealing with the crisis: The authorities have indicated that their short-term priority is to reduce inflation through a combination of monetary and, to a lesser extent, fiscal policy. If this proves successful in the next 6 months, the authorities may have to switch quickly to a more expansionary stance in the new Financial Year (FY). This includes speeding up disbursement of existing commitments, especially in infrastructure investments. It would be important to avoid an excessively pro-cyclical fiscal policy, while ensuring sustainable funding for the key development expenditures. Because of the simultaneous Balance of Payment (BOP) shock, extra foreign financing of the budget may prove important. PREM Guidance Note on the Financial Crisis, World Bank; Tanzania, Planning and Budget Guidelines 2009/ /12 9

26 2.26 The Government s fiscal framework for 2008/ /11 aims at halving the overall budget deficit (before grants) from 11.1 percent in 2007/08 to 5.8 percent in 2010/11. Domestic revenue is expected to increase by 3.8 percentage points of GDP over the next three years from 15.6 percent to 19.4 percent, while overall expenditures will slightly decrease from 26.7 percent in 2007/08 to 25.2 percent in 2009/10. Table 4 gives a summary of the structure and trend of the main fiscal aggregates over the medium term. Table 5: Fiscal Framework, 2005/ / Prel Budn. Act. Budg. Act. Budg. Act. Budg. Proj. Proj. Total Domestic Revenue Grants Total Expenditure Total Primary Expenditure O Overall Balance (excl.grants) Overall Balance (incl. grants) Source: MoFEA & Authors Calculations Macro-framework: revenue and financing 2.27 The macro medium-term projection in the 2008/09 MTEF appears optimistic on the revenue side and pessimistic on the foreign aid side The authorities have projected a strong increase in domestic revenue to mitigate long-term vulnerabilities to an anticipated decline in aid over FYs 09/11 period. In 2008/09, domestic revenue is projected to further increase to 18.5 percent of GDP. The authorities anticipate that the expansion of the tax base induced by private sector growth, the implementation of the Tanzania Revenues Authority s (TU) five-year corporate business plan and a new strategy to increase revenue collection from natural resources (including the mining sector) would lead to a 2.5 percentage point surge in domestic revenue in 2008/09 and a half point increase over the following 2 years While the growth in the revenue to GDP ratio is commendable, it represents a very challenging target. This 2.5 point annual increase seems highly ambitious against the already sizable one percentage point annual average increase in domestic revenue achieved since 2002/03. The targeted increase in May 2008, under the Policy Support Instrument (PSI) economic program supported by the IMF, was more modest, as the revenue-to-gdp ratio was expected to reach 16.2 percent of GDP in 2008/09. In the March Plan and Budget Guidelines, the authorities had anticipated a 15 percent annual increase in revenue collection, implying that revenue collection would reach 16.7 percent of GDP in 2008/09. This will require careful monitoring and a strategy on how to adjust to eventual shortfalls The government has cautiously revised aid projections over the next 3 years, anticipating aid would be almost halved by The pattern of external aid is mixed: over the last 2 years aid was significantly over budgeted; over the 2003/ /05 period aid was under-budgeted. Failure to properly assess aid flows has hampered the budget management as total expenditure was accommodating aid ups and downs. 10

27 2.31 In 2008/09, the GOT anticipates that total foreign aid inflows as a percentage of GDP would decrease by 2 points. This 4.7 percent decline is largely attributed to a decline in grants (-14.8 percent) despite an increase in foreign loans. The structure of aid inflows in terms of aid modalities illustrates a decline in budget and project support (inclusive of MCC and Multilateral Debt Relief Initiative - MDRI) and an increase in basket support as budgeted in 2008/09 relative to 2007/ Over 2009/ /11, the GOT has forecasted a further decline in aid flows and plans to scale down expenditures as of 2009/10. Over the medium term, the budget framework indicates that expenditures would be gradually scaled down from 28 percent to 25 percent of GDP as aid decreases from 9.4 percent to 6 percent of GDP between 2008/09 to 2010/11. The projected medium-term level of foreign aid flows seems likely to be higher than projected as new donor projects come on line and more robust development partner pledges are made An analysis of the medium-term reliability of the MTEF revenue projection (Table 6) bears out the following conclusions: a) Budget support is, on average, the most reliable form of funding for the budget, followed by domestic revenue. b) Both project and basket funding are very difficult to project. c) The first outer year projections in 2006/07 were more reliable than the second outer year. Table 6: Medium-Term Projection Analysis 06/07 MTEF 06/07 MTEF 007/08 MTTEF Year 2 Accuracy Year 3 Accuracy Year 2 Accuracy Domestic Revenue Aid Inflows o/w GBS o/w Projects o/w Baskets Total Expenditure o/w Recurrent o/w Development (local) Source: Budget Digest 2006/07, Budget Speech 2007/08, Budget Frame 12/10/2008, MOEA. Macroeconomic framework: expenditures 2.34 The approved medium-term expenditure framework indicates a 1.2 percentage point of GDP increase in overall expenditure in 2008/09 followed by a sharp fall of 3 percentage points in 2009/10. The lower level of expenditure follows the pessimistic assumption on aid trends (discussed above) together with a one percent annual reduction of the overall deficit between 2009/10 and 2010/ A credible medium-term budget framework would indicate how surplus aid will be saved, or shortfalls made up through domestic borrowing or spending cuts. Hence, it would be important for the GOT to develop a Priority Investment Plan (PIP) in line with MKUKUTA objectives, with clear indication of (i) programs that would be protected and financed through domestic borrowing in case of aid shortfall and (ii) a pipeline of high return.projects that could absorb any aid surplus Table 5 indicates that recurrent expenditure projections are much more accurate than local development expenditure, which may partly being explained by the fact that it is sometimes tied to donor project inflows via counterpart funding.

28 2.37 Figure 6 indicates that, over the last 2 budgets, the GOT has increased the share of local funding of the development budget in key sectors, such as roads or water, which gives the Government greater flexibility to accommodate the fluctuations in project funding. The increased funding of the development budget from local resources has also restricted possible donors from imposing their budget priorities through earmarked project funding. Without getting into the merit of the reduction in the water sector budget, the large domestic funding in 07/098 allowed the Government to reduce the water sector budget in FY08/09, notwithstanding a large increase in donor funding for the sector. Fieure 6: Local and foreign fundine in Infrastructure and Water budgets MolD Budget I UForrx/ BIOOO 1 lijo00 I WO 11?1621 I ELaciI lWO I li4000 I 360W RECOMMENDATIONS 2.38 Improve the reliability of the macro-framework by sustaining efforts to develop capacity in macro-fiscal policy and macro-modeling. The upgraded MACMOD-TZ should serve as an important tool for developing the credible macro framework as well as budget frame The share of capital investment in the budget has stagnated at around 25 percent, although the development budget is around 35 percent of the budget. In order to reach the government target of economic growth of 8-10 percent over the medium term, the share of private investment has to be complemented by a significant share of public investment. It is important that the government improve the growth orientation of the budget by increasing the share of public investment The government should strive to reduce the inflationary impact of the budget by controlling wage growth (including allowances) in line with medium-term pay policy. Ln the 2008/09 budget, the share of the wage bill (including allowances) increased to 9.5 percent of GDP, which is way out of the mediumterm pay policy targets. The new medium-term pay policy, which is being developed, should try to ensure that wage bill growth is consistent with the macroeconomic framework The government should increase speed and depth of structural reforms on the supply side to increase the economy's capacity to absorb aid. This particularly important for infrastructure sectors like transport, water, and energy The government fiscal framework targets a 2.5 percent of GDP increase in domestic revenue in 2008/09 compared with 2007/08. However, given the world economic and financial crisis, this target seems ambitious. This suggests that there is a need for improved forecasting of domestic revenue Given the observed unreliability of aid, it is important that the government develop a Priority Expenditure Plan in line with MKUKUTA objectives, with clear indication of (i) programs that would be protected and financed through domestic borrowing in case of aid shortfall and (ii) a pipeline of high return projects that could absorb any aid surplus. 12

29 2.44 The government needs to engage with Development Partners over medium-term projected level of financial assistance, with a view toward avoiding abrupt reduction The government could consider using domestic financing for a limited amount of expenditures smoothing, as long as real domestic interest rates remain low The government is encouraged to continue providing sizeable local funding to the development budget in key sectors to reduce vulnerability. 13

30 3. BUDGET ALLOCATION ALLOCATION BETWEEN MKUKUTA AND NON-MKUKUTA SECTORS 3.1 The share of MKUKUTA spending, net of MDA wages, has decreased since last year. To assess properly the share of MKUKUTA against non-mkukuta expenditures, it is important to consider that the GOT has exclusively allocated to MKUKUTA the other charges or development expenditure categories, therefore excluding transfers to local authorities and wages. 0 Basic salaries and pension paid, for example, to teachers, medical staff (nurses, doctors) and judges is excluded from MKUKUTA, although they are essential priority expenditures to deliver social services. As more MKUKUTA infrastructure, such as schools and health centers, is built, the GOT must recruit more personnel to ensure adequate service delivery. 0 0 Transfers to LGAs, which represent about 20 percent of total budget, are not coded according to the MKUKUTA spending and are therefore considered non-mkukuta spending. Also, a large amount (4.9 percent of total expenditures) is budgeted this year in Treasury - vote 21 (classified as non-mkukuta) to pay wage arrears and adjust salary along the revised pay scale. 3.2 Taking into account these adjustments, two different breakdowns have been calculated; one includes only the LGA spending and excludes MDA wages (the basic ones as well as the wage adjustment and arrears included in the Treasury - vote 21); a second breakdown includes LGA expenditures and the whole MDA wage bill (with the arrears and adjustment in vote 21 redistributed to their respective MDAs (Table 7). In the first case, a reduction in the budget share of MKUKUTA is noticeable. In the second case, there is actually an increase in the share of MKUKUTA by half a percentage point. As mentioned above, about 70 percent of the increase in total basic wage in FY07/08 is due to salary increases rather than new hires, so the increase is MKUKUTA spending in the second definition reflects this fact. Table 7: MKUKUTA and non-mkukuta expenditures (shares of total) 2007/ /09 Incl. LGA, excl. MDA wages MKUKUTA Non-MKUKUTA Total Incl. LGA and MDA wages MKUKUTA Non-MKUKUTA Total Source: MoFEA & Author s Calculations 3.3 The diminution in the MKUKUTA share on non-wage expenditures is of concern, particularly given the still modest progress toward poverty reduction, as evidenced by the recent Household Budget Survey, which would advocate for additional efforts toward the implementation of MKUKUTA. On the other hand, because of the large growth in the overall budget expenditures, the slightly reduced MKUKUTA share translates into a sizeable nominal increase in the MKUKUTA expenditures. 14

31 Allocation to key non-mkukuta spending agents (votes) has remained stable with the exception of Treasury and Ministry of Finance and Economic Affairs, (Table 8) which have increased substantially, even after the salary adjustments have been netted out from allocation to the Treasury. Table 8: Selected Non-MKUKUTA budget allocations (shares of total) Vote code Vote name 2007/ /09 Difference 22 Public Debt Police Force Prison Service President's Office Foreign Affairs 1.o 1.o 38 Defense National Service National Assembly 42 Fund Ministry of Defence 1.o & 50 Treasury & MoFEA Sub-total Total budget Source: MoFEA & Author's calculation ALLOCATIONS BETWEEN MKUKUTA CLUSTERS 3.5 The Government has made commendable efforts to use the three clusters of the MKUKUTA as the strategic level of budget allocation. However, in the past three years there has not been a clear direction toward strategic MKUKUTA allocations. The uncertainty seems to be between a 40:40:20 allocation across clusters, and one more skewed toward supporting economic growth resulting in a 50:30:20 distribution (Table 9). For instance, the FY06/07 guidelines projected a 49:30:2 1 allocations for FY07/08, but the FY07/08 budget guidelines went for a 42:41:17 profile. Similarly the FY07/08 guidelines projected a 43:37:20 allocations for FY08/09, but the FY08/09 budget guidelines establish a 48:34: 18 distributions. Table 9: Budget Guidelines: Cluster Allocation i I BG 06/07-08/09 BG 07/08-09/10 i BG 08/09-10/11 / 06/07 07/08 08/ /08 08/09 09/10 08/09 09/10 10/11 Cluster I i I Cluster I ~ Cluster 111 I ~ Note: Cluster I: Growth of the economy and reduction in income poverty Cluster 11: Improvement of quality of life and social being Cluster 111: Governance and Accountability Source: MoFEA, Plan and Budget Guidelines (various) , Votes 21 & 50 have been combined because they used to be one vote before 2008/09. Wage adjustments for major sectors in 2008/09 have been taken out of vote

32 3.6 The difficulty in strategically using the MKUKUTA clusters for budgeting is further complicated by the classification of MKUKUTA clusters, which, as mentioned above, does not include basic salaries and transfers to local government. 3.7 Accounting for transfers to LGA but excluding MDA wages provides a different view of MKUKUTA cluster allocation, in particular raising markedly the share of cluster 2, at the expense of cluster 1 and 3, as well as showing a substantial stability of the shares of the various clusters (Table 10). Allocations /09 YO of Mkukuta I YO of Overall YO of Mkukuta I YO of Overall I I Cluster I 32.1%! 22.7% 32.8%! 23.2% I i Cluster I1 41.0% I 29.0% 28.0% I Cluster % 9.2% I 7.9% I I Cross Cutting 52% I 3.7% 4.2% I 3.o% Total MKUKUTA 91.3% j 64.5% 87.6% I 62.0% Non-MKUKUTA I 35.5% I 38.0% Overall I I I 10% I 10% 3.8 Similarly, the inclusion of the MDA wage bill rebalances the shares in favor of Cluster I1 and show a substantial stability of the overall shares (Table 11). Allocations 2007/08 YO of Mkukuta I YO of Overall I Cluster I 33.1% I 23.4% Cluster I1 45.0% 3 1.8% Cluster % 11.7% Cross Cutting 5.4% 3.8% Total MKUKUTA 10% i 70.6% I Non-MKUKUTA I Overall I 29.4% 10% YO of Mkukuta 34.1% j YO of Overall 24.1% 45.5% 32.2% 16.0% 11.3% I 4.4% 3.1% 10% I I I I 70.8% 29.2% 10% ALLOCATION BETWEEN MAIN SECTORS 3.9 The only MKUKUTA sector that has seen its share of total expenditures increase significantly is Transport, which increased by 1.5 percentage points (Figure 7). The Heath sector budget has remained broadly stable, while both Education and Agriculture shares have decreased by around 1 percentage point. The Water sector experienced the most drastic reduction of around 2 percentage points. Altogether these changes imply a decrease of 2.5 percentage points for these MKUKUTA priority sectors, which have been allocated to other MKUKUTA sectors The increase in the Transport budget conceals a much sharper increase in the Road sub-sector budget, which has almost doubled over the last 3 years, moving from 6.5 percent of the total budget in 2006/07 to 8.7 percent in 2007/08 and 11.5 percent in 2008/09. Consequently, the Transport Sector Investment Plan (TSIP) requirements have largely been met, at a level of 96 percent in 2007/08 and 101 percent in 2008/09. The requirements in TSIP are US$ million and US$ million in 2007/08 and 2008/09, respectively. In comparison, the allocated budget for Roads is respectively US$ 16

33 621 million (719, million TSh) in and US$ 704 Million ( million TSh) in The education sector share has experienced its second decrease in two years, coming down from 20 percent of the total budget in to 19.4 percent in 2007/08 and 18.5 percent in Figure 7: Sector Share for selected MKUKUTA sector Education Health Water Agriculture Roads Fiscal Years Education Health Water Agriculture Roads (Sector Share of Total Spending) :007/ % 10.6% 5.4% 5.2% 12.8%,008/ % 10.8% 3.3% 4.0% 13.1% (Sector Share of Total Priority Spending) :007/ % 19.9% 10.1% 9.7% 24.0%,008/ % 21.7% 6.7% 8.1% 26.4% (Sector Share of Total Spending excl interest) :007/ % 11.2% 5.7% 5.4% 13.5% :008/ % 11.2% 3.4% 4.2% 13.6% qote: Priority spending is defmed in this table as Road, Water, Agriculture, Education and Health Agriculture ALLOCATION WITHIN SECTORS 3.12 The specific Government priorities within the agriculture sector in Tanzania (Agricultural Sector Development Strategy) include (i) supporting the local development process to increase access to services and expand agricultural investments; (ii) increasing agricultural productivity through the adoption of, and investment in more productive technological packages; (iii) increasing the area under irrigation and promoting greater water use efficiency; (iv) improving access to inputs and appropriate technology; and (v) maintaining food security through a national food reserve. Overall, the sector budgets are largely in line with its stated priorities, as indicated in Figure 8, but it is not possible to make a more accurate conclusion, since the cost of the sector strategy has not been established There are, however, several emerging trends that require attention. During 2006/ , the budgets for input programs and grain reserves increased, while the expenditure for research and training, policy and planning, and animal disease control decreased. The latter expenditures are growth-enhancing and are for core public goods, without which not only will the performance of the sector suffer in the long-run but, moreover, the effectiveness of input use, increasingly promoted by the government, is likely to be constrained. 17

34 Figure 8: Functional composition of the 2008/09 approved agriculture sector budget Livestock development 2% Fisheries = I 7 Animal disease control Comdty marketing brigation 2% Food securityigrain reserve 6% I Research and Training 9% Admnistration, information policy and planning and CO-nlcatlon 3% 3% Local Government transfers and support 40% Education 3.14 The distribution of the education budget shows around 66 percent of the budget going to pretertiary education (Figure 9). In the 2008/09 budget, we observe stability of the primary education allocation, a decrease in the secondary education funding, and an increase in tertiary education funding. The decrease in secondary education allocation reflects the fact that most construction work for secondary education infrastructure has been completed, hence there is less allocation for development budget. Conversely, the increase in the allocations for higher education reflects the extent of construction work in higher education infrastructure (especially University of Dodoma) as well as increased allocations to Higher Education Loans Board for student loans due to expanded admission at higher learning institutions. However, it is not clear how this evolution matches the stated education sector priority to achieve enhanced enrollment, attendance and completion of preand primary levels. I Figure 9: Composition of total education expenditure by Levels Health 3.15 In health, it is difficult to deconstruct the budget precisely according to the different levels of care or according to the key priority programs of the government. It appears that the health budget is 18

35 more or less evenly distributed between centravhospita1 care, primary care and medical supplies (Table 12). Depending on the distribution of medical supplies across levels of care, the orientation and strategic alignment of the health budget could look very different. Table 12: 2008/09 health budget by level of care and type of expenditure (share) Ministry (central) 21.l0/0 Administration 4.8% Other priorities 16.3% Hospitals 13.6% Hospitals: national referral /specialist 6.5% Hospitals: regional; 3.1% Hospital: district 4.0% Medical supplies 32.7% Local/regional expenditures (primary healthcare) 32.6% Total 10% Water 3.16 The internal consistency of budget allocations with government policy objectives in the water sector is improving. The budget reflects changes in the role of the government, both at the level of the Ministry of Water (MOW) and local governments - with more funding allocated to local governments in line with a more decentralized service delivery (Figure 10). In the MOW, the departments focusing on coordination, policy and planning and water resource management have seen their budgets increase sharply. Figure 10: Government allocation to rural & urban water sub-sectors (TSh. millions) 120, ,000 80,000 60,000 40,000 20,000 Urban Rural Other 1 Devt. Transfers Devt. Capital Devt. Recurrent Portion Recurrent Costs The largest part of the MOW budget is allocated to the urban water supply and, to a lesser extent, to the water resource management sub-sectors. A positive development is that a larger part of overall government funding is currently allocated to the rural sub-sector, which in view of the total population living in rural areas and their lower levels of access, results in a more equitable distribution of funds. Nevertheless, despite the decline in bias toward rural water supply (and sanitation), the average budget allocation to the sub-sectors is still tilted in favor of the urban subsector, which receives 7,767 TShs annually for every urban resident compared with 3,550 TSh per year for rural residents. 19

36 Social sectors: common issues 3.18 A deteriorating trend in the budget composition of the health and education sectors is the rise of allowances. The Ministry of Education and Vocational Training and the Ministry of Health and Social Welfare have entered the group of top five beneficiaries of allowances in 2008/09 budgets, coming from the sixth and seventh ranking the previous year. Allowances in these two ministries have increased both in share of total allowances and also as a percentage of ministry's personnel emoluments in 2008/09 compared with 2007/08, It is not very clear whether the activities in the two ministries involve paying a lot of allowances, as is the case for the National Assembly Fund and Ministry of Foreign Affairs. In the water sector, only 55 percent of total staff expenditure is made up of wages and salaries (up from 50 percent in 2006/07); almost all of the remaining staff costs consists of personal allowances. This translates into an overhead cost on basic salaries of 82 percent. This constitutes a decline in the past three years, as in FY2006/07 the overhead costs were higher, at 97 percent, but the current number is still very high and indicates that the pay reform process still has some way to go in the sector In the water sector, the resources for maintenance represent one third of the total budget and surpass the total resources allocated to the constructionhehabilitation of infrastructure. By contrast, maintenance spending represents only 1 percent of total spending in the education and health budget. This latter trend is particularly worrisome given the dilapidated state of much of Tanzania's social infrastructure. As the GOT is stepping up its efforts to build classrooms and health centers, it will be critically important to also increase the share of budget resources allocated to the maintenance of these new buildings, not to mention the stock of existing facilities (Figure 11). Figure 11: Maintenance and Infrastructure spending in selected sectors (in Bil. TSh.) Roads 3.20 The budget allocation in transport is generally in line with sector priorities, with trunk road sub sector as a priority in the transport sector. However, local roads have not been appropriately discussed in the sector strategy and, as a result, are receiving little budget allocation: the budget allocations for trunk and regional roads were, respectively, five and eight times higher in FY07/08 and FY08/09 than the one for local roads, and it is clear that funds for local roads are still very low compared with actual requirements In order to secure enough resources for the maintenance budget, the Fuel Levy, which is the main source of finance for the Road Funds, was increased. This significant increase in the Road Fund has led to a dramatic decrease in the gap between maintenance needs and actual funding in the road sector (Figure 12). 20

37 Figure 12: Road maintenance needs and budget (Bill. TSh) 2001 / / / / / / / /09 -t Needs -2- MTCE BUDGET Gap Source: Road Fund Report in the 2"d Joint sector Review, October 2008 ALLOCATION BETWEEN CENTRAL AND LOCAL GOVERNMENT 3.22 The Decentralization by Devolution (D by D) process is carried over through five ministries (health, education, water, infrastructure and agriculture) and Prime Minister's Office - Regional Administration and Local Government (PMO RALG) The share of resources devoted to LGA expenditures as a proportion of total GOT budget has slightly decreased. The share of LGA expenditures in the 2008/09 budget (excluding CFS) is 20.8 percent, compared with 21.5 percent in 2007/08. However, adding salary adjustment for agriculture, education, health and water staff in LGA into the total spending, the share of LGAs in the 2008/09 total budget reaches 21.4 percent, which is still slightly lower than the previous year's figure. This shows that there has not been any movement in transferring more budgetary resources to LGAs in the 2008/09 budget. Should this trend persist, in the next 2 years Government will not reach its objective of transferring 25 percent of budgetary resources to LGAs Wages represent 61 percent of the overall LGA budget. Their share slightly increased as compared with 2007/08. However, the composition of the wage bill has changed, as the share of allowances has increased by 2 percentage points of the total wage bill at the expense of the resources allocated for the payment of basic salaries. The share of total LGA recurrent spending has seen a 2 percent increase, from 83.2 percent to 85.percent, following a 3 percent rise in the maintenance budget Goods and services represent 17.5 percent of total expenditures, of which 10.3 percent are allocated to the purchase of education and medical supplies. The proportion of the LGA budget allocated for the purchase of goods and services remains stable as compared with 2007/ The share of capital expenditure represents barely 12.7 percent of total LGA budget as the bulk of the constructionhehabilitation work is carried out at the center level. Infrastructure spending has dropped since last fiscal year from 9.8 to 7.7 percent as the share of studies has surged; in 2008/09, it represents one third of capital expenditure. 21

38 RECOMMENDATIONS 3.27 The shares of MKUKUTA allocations in the 2007/08 and 2008/09 budgets are 70.6 percent and 70.8 percent, respectively. The slight increase in share of MKUKUTA in 2008/09 is due to an increase in ministries, departments and agencies (MDA) wages, since MKUKUTA shares without MDAs decline from 64.5 percent in 2007/08 to 62.0 percent in 2008/09. It is therefore important that the government have a clear commitment to protect MKUKUTA allocations The applicable definition of MKUKUTA allocations applies only to other charges and the development budget. The MDA wages and transfers to LGAs are not allocated according to MKUKUTA. However, MDA wages and transfers to LGAs are important elements for implementing MKUKUTA, especially public service delivery in health and education. It is important that MKUKUTA definitions be redefined to include MDA wages and transfers to LGAs in order to be strategically relevant Although the Planning and Budget Guidelines provide strategic direction in preparation of the budget and the MTEF, they don t seem to show a clear medium-term strategic direction in terms of allocation among three clusters. Allocations have moved more toward 40:40:20 shares despite government intentions of boosting economic growth. It is important, therefore, that the government express a clear medium-term strategy, which is economic growth through MKUKUTA cluster allocation, and strive to implement it Allocations to the transport sector, in particular to the trunk road sub sector, have continued to increase in recent years, with substantial funding from local resources. The Transport Sector Investment Plan has proven a useful anchor for medium-term budgeting in the transport sectors by providing a costing of the sector strategy and some form of prioritization. This experience could be extended to other sectors, such as agriculture Although the share of allowances in the overall wage bill has decreased over the past two years, for some sectors, the share has increased. For instance, the education and health ministries have entered the group of the top five allowance receiving ministries. Most services provided by these sectors are at the LGA level, but most duties facilitating allowances are paid to staff at the central level (ministry). Hence there is a need to control the proliferation of allowances in the social sector ministries (especially the health and education ministries) Slightly less than 1 percent of the allocations for social sectors are allocated to infrastructure maintenance. With the current government effort toward construction of social infrastructure, such as classrooms, teacher houses, and health centers, maintenance cost should go up in the future. This implies that the government should also increase the share of maintenance in health and education budgets in order to ensure these infrastructures good status Although the government is committed to decentralization by devolution policy, transfers to LGAs have stagnated at around 20 percent of the budget over the past two years. This poses a serious challenge to the government in terms of achieving its target, which implies further effort would be required in order to reach the 25 percent objective of LGA transfers. 22

39 Timeliness and completeness 4. IMPACTRESULTS BUDGET EXECUTION PERFORMANCE 4.1 A key factor affecting the performance of public services funded by the budget is the timeliness and completeness in the release of funds to the service delivery units. This section analyzes overall budget execution in term of timeliness and completeness of release of funds to spending units. More emphasis is also put on development budget execution which has performed poorly over the recent past. 4.2 Recurrent budget execution continues to perform well in 2007/08, both in timeliness and completeness. This indicates that the basis for delivery of routine services should be good, although two important qualifications apply: (i) some important recurrent resources for service delivery, such as drugs, are budgeted under the development budget; (ii) at the LGA level, execution actually measures releases of funds from the central government to LGA, and not actual spending. It is therefore important to have more information on actual releases at the service delivery level, for instance through the regular use of such instrument as Public Expenditure Tracking Surveys. 4.3 There is a worsening in the overall execution rate and timeliness of the release of development funds, as shown in Figure 13 with the execution rate for the development budget (both regions and MDAs) falling below 50 percent in 2007/08 from slightly over 50 percent in 2006/07. Figure 13: Development Budget Execution /07 and 2007/ / ICOngmal estimates +Exchequer releases Allocation.IC-Expenditure Source: MoFEA Expenditure flash reports and author s calculations.icongmal estimates +Exchequer Allocation -Expenditure releases 5 PETS have not been done over long period 23

40 ~ Overall.. 1.4% ~ I 4.4 An execution index calculated from the IFMIS rather than the flash reports data shows a higher execution rate of 56 percent for the development budget (an execution gap of 44 percent), but also indicates a significant worsening in the execution rate of the development budget for 2007/08 for MDAs (Table 13). Recurrent total ~ MDAS ~ I Table 13: Overall budget execution gaps I 2006/07 I 2007/08 execution gap Deviation index6 1 Overall execution gap Deviation index I i -0.1% 9.1% 4.5% I 1.O% 6.4% I Regions I.. I -3.o% 16.2% I Development total.. I 44.4% 46.0% MDAs 33.0% 36.2% j 49.6% 49.8% Regions 16.1% 25.2% Source: MoFEA, IFMS Itemized expenditure by warrant holder and author s calculation I Economic Nature and Funding Source 4.5 The previous analysis reveals that the execution of the development budget stills remains a major challenge. This section attempts to identify the causes of the problem. The development budget has two distinct features: (i) it is mostly funded by donors through project and baskets: and (ii) although it contains a good deal of current expenditure, the bulk of the development budget is made of capital expenditure such as infrastructure building and rehabilitation, and equipment acquisition. Execution of capital investment programs requires superior planning and implementation capacities as compared to current spending programs. This analysis separates the effects of the two features in order to understand more what is driving the low execution of development budget. 4.6 To separate planning and implementation capacities issues from funding problems in budget execution, development expenditure is reclassified by true economic nature. Specifically, current spending related to development programs (wages, goods and services and maintenance, etc.) were distinguished from actual capital spending (infrastructure construction and rehabilitation, project related studies, equipment acquisition, etc.). Execution rates were then computed and analyzed according to the two dimensions: (i) by economic nature (current vs. capital spending), and (ii) by funding sources (local vs. foreign) (table 12). 4.7 The analysis reveals three features of development budget. First, there is a sizable amount of current spending in the total development budget, accounting for 47.5 percent of the 2007/08 (Table 14). The weight of current spending in sectors development budget differs widely, with health having the highest share (83 percent) while roads having the lowest share (9 percent). 4.8 Second, the bulk of current spending in the development budget is financed by foreign resources (41.3 percent foreign vs. 6.2 percent local while in capital spending local funding is dominant. The picture varies widely across sectors. While in roads and energy sectors local funding of development budget is dominant, agriculture, water, education and health sectors foreign funding is dominant. The analysis reveals that foreign resources are primarily directed toward agriculture, water and the social sectors whereas local funding is relatively significant in roads and energy. Deviation index is the ration of sum of the weighted absolute difference between original budget estimates and actual expenditure to total original budget estimates. In this table it is calculated separately for recurrent and development budgets as well as MDAs and regional votes. 24

41 Table 14. MDAs Development Budget Execution /08 Sectors Overall Agriculture Educatlon Health Energy Roads Water :Itern :Current :Capltai!Total :Current :Capital :Total ;Current ;Capital :Total :Current :Capital!Total :Current :Capital :Total :Current :capital :Total :Current :Capital Shares of total Source: MoFEA, IFMS data and author s calculations Execution rates Forelgn Looal Total Forelgn Looal Total : ! ; : : ; : ; : : i : j : j ; : Third, the execution rates among foreign and locally financed current development spending are not significantly different (47.5 percent for foreign vs percent for local). Note however that these rates are still far lower than the execution rates of the recurrent budget (48.4 percent for current spending in the development budget vs. 98 percent in the recurrent budget). This suggests that the nature of expenditure matters than funding source. This is shown by the overall execution rate for true capital spending (47.4 percent ) which is not significantly different from that of current spending (48.4 percent) in the same development budget. The analysis reveals further that, execution rate is far lower for foreign funded capital spending (23.5 percent) compared to the locally funded (67.8 percent) within the development budget While the previous analysis links the problem of low execution to funding modalities, the sector analysis seems to highlight difficulties related to sector capacity to execute large investment projects. For example, although the bulk of the development budget in agriculture, education, and health sectors is financed by foreign resources, execution rates are relatively high. By contrast, sectors where foreign financing is less dominant but where capital spending accounts for the bulk of expenditure such as energy, roads and water execution rates are significantly lower. These trends underscore the importance of the issue of investment planning and implementation capacity The two sets of conclusions confirm that the problem of the under-execution of the development budget is a compounded outcome of the two features of the budget. First, is the problem of capacity to execute capital investment projects. Second, are the funding patterns and modalities. Weak capacity to execute capital spending may be attributed to inadequate capacity in key sectors to design and implement quality projects, from feasibility and economic evaluation studies to the physical implementation and monitoring. These aspects, in turn, reduce the capacity to absorb funds. The problem worsens when releases are untimely and resource levels are unpredictable - a phenomenon noted often in the case of donors funded interventions. Underlying Causes of the Spending Gaps 4.12 Under-reporting of spending for projects: Most of large infrastructure projects, especially in roads and energy, are funded by donors through direct to project funding modality (D-Funds). This 25

42 funding modality allows for the budgets of these projects to be recorded in the IFMS but release and spending data is not integrated in the system. It is estimated about 84 percent of such projects don t have their release and spending data integrated in the IFMS. This is major problem for the Ministry of Infrastructure Development (MoID), where, for instance in 2006/07 about US$ 100 million was wrongly-left unaccounted for according to EMS The problem of under-reporting is partly related to disbursement procedure where releases and spending for D-funds are supposed to be recorded in the IFMS through Dummy Exchequer. A recent study7 has concluded that this system is not functioning well and that very little data is captured by dummy exchequer. Although dummy exchequer system is good technically, the main actors (for example MOD and TANROADS) do not appear to have incentives to apply it adequately Delays in release development funds: The spending patterns of development budget are heavily linked to release pattern of funds. Funds that are released late in the year are likely not to be executed on time. Similarly large budget reallocations during the latter part of the financial year are likely to have a similar effect. For instance, whilst LGA budget execution performance was generally good in 2006/07, the last quarter of the same year experienced poor performance. The data indicates that substantial amounts of funds disbursed to LGAs in that quarter-particularly for recurrent expenditures-were not utilized The releases of development funds to spending agencies are on quarterly basis subject to submission of quarterly progress reports. For large contracting sectors certificate of completion of the project is also a requirement to trigger release of funds from the Treasury. These progress reports serve as tool for assessing the need for the release of funds within a cash management context. However, submission of these reports in the first quarter of the financial year tends to be a significant challenge for the spending agencies. This is partly because they are still engaged in finalization of the budgeting process in the first quarter. Non release of development funds in the first quarter due to lack of progress report has been a major cause lower execution of development budget in the first quarter of fiscal year. A potential way to overcome this problem is for the MoFEA to waive this requirement in the first quarter; especially to MDAs because they remit unspent funds back to treasury at the end of the financial year and begin the first quarter of the next financial year with clear balances Basket funding requirements: Basket funds (loans and grants) accounted for 18 percent of the total development budget in. 2006/07. They are a significant source of development funding for some key sectors (health, agriculture, education, water, local government, etc) and most of the core reform programs (PFMW, Local Government Reform Program - LGW 4.17 Legal Sector Reform Program - LSW, etc). Most basket fund mechanisms involve a two step disbursement process. Firstly, funds are deposited by DPs into a holding account once a specific set of requirements have been fulfilled. Secondly, the funds are withdrawn by government in to the exchequer spending account after the fulfillment of another set of requirements. The precise requirements are specific to each basket funding mechanism. Delays in the flow of funds can occur at either or both of these stages The sample of three basket funds (health, agriculture and Local Government Capital Development Grant - LGCDG) show inconsistent approach in the frequency of approval for disbursements; and the conditions for approval of disbursements to and from holding accounts, even 7 Integrating Aid into Planning and Budgeting Mechanism of the Government of Tanzania, University of Dar es Salaam, March

43 though all the funds operate within the same financial management context. It is true that all the basket agreements require effective implementation of monitoring and audit processes. However, some are specifically linked to disbursements while others are not. In the sample of baskets, an approved annual plan, budget and cash flow plans prior to disbursement are requirement. They also require periodic monitoring of implementation and performance before disbursing. However, the type and frequency of reporting and audit requirements vary significantly. For instance, reporting requirements vary so that at any quarter in a financial year, a planning officer in local government council may be required to prepare performance reports for: 0 0 Agriculture Sector Development Program (ASDP) -the most recent quarter (n-1) LGCDG -the quarter before the last complete one (n-2) HEALTH - at six month intervals (n-3) 4.19 Since the most of the performance reports require aggregated data from all of the MDAs, regions and local councils that receive basket funding at short intervals, the quality of these reports will necessarily be affected, diminishing their usefulness as a management tool. Additionally, requiring all of this reporting to take place on a quarterly basis become a big challenge and impairs disbursement of the funds. For instance, in health basket where reporting requirement is based on review of performance on longer term basis (six month), disbursements from donors were 75 percent in first quarter of On the contrary in ASDP and LGCDG baskets less than 25 percent of the budget was released in the same period Demanding administrative procedures: A major obstacle for the timely flow of funds that has been identified by this analysis is the administrative procedures associated with processing payments within spending agencies (Box 2). Some spending units have developed a large number of administrative steps required to process a regular payment that can take two weeks to a month if one would count one working day for each step in the process depicted in the Box 1. The large number of administrative steps for the verification of transactions such as the ones depicted below mark a departure from the standard (more streamlined) procedures set by the MoFEA, and may reflect a weak internal audit function within the vote. In addition to this, there is lack of clarity of the requirements of various steps as there is no operational or procedural manual that guides this process within the spending agencies. 27

44 Box 2 Steps involvement in securing payment within the Ministry of Health Subsequent to the procurement process, once a contract is in place, the follow steps need to be taken in order to secure the Local Purchase Order (LPO) and to trigger payment: 1. Following a request from a line department, PMU prepares request for LPO 2. Director / Department head signs off on request for LPO 3. Permanent Secretary signs off on request for LPO (if amount exceeds TSh 1 million) 4. Accounts Department enters the purchase into the (manual) vote book 5. Request for LPO is sent to the Chief Internal Auditor (CIA) for examination 6. Request is sent to the machine room and entered into the IFMS 7. The LPO is generated 8. LPO is sent to PMU for signatures 9. LPO is sent to CIA (examination) for signature 10. LPO is sent to Chief Accountant for signature 11. PMU delivers the LPO to the supplier 12. Upon delivery of goods or services, a delivery notice and invoice are submitted to PMU 13. The invoice with supporting documentation is passed on to Chief Accountant 14. The payment request is entered into the IFMS by the machine room 15. The check is printed by the Treasury (Accountant General) and provided to the supplies by PMU Underlying Causes of the Release Gaps Weak cash flow management system: The cash management system is still nascent and has not been fully implemented. Monthly cash flow projections are not been produced by all the spending agencies, which makes it impossible to compile the full cash flow picture and to allocate resources accordingly. Even for those spending agencies which have produced cash flow plans they are credible in the sense that they tend to frontload their resource requirements, making them difficult to implement under a cash budgeting framework. Moreover, the draft report on the implementation of the cash management system indicates that the lack of an established administrative structure and expertise to handle the roll out of the system are the underlying causes for the lack of a coherent cash management system. These may largely be teething challenges to this new process Nevertheless, strengthening the cash management unit needs to be to be a priority for the government. This includes spending units producing realistic cash flow plans at the beginning of the financial year and tracking what happens to them during the year. In addition, the integration of the cash flow forecasts with the relevant module of the EMS should be pursued in order to enhance the coherence of the systems. Currently, the cash management or procurement plans are note captured in the IFMS, where a pro rata allocation (divide by twelve) of the budget acts as the monthly budget plan. Since monthly releases do not tend to follow a cash flow plan, a time consuming resource allocation process takes place each month, causing delays in the deployment of resources Unrealistic budgeting: Reliability and relevance of the MTEF projections and budget estimates are important elements in execution of the budget. Since foreign resources finances more than 65 percent of development budget its predictability, both in the budget and MTEF, is key. Projects and basket funds are the most unpredictable resources, leaving budget support as the most predictable resources. The problem of unpredictability results into either over-budgeting or under- The Implementation of Cash Management System, Ministry of Finance, Tanzania, May

45 budgeting. For instance, the local government development budget execution performance in was affected by over-budgeting for the Primary Education Development Program (PEDP). While the problem of medium-term aid predictability is widely recognized, some options to deal with this problem have been put forward, including the scenario analysis. It is important that the government develop some scenarios in response to more or less than expected aid. These scenarios should not be limited to priority investment programs but also should include options on how to adjust recurrent expenditures. EFFECTIVENESS OF PUBLIC SERVICE DELIVERY 4.24 Critical to the evaluation of the budget is analysis of its impact on service delivery. Although this kind of effectiveness analysis requires more micro-level analysis, such as quality surveys, expenditure tracking surveys, and other diagnostic tools, it is possible at a more aggregate level to derive useful conclusions from matching performance indicators with the evolution of spending. Education 4.25 In the education sector, the Ministry has provided a first-rate performance report for FY2007/08 that is comprehensive, up-to-date, and candid in this identification of issues. Such an excellent report provides a good basis for linking budgets and results In primary education, the ministry s data show that the net enrollment ratio stalled in 2007/08 after many year of increase (Figure 14). More worrisome, the percentage of cohort completing Standard VI1 has dropped to 65 percent in 2007/08 from 72 percent 2006/07 and 77 percent in 2005/06, indicating a trend that is inconsistent with the universal primary education Millennium Development Goals (MDG). There are minor gender differentials. There has also been a consistent worsening of the pupil-teacher ratio in primary, which has now reached 54: 1, although the primary qualified teacher-pupil ratio has shown a significant improvement indicating that the proportion of qualified teachers is increasing At the secondary level, overall enrollment has continued to increase, to around 25 percent for lower secondary and 1.4 percent for upper secondary. There has been a steady increase in completion rates of lower secondary, up to 94 percent; however, the completion rates of upper secondary education collapsed from 93 percent to 73 percent in 2007/ The HBS also provides data on primary and lower secondary net enrollment. Those data, at 84 percent and 15 percent, respectively, are lower than the ministry s data, reflecting possibly an inaccurate registration of pupils age at the schools and some over-reporting of enrollment as well as normal statistical discrepancies among different sources The data indicate that the objectives of universal primary enrollment and universal primary completion are still problematic and that primary education still needs sustained funding if it is to reach these objectives. The data also raise the question of why, with continuous high levels of funding, some key indicators of results, such as net enrollment and completion rates, tuned in the wrong direction in 2007/08. Finally, the collapse of the completion rate in upper secondary education, if confirmed, might be a symptom of low quality and unsustainable expansion of secondary education in recent years. 29

46 Figure 14: Net enrollment ratios - primary, lower secondary, , HBS Source: NBS. Dar es Salaam Other urban areas Rural areas Mainland Tanzania 1 C@3 Primary NER I Lower secondary NER 1 Health 4.30 The performance report of the health sector provides some indication of the recent evolution of the health situation in Tanzania, although very few indicators refer to the 2007/08 period. While the top five causes of morbidity remained stable in relative terms from 2000 to 2006, the top five causes of infant mortality have been retreating, in particular malaria and anemia. The only data relative to 2007/08 show a reduction in the immunization coverage below the MKUKUTA target of 85 percent, which had been reached the previous year Complementary information on the recent health status of the population also comes from the 2007 HBS. It shows similar proportions reporting illness, and similar patterns by area (rurabother urban>dar) between 2001 and It also shows no changes in proportions consulting any provider when sick (69 percent), but an increase in consultation at government facilities (55 percent to 65 percent). It reports a small increase in proportion satisfied with services in government facilities and fewer reporting lack of drugs The snapshot of the effectiveness of the health sector is mixed. With a sustained increase in the health sector budget in the last 3 years, from 10 percent in 2005/06 to 11 percent in 2008/09, the data seem to imply that, although infant mortality is declining, the overall morbidity burden of the population has not improved, casting doubts on the effectiveness of primary health prevention. The curative care system seems to be performing better, with fewer people report lack of drugs in government health facilities. Water 4.33 The water sector performance report contains a comprehensive set of up-to-date indicators to track spending effectiveness. The documents show an increase of around 10 percent in the number of rural water points built by December 2007, and a larger increase in the number of urban household connections. The rural water supply coverage is estimated at 57 percent in 2007, and the overall coverage of water services in all urban areas including Dar es Salaam stood at 77 percent in June 2008 (Figure 15). The report concluded that the MKUKUTA objective for rural water coverage was off-track, while the urban coverage objective is possibly on-track, though the reduction of the water sector budget this year may make this more problematic. The estimated access to a protected source of water in 2007 is 41 percent in rural areas, 83 percent in Dar es Salaam, and 77 percent in other urban areas. The lower estimates for rural areas are likely to be more realistic than MOW data that assume a ratio of 250 people served per water point. 30

47 I Figure 15: Access to safe drinking water supply per percent of population Source: NBS. Roads 4.34 The monitoring framework of the road sub-sector from the joint infrastructure sector review provides a comprehensive and up-to-date set of output indicators. These show good progress in the share of roads in good conditions for almost all categories of roads. The document also reports an increase in the length of paved roads by around 500 km in 07/08. These improvements show that the increased allocation to the road sector both for maintenance and construction is being put to good use, although the issue of cost efficiency, given the recent sharp increase in construction material prices, needs to be reviewed. Agriculture 4.35 The agricultural sector review report provides some physical indicators of the sector, but no output or outcome indicators. The only exceptions are data on crop yields, which are however not up to date for 2007/08 and cover only 3 years, too short to make any inference between agricultural yields and fertilizer use (which is subsidized by the budget). EFFICIENCY IN PUBLIC SERVICE DELIVERY 4.36 Data on efficiency are not commonly produced in many sectors, and more broadly, the budget process in Tanzania seems largely unconcerned about this issue, in terms of both budgetary allocation and management. However, given the huge financial needs and the limited resources, increasing efficiency is an important tool for achieving a better budget policy. Roads 4.37 Figure 16 indicates that in the road sector, because of inflation of the price of construction materials, the unit costs of road construction and maintenance have gone up substantially, by almost 40 percent in four years for routine maintenance and rehabilitation, implying that larger budgets do not necessarily translate in potentially greater outputs. 31

48 Figure 16: Unit Cost per km in Routine and Recurrent Maintenance FY FY FY2005l06 FY FY Trunk Road Regional Road Education 4.38 In education, one key indicator of internal efficiency of the system is the repetition rate, which impacts the cost per student to complete a course of study. The primary repetition rate had been going down since 2003 but has increased significantly from 4.4 percent to 4.9 percent between 2006/07 and 2007/08, pointing to a potentially costly decrease in efficiency of the primary education system. Equity in Public Service Delivery 4.39 Inequity in budget allocations persists across LGAs, according to 2008/09 budget figures. There is little evidence of progress toward addressing the inequity in key sectors such as of health and education The persistent inequity is driven by uneven personnel emoluments budgets stemming from the difficulty of attracting staff to particular LGAs. Government is trying to solve this issue and routinely posts staff to underserved LGAs. However, very few of the staff members recruited actually take up their postings, and consequently, the PE budgets continue to reflect staff in posts and not the desired needs-based level of staffing LGA education personnel emolument block grant budgets for 2008/09 per child aged 7 to 13 are shown below relative to the budget for 2007/08 and the formula based allocation for 2008/09 (Figure 17). The graph shows an increase in the budget with an almost identical pattern of inequity. Annual personnel emoluments expenditures per 7-13 year old vary from TSh. 33,000 to TSh. 175,000 across LGAs. If it were possible to apply the formula, then personnel emoluments expenditure in 2008/09 would have averaged TSh. 66,000 across LGAs. 32

49 Figure 17: Education personnel emolument expenditure per child aged 7-13 across LGAs 200,000 E 180, ,000 " 140, ,000 P I 100,000 80,000 60,000 c 40,000 20, I The situation is the same for health sector personnel emoluments block grants budgets, except that there has been only a marginal budget increase between 2007/08 and (Figure 18). The health sector per capita PE budget varies from TSh 1,400 to TSh 14,000 across LGAs. It should also be noted that the variation in the health formula line is explained by the fact that the health block grant formula is based on only 70 percent of the LGA population. The remaining 30 percent is driven by the number of poor residents (1 0 percent), district medical vehicle route (1 0 percent) and underfive mortality (1 0 percent). Figure 18: Health personnel emolument expenditure per capita across LGAs Budget b Budget Highest spend - Lowest 4.43 Given the uneven expenditures, there are certain LGAs benefiting and certain LGAs suffering (Figure 19). Some information is provided below. It should be noted that a poor rank for education (health) correlates with a poor rank in health (education),which suggests that the problem is with staffing in general and not with a particular sector. 33

50 Figure 19: LGA ranked in terms of per capita expenditure budgeted for 2008/09 Top 10 served LGAs Per capita aged 7-13 education budget 2008/09 Njombe Urban (Iringa) Ileje (Mbeya) Kibaha Rural (PwanUCoast) Songea Rural (Ruvuma) Kyela (Mbeya) Songea Urban (Ruvuma) Kibaha Urban (PwanVCoast) Mwanga (Kilimanjaro) Morogoro Urban (Morogoro) Pangani (Tanga) Top 10 served LGAs Per capita health budget 2008/09 Pangani (Tanga) Mafia (PwanVCoast) Kisarawe (Pwani/Coast) Mwanga (Kilimanjaro) Kibaha Urban (PwaniKoast) Ludewa (Iringa) Kibaha Rural (PwaniKoast) Iringa Urban (Iringa) Ukerewe (Mwanza) Bukoba Urban (Kagera) Bottom 10 underserved LGAs Per capita aged 7-13 education budget 2008/09 Nkansi (Rukwa) Bahi (Dodoma) Monduli (Dodoma) Kondoa (Arusha) Bariadi (Shinyanga) Kigoma Rural (Kigoma) Ngorongoro (Arusha) Tabora Rural (Tabora) Bukombe (Shinyanga) Chato (Kagera) Bottom 10 underserved LGAs Per capita health budget 2008/09 Bukombe (Shinyanga) Kasulu (Kigoma) Rorya (Mara) Mkinga (Tanga) Kilindi (Tanga) Tabora Rural(Tabora) Bariadi (Shinyanga) Kigoma Rural (Kigoma) Sikonge (Tabora) Nanyumbu (Mtwara) 4.44 The distribution of the recurrent road budget, excluding maintenance, among LGAs is not geographically equal, as it does not correlate with the length of the roads in the district (Figure 20). Figure 20: District recurrent budget and road length ' r e s Recommendations 4.45 Improve key elements of the budget execution cycle, which includes streamlining basket funding modalities; streamlining of internal MDA procedures; improving cash management system; and improving capacity to implement capital investment projects. 34

51 i. Streamlining of basket funding modalities: Despite basket funds operating under the same financial management system, there are a lot of inconsistencies in frequency of approval and the conditions for approval of disbursement of funds from the holding accounts. While all basket arrangements require effective implementation, monitoring and auditing processes, not all these processes are linked to disbursement. Developing a general standard procedural requirement for a well functioning basket fund arrangement is a key to effective execution of a development budget. ii. Streamlining of internal MDA procedures: While administrative procedures for triggering payment to suppliers are necessary, such measures need to be streamlined. It is understood there are some standard administrative procedures set by MoFEA for effecting payment to suppliers, but some MDAs have added more procedures, which slows down the payment process and execution of the budget. Ensuring MDAs adhere to standard streamlined procedures for effecting payments to suppliers is important for effective execution of the budget. iii. Improvement to the cash management system: It is important that spending agencies produce realistic cash flow plans at the beginning of the fiscal year and that MoFEA completes tracking of what happens to these cash flow within the year. This is a key step in order to improve predictability of funds to spending units and also to improve the cash management process. iv. Improvement in capacity to implement capital investment projects: It is important to improve the capacity of the key sectors to design and implement quality projects, from feasibility and economic evaluation to procurement and physical implementation as well as monitoring, especially for sectors where development projects have a large component of capital spending, as in roads and energy Improve quality of expenditure flash reports (including reliability of the data) given that they are key instrument for tracking on a monthly basis the within-year budget execution. This report is prepared by the ACGEN's office but compiled manually from reports generated automatically from the IFMS. Automation of this report is recommended so that is generated from IFMS directly Make more regular use of Public Expenditure Tracking Surveys to obtain better information about effective budget execution at the service delivery level. The sector where PETS are suggested to be carried out regularly include education, health, water, and agriculture. The PEFAR 09 work program has already included water and education as sectors where PETS will be carried out Continue with plan to develop significant results indicators for the budget, and restructure budget programs to be aligned to such indicators. It is important that MDA and LGA budgets and MTEFs are organized by programmatic approach, with programs drawn from the SP, and that result indicators be linked to programs. This will make it easy to link amount of resources (input) and outcomeh-esults achieved (output) Make the discussion of results indicators a central element of the sector budget scrutiny. This is a key step to ensure that sectors link resources to results or outcomes, since sectors will have to justify use of the previous year's resource in terms of results or outcomes achieved before making requests for following year. 35

52 4.50 Introduce key efficiency indicators along with results and quality indicators in budget management. It is important to introduce efficiency indicators in order to ensure that scarce budgetary resources are spent where more results can be achieved To try to solve the inequity issue, further efforts must be made to attract and retain staff in underserved LGAs. Three suggestions have been made by LGA-based government staff during recent field visits: i. ii. iii. The proposed and currently being considered national incentive scheme is fasttracked and implemented. An equalization grant is allocated to under-served LGAs that are earmarked to efforts focused on the attraction and retention of staff. All LGAs receive their PE budget based on formulddesired postings. Where staff members do not take up their postings, the full funds are still released and then can be used as other charges to assist in the attraction and retention of staff (for example: equipment for starting a new home, better working conditions, training). 36

53 PART 11: PFM ASSESSMENT 37

54 5. INTRODUCTION OBJECTIVE OF THE PUBLIC FINANCE MANAGEMENT PERFORMANCE REPORT 5.1 The objective of this Public Finance Management Performance Report (PFM-PR) is to update previous PEFA indicators and their associated reports to November Process of preparing the PFM-PR 5.2 In 2004 a PEFA assessment exercise was undertaken by the development partners (DP), using PEFA indicators and scoring methodology before the PEFA Secretariat issued its standard Manual for PEFA-PFM Performance Measurement Framework in June Later, using the standard PEFA Manual, the PEFA indicators were revisited and scored by different DPs in preparing their individual PFM-related instruments and reports. Also, the DPs have updated PEFA indicators twice, but these updates were limited to the scoring of indicators only, and no comprehensive PFM- PR reports were prepared. In May 2007 the Public Finance Management Working Group (PFMWG) of the DPs, with the assistance of a consultant, conducted a full PEFA exercise. Its report, including a PFM-PR was drafted in July Comments were received from the Peer Reviewers, the PEFA Secretariat and other stakeholders. 5.3 With the assistance of a short-term consultant recruited jointly by the World Bank and EC Delegation, and in cooperation with government officials, the 2007 field work data and accordingly its July 2008 PFM-PR was updated in November 2008, and the stakeholders' comments have been addressed. In the absence of a full PFM-PR in the indicator updates prepared in the past, a comparison and explanation of differences between the scores of the PEFA indicators in this report with those recorded in past reports by using summary tables only, may not prove to be firm and/or de~endable.~ The methodology 5.4 In line with the objective of this PFM-PRY the PFMWG with the assistance of a short-term consultant updated and finalized this report, preparatory work of which had been conducted earlier as mentioned above. Comments of the institutions on the earlier draft were instrumental in finalizing this report. As in the July 2008 version of the report, the government officials have facilitated the updating and finalization of the report, and whilst not having an input in the scoring of the indicators, they have been central in supplying the information and updates. In this regard, whenever possible, evidence in the form of reports was obtained and quantitative data sought. Also the team discussed in detail the qualitative descriptions of events and processes with the officials. The scope of the assessment 5.5 PEFA exercises in Tanzania were initially planned to be separately conducted for the central government, local governments, and public enterprises. Whereas an exercise was conducted in 2006 for local governments in the context of a donor fiduciary assessment and a report was prepared, application of the PEFA Manual to the public enterprises was not fully possible. This report covers It should also be noted that comparing the status of PEFA indicators in this report with those of the 2004 PEFA exercise is not helpfil due to different methodology and definitions that were used in 2004 (PEFA pilot period) before the final manual was issued by the PEFA Secretariat in June

55 the central government operations, and the issues of local governments and public enterprises are only visited in relevant indicators as recommended by the PEFA manual where they have fiscal relations with the central government, and in the context of fiscal risk assessment and transparency and timeliness of fiscal transactions. 39

56 Country context 6. COUNTRY BACKGROUND INFORMATION ECONOMIC SITUATION 6.1 Tanzania is a low income country with a population of 41.5 million people in The country s economic performance has been stable. Grounded in prudent macroeconomic policies, growth averaged 7 percent during , outpacing the average for sub-saharan Africa. Inflation remained moderate during this period, although recent global fuel and food price increases pushed inflation to 9 percent in the first quarter of Government spending has experienced extraordinary growth since 2001, financed by a significant broadening of the revenue base and scaled-up donor assistance. By limiting the government s use of. domestic financing, fiscal policy helped to ease inflationary pressures and provided room for a rapid expansion of credit to the private sector. Extensive debt relief and a major build-up of international reserves have reduced external vulnerabilities. Overall government reform program 6.2 Tanzania s National Strategy for Growth and Reduction of Poverty (NSGRP), known as the MKUKUTA (Mkakati wa Kukuza Uchumi na Kupunguza Umaskini Tanzania) was designed in June 2005 for implementation over the period It is the successor to Tanzania s Poverty Reduction Strategy Paper (a first generation Poverty Reduction Strategy Paper (PRSP), formulated in 2000) and builds on Tanzania s Development Vision 2025, especially in its emphasis on growth and long-term strategy for reducing aid dependence. MKUKUTA has an increased focus on equitable growth and governance, and is an instrument for mobilizing efforts and resources toward targeted poverty reduction outcomes. MKUKUTA includes targets for poverty reduction outcomes that are consistent with, and indeed in many cases go beyond, the MDGs. MKUKUTA identifies three clusters of broad outcomes: (i) growth of the economy and reduction of income poverty; (ii) improvement of quality of life and social well being, and (iii) enhanced governance and accountability. Rationale for PFM reforms 6.3 Continuous PFM reforms are recognized as key to achieving the aims of the MKUKUTA. The PFM reforms cover all stages of the system from planning and budgeting to budget implementation, control, auditing, and external oversight. The Government has announced that it seeks to achieve allocative efficiency by (a) ensuring aggregate fiscal discipline and accountability, (b) allocating resources in accordance with government priorities, and (c) promoting efficient service delivery through enhanced predictability and availability of medium-term resources for the MDAs. Io Based on the country data and statistics information available on the World Bank website for 2006 and calculating an annual population growth rate of 2.5 suggested in the same source. I IMF Website-Tanzania: Third Review Under the Policy Support Instrument-Staff Report, June

57 Fiscal performance BUDGETARY OUTCOMES 6.4 Between 2005/06 and 2007/08, Tanzania s overall fiscal performance has improved. Domestic revenue grew by 3.4 percentage points to 16.0 percent of GDP. External grants (direct budget support and program support, including basket grants and project grants) grew by 1.6 percentage points, reaching 7.0 percent of GDP. Expenditures remained stable in this period, and fiscal deficit and its financing as a percentage of GDP was reduced from 5.0 percent to 1.6 percent of GDP. Thanks to external concessional loans for financing the deficit, net domestic borrowing in 2007/08 became negative, thereby reducing stock of domestic debt (Table 15). According to the 2008/09 budget speech, a net zero domestic financing policy will continue in 2008/09. On the other hand, the government faces a fiscal risk with about 30 percent of its total revenue coming from external grants. Moreover, given the fact that from a macro-fiscal perspective there is no room for domestic non-inflationary borrowing, almost the entire budget deficit is financed by concessional external loans. Total revenue - Own revenue - Grants Total expenditure - Non-interest expenditure - Interest expenditure Primary deficit Aggregate deficit (incl. grants) Adjustment to cash13 Overall deficit Net financing - external - domestic Actual Actual Actual O O Allocation of resources 6.5 In the absence of a functional classification of expenditures in Tanzania s budget and accounting systems, it is difficult to provide a clear picture of the allocation of government revenue to internationally-recognized government operations (indicator P-5 for details). A data bridging exercise from the existing administrative and economic classifications to a standard functional classification does not seem to be very helpful either, due to several assumptions that may not be stable and reliable over time. Indirectly, however, the economic classification of expenditures (Table 16) demonstrates a substantial increase in development expenditures, and knowing that these expenditures are mainly allocated to social and economic services, it may be concluded that the composition of expenditures over the last three years has moved these kinds of expenditures from 5.7 percent to 8.0 of GDP. This l2 2007/08 data is preliminary in accordance with the data source. l3 Unidentified financing (+)/expenditure (-). Includes expenditure carryover from the previous year 41

58 also may be verified by increased external grants and concessional loans, which are typically directed to social and economic services, especially those which are based on programs and projects. However, since recurrent expenditures for operations and maintenance are not classified on a functional basis, this alone does not provide a full picture. 2005/06 Actual Total expenditure 23.0 Recurrent expenditure / /08 Actual ActualI Interest payments I -Goods and services and transfers I 12.1 I 11.1 I 8.8 I Development expenditures - Domestically-financed -Foreign- financed The legal framework THE LEGAL AND INSTITUTIONAL FRAMEWORK FOR PFM 6.6 The roles and responsibilities, accountability of spending agencies, transparency requirements, and sanctions arrangements are specified to different extents in various pieces of legislation: the Constitution; the Public Finance Act; the Public Procurement Act; the Local Government Finance Act; the Loans, Grants and Guarantees Act, and a new Audit Act. 6.7 Chapter 7 of the Constitution of the United Republic of Tanzania (1977) outlines the legislative function and the roles of various bodies involved in the management of public finances, specifically the National Assembly (legislature), the President (executive) and the Controller and Auditor General. 6.8 The Public Finance Act (2001, revised in 2004) and its subsidiary instrument (regulations 2001, revised in 2004) defines in great detail the roles, functions and responsibilities in management of government revenue and expenditure (the Minister of Finance, the Paymaster General, the Accountant General, the Accounting Officers and Warrant Holders in ministries, departments and agencies, as well as the Controller and Auditor General). They also define the accounting, control and reporting systems. 6.9 A new Audit Act passed the National Assembly and was made public in a September 2008 gazette. It is expected to bring about wide-ranging improvements in the external auditing task in the future, including further independence of the NAO, enhanced engagement of the PAC of the National Assembly in the oversight processes, and response of the executive government to the NAO s findings. Presently, the NAO is drafting the enabling regulations of the new law The Public Procurement Act (2004) repeals the Public Procurement Act of with a view to make better provisions for the conduct of public procurement with the establishment of the public procurement regulatory authority, tender boards, principles and methods of procurement and dispute l4 2007/08 data is preliminary actual in accordance with the data source. 42

59 settlement. The enabling regulations of the act were updated in 2005 with a focus on the selection and employment of consultants and outlines of specific guidelines for their selection, recruitment, and payment The Local Government Finance Act of 1982 (as amended in 2000) and the Local Authority Financial Memorandum of 1997 require each council to advertise in the media andor post information on the council notice boards key information, including: receipts of funds from the government, expenditure, statements, budgets and signed audited accounts, and tender advertised. They also require allowing the public to attend the full council meetings The Loans, Grants and Guarantees Act (1984), amended in 2003, defines roles, functions and responsibilities in public sector contracting of loans, issue of guarantees, and receipt of grants. This covers the entire United Republic of Tanzania (URT), though some sections are silent on the position in Zanzibar Since 2002 a number of taxation acts were updated, including the Income Tax Act; Value Added Tax Act; Tax Revenue Appeals Act; Gaming Act; Vocational Educational and Training Act; Foreign Vehicles Transit Charges Act; Hotels Act; Motor Vehicles (Tax on Registration and Transfer) Act; Stamp Duty Act; Road and Fuel Tolls Act; Port Services Charges Act; Airport Service Charges Act; and Tanzania Revenue Authority Act. The institutional framework 6.14 As noted in the IMF Report on the Observance of Standards and Codes (2002), fiscal management responsibilities are defined on the basis of a clear separation of roles between the executive, legislative and judicial branches. The Constitution assigns the responsibility for fiscal matters to the executive and legislature; and it also provides the legal basis for appropriating and spending public funds. The National Assembly approves the state budget enables laws for the imposition of taxes, and authorizes expenditure out of the Consolidated Fund. The Cabinet of Ministers, on the basis of authority conferred by the President, is responsible for formulating the budget and submitting it to the National Assembly for approval. The judiciary process and procedures, including the composition of the courts, are defined in the Constitution The MoFEA oversees budget preparation and execution. Each year in June, it presents to the National Assembly the Budget Speech, which contains the government s fiscal revenue, expenditure, and financing policies and plans. The Ministry monitors fiscal developments during the year and reports to the National Assembly. The Ministry also formulates and manages revenue policies and legislation that are presented to the legislature. Its responsibilities include preparing the central government budget; developing tax policy and legislation; managing government borrowing on financial markets; determining expenditure allocations to different government institutions; and transferring central grants to local governments. The key features of the PFM system 6.16 Tanzania has a few PFM features that need to be mentioned, as they have some impact on the analysis of the indicators First, apart from the central government MDAs as the first level organizational classification of the government budget (known as vote), there exist some AGAs, which are regarded as central government agencies under some MDAs, and which enjoy more financial freedom after receiving their bulk transfers from their parent MDA. The AGAs might pose some fiscal risks by their decisions 43

60 in the form of future larger transfers from the government budget. Their spending is not subject to the same rules and scrutiny that apply to MDAs. For one thing, some AGAs do not use government standard budget classifications Second, the LGAs receive about 95 percent of their resources from the central government under different arrangements, almost all of which can be classified as conditional grants. In other words, the central government has delegated several of its functions to the LGAs while policy and financial aspects of these functions remain at the central level. The LGAs, therefore, in the PFM context, should be regarded as the extended arms of the central government s agencies, and are treated as MDAs with regard to their budgeting, payment and accounting systems. Given this and considering the fact that LGAs do not have borrowing power (except by specific law); they do not pose fiscal risks to the central government, as they financially are treated as MDAs Third, the PES generally have commercial status in their budgeting and accounting systems, and therefore can pose potential and actual fiscal risks. This may arise from a need for transfer from the central government s budget to them. both on current and capital accounts in the form of subsidy or capital injections. In Tanzania AGAs and PES collectively are referred to as Parastatals, though they are different organizations by international standards Fourth, the appropriation structure that is approved by the National Assembly is brief and broadly classified (normally one vote for one ministry). It provides wide authority to the government to change its operations without reference to the legislature. Although some detailed ministerial budgets are prepared and widely disclosed, these can be changed in the course of budget execution by the executive branch, mainly by the approval by the MoFEA. 44

61 7. ASSESSMENT OF THE PFM SYSTEMS, PROCESSES AND INSTITUTIONS BUDGET CREDIBILITY Aggregate expenditure outcomes compared with original approved budget (PI-1) 7.1 The ratio of actual to budgeted total expenditures (excluding interest payments and externally-financed development expenditures, both of which are subject to substantial variations in most countries) in 2005/06 and 2006/07 have been within normal range (Table 17). In 2007/08 this ratio has exceeded 10 percent. The budget execution reports of the MoFEA reveal that most variations have been from spending on goods and services in the recurrent budget and domesticallyfinanced development expenditures. A draft World Bank policy note identifies the main reasons for budget implementation deviations of both recurrent and development budget^.'^ Table 17: Comparison of Originally Budgeted and Actual Expenditures in billions of TSh / / l08 1. Originally total budgeted expenditures (excluding interest I 2,845.1 I 3,563.7 I 4,260.2 payments and foreign-financed development expenditures). 2. Actual total expenditure (excluding interest payments and foreign-financed development expenditures). 3. Absolute difference 4. Percentage deviation 2, , , % -3.9% -13.1% Indicator PI-1 Aggregate expenditure out-turn compared to original approved budget Score B Brief Explanation and cardinal data used In last three years, only in 2007/08 did the actual total expenditure deviate from budgeted expenditure by more than 10 percent. Composition of expenditure outcomes compared with original approved budget (PI-2) 7.2 There are some big differences between original budgets and outcomes in almost all MDAs. This is partly due to the distribution of contingencies, more specifically to salary increases during the year from a provision under the MoFEA vote to the MDAs. Most of these differences are due to discrepancies on the development expenditures and non-salary expenditures. In general, development expenditures are over-budgeted or, in the case of those financed from domestic resources, are cut back during budget implementation. Cash restrictions force the government to reduce expenditures on goods and services in the course of the budget year in almost all MDAs. On the basis of the figures provided by the AGD on the actual expenditures of the main votes, and comparing them with the original budgets, these variations have been calculated according to the PEFA Manual (Table 18). l5 Draft Budget Execution Analysis for 2006/07, Policy Note, World Bank, September

62 Variance in overall expenditure (as defined in PI -1 above) 3.7% 3.9% Variance in expenditure composition (in percent) % 12.9% Excess of variance in expenditure composition to overall primary expenditure (percentage points) 28.8% 9.0% % 20.8% 7.7% 7.3 Table 19 indicates that in 2005/06, the variance in expenditure composition exceeded the variance in overall expenditure by 28.8 percent, but this has returned to below 10 percent in the last two years. The unusual variation rate in 2005/06 is related to the votes associated with the ministries of Infrastructure, Health, Education, Water, and Energy and Mining. This deviation is explained in part as the result of creatindmerging a few MDAs after the elections in the course of the fiscal year, the accounts o f which are not fully separated or adjusted in the actual figures for that year. Indicator Scores Brief Explanation and cardinal data used PI-2 Composition of expenditure outcome compared with original approved budget C In 2005/06, variance in expenditure composition exceeded overall deviation in primary expenditure by more 10 percentage points. Aggregate revenue outcome compared with original approved budget (PI-3) 7.4 Actual domestic revenue collection compared with its estimates in the originally approved budget has consistently over-performed in the last three years, with surpluses ranging from 3 percent to 13 percent of the approved budget (Table 19). This performance can be noted in almost all domestic revenue components of tax and non-tax categories in the budget execution reports of the MoFEA. Original budget (total domestic revenue) Actual revenue collection % of actual collection to original budget Table 19: Domestic Revenue Performance in billions of TSh / , , , , , , % 113% 104% 7.5 It appears that traditionally, governments in Tanzania are rather conservative in revenue projections, thereby providing a safeguard to unexpected in-year expenditures, and preventing possible unwanted budget deficits. It should also be noted that tax policy and administration reform in recent years, along with better economic performance, have helped enhance revenue performance. Figures refer to the sum of absolute deviations for the largest 20 votes as a proportion of total budgeted expenditure, excluding debt service and externally-financed operations. 46

63 Indicator PI-3 Aggregate revenue outcome compared with original approved budget Scores A Brief Explanation and cardinal data used In the last three years, revenue collection was consistently above the approved budget. Stock and Monitoring of Expenditure Payment Arrears (PI-4) 7.6 Stock of expenditure payment arrears and any recent changes in the stock. Payment arrears in the last two years have increased. It is widely understood that without monthly cash allocations, the Integrated Financial Management System (IFMS) does not allow for expenditure commitment. However, payments still could be delayed for any reason, including those on expenditures without repeated contracts, such as utilities or non-completion of payment documents at the end of accounting period or supplementary legal claims associated with previous contracts due to price escalations. In 2005/06, the stock expenditure arrears as a percentage of total expenditures was minimal, but in both 2006/07 and 2007/08, these have substantially increased. 1.Total expenditure^'^ 2.Stock of arrears at year-end 3.Percentage of 2 to , , , % 3.9% 7.8% 7.7 Availability of data for monitoring the stock of expenditure payment arrears. Data on payment arrears are collected at the year-end, when final accounts and government financial statements are prepared for external auditing. The AGD receives financial reports from the MDAs in which they are required to report any arrears with a footnote explaining the reason for the accumulation. In a related matter, periodic audits for specific MDAs also are undertaken to verify their arrears prior to clearance during the year, the action which by itself indicates the accumulation of payment arrears. Indicator PI-4 Stock and Monitoring of Expenditure Payment Arrears I Score c+ I Brief Explanation and cardinal data used (i) The stock of expenditure payment arrears constitutes 2-10 % of total expenditure, and there is no evidence that they have been reduced significantly in the last two years (C). (ii) Data on the stock of arrears is consolidated annually, but they do not have an age profile (B). l7 As defined and recorded in table

64 Classification of the budget (PI-5) COMPREHENSIVENESS AND TRANSPARENCY 7.8 Since 2001, the government has introduced a Government Finance Statistics (GFS)-based economic classification to the budgets of all MDAs. This classification is also fully incorporated into the IFMS and the quarterly budget execution reports. However, use of a standard and internationallyaccepted functional classification of expenditures remains absent from the budget documentation and execution reports with respect to both recurrent expenditures and development expenditures. One challenge has been the continued focus on a traditional sector classification in the development expenditures, which does not match with a GFS-based functional classification. This needs to be addressed along with the introduction of a functional classification for recurrent expenditures., 7.9 In the budget preparation guidelines, the MDAs are requested to classify expenditures according to the NSPRG clusters: (i) growth of the economy and reduction of income poverty; (ii) improvement of quality of life and social well being, and (iii) governance and accountability. This seems to be an innovative but unusual expenditure classification that is not directly linked to standard budget classifications. Some selective program classifications, especially in externally-financed operations, can be found in some MDA development budgets, but none of these classifications is linked to a standard functional classification. Government has announced that it intends to introduce a functional classification in the 2009/10 budget. Indicator PI-5 Classification of the budget Scores C Brief Explanation and cardinal data used The budget documentation and execution is based on GFS-based administrative and economic classifications. No GFS-based functional classification is part of the budget documentation and reporting system. Comprehensiveness of Information Included in Budget Documentation (PI-6) 7.10 The annual budget documentation, including the Budget Speech submitted to the National Assembly, includes the following data: 1. Macroeconomic assumptions, including growth rate and inflation, are briefly stated in the budget speech and further supported by a detailed economic review for the previous calendar year. These are submitted separately to the National Assembly, but not as part of the budget documentation. 2. Debt stock is included in the economic review volume, mentioned above. 3. Prior year s budget outcomes in the same format as the budget proposal. 4. Current year s revised budget in the same format as the budget proposal. 5. Budget proposal data in summary and details in several volumes. Missing information benchmarks from the budget documentation are: 1. Fiscal deficit as defined by the GFS or other internationally recognized standards. 2. Deficit financing and its anticipated composition. 3. Financial assets, including data for the current and previous years. 48

65 4. Budget implications of new budget initiatives for expenditures, though these are mentioned for revenue. Interestingly, immediate item 1 and 2 above are included in the quarterly budget execution reports, but not in the budget documentation submitted to the National Assembly. Indicator PI-6 Comprehensiveness of information included in budget documentation Scores B Brief Explanation and cardinal data used 2008/09 budget documentation fulfils 5-6 out of 9 benchmarks of the required information. Extent of unreported government operations (PI-7) 7.11 The level of extra-budgetary expenditure (other than donor-funded projects) that is unreported. Budgets and accounts of AGAs, with the exception of transfers to them from their parent ministries budgets, remain outside government budgeting and accounting systems. This is because AGAs are treated as parastatals and are classified alongside PES, even though they are not public corporations and are financed from earmarked revenue and transfers from the government budget. The MoFEA has the authority to collect budget execution reports from AGAs, but this is normally associated with delays and is not published. AGAs range from universities to the Road Agency or National Parks Agency and several non-for-profit organizations owned and operated by the government. No data is published to indicate the size of unreported expenditures by AGAs that are financed from own sources; but taking into account their size and number, along with available data in the MoFEA, such expenditures are estimated to be between 5-10 percent of total government expenditure Income/expenditure information on donor-funded projects that is included in fiscal reports. While income/expenditures of loan-financed operations are included in the fiscal reports using data available in the debt management unit of the MoFEA and the MDA final accounts, there are still some unreported expenditures using external grant funds. The AGD has recently increased its attempts in this regard through advising MDAs and providing them with specific forms for reporting such transactions on a monthly basis. The AGD then enters this information into the IFMS centrally. However, there are some unreported income/expenditure data from grants where donors directly are spending and/or providing goods and services to the MDAs or LGAs. The level of the latter is far less than 50 percent of all donor grants, even if NGO spending, which is not classified as official aid flow, is taken into account. Indicator PI-7 Extent of unreported government operations Scores C+ Brief Explanation and cardinal data used (i) Extra budgetary spending is estimated to be between 5-10 percent of total government expenditure (C). (ii) Complete income/expenditure information is included in the fiscal reports for all loan-financed operations and at least 50 percent of grant-financed operations (B). Transparency of Inter-governmental Fiscal Relations (PI-8) 7.13 In Tanzania, about 95 percent of LGA (presently 133 urban municipal and rural district councils) operations are financed by different types of transfers from the central government. This demonstrates the low revenue base of the local governments that in fact is limited to small amounts of municipal taxes and service charges. The transfers from the central government to the LGAs are 49

66 mostly sector-based; therefore, they can be classified as conditional grants. In other words, the LGAs must observe sectoral ceilings and limitations but may implement with reasonable flexibility while following the central government s guidelines. In 2007/08 these transfers consisted of: Recurrent block transfers (60 percent), sector basket funds and ministerial subventions (1 0 percent), and development grants and funds (30 percent). Though under different names, almost all of these transfers are sector-based and fully conditional, with the exception of only 8 percent of recurrent block grants associated with a general purpose grant allocated for the improvement of local government capacity Transparent and rule-based systems in the horizontal allocation of unconditional and conditional grants from central government (both budget and actual allocations). In each of the above types of transfers and in each sector covered by them, several quantitative measures, such as population and its composition, number of villages, number of rural population, and some sectorspecific measures initiated by sectoral ministries, are taken into account and budgetary amounts are calculated in great detail. However, these do not constitute a firm formula, and because of the conditional nature of the transfers, sectoral policies need to be taken into account as well. The LGAs do not have a say in these measures, but they are announced in the local government s budget preparation guidelines and are published in the website of the Prime Minister s Office. As for actual allocations, the LGAs are treated like central government MDAs, meaning that in the course of the year their allocations are subject to the government s cash position Timeliness of reliable information to LGAs on their allocation from central government for the coming year. Information on the ceilings of LGAs budgets is known normally three months before the beginning of the fiscal year, and in this regard they are treated as MDAs. However, the LGAs are required to receive their Councils approval before sending their budgets to MoFEA, and then a round of discussions begins between MoFEA and the LGAs. In the event that some other outcome (either ceiling or budget composition) emerges as a result of these negotiations, then the concerned LGAs are required to receive Council approval again. In other words, initially timely data is available to the LGAs on their ceilings, but these are not firm and may change during their budget discussions with the MoFEA. All this is due to the conditional nature of the transfers and the engagement of sectoral policies mandated by MDAs. A balance always needs to be made between central policy-making and decentralized execution, and within changing financial means Extent to which consolidated fiscal data (at least on revenue and expenditure) is collected and reported for general government according to sectoral strategies. Reporting on the local governments operations is of two types. First, when the central government makes monthly transfers to the LGAs, it reports the transfers as its outlays, which can be considered satisfactory in the context of in-year fiscal reporting for general government. Second, for their minimal local funds, as well as for their sectoral distribution of their operations financed from transfers, the LGAs report to the Local Government Working Group, a body comprising representatives of the Prime Minster Office, MoFEA, and relevant sectoral ministries. This data is available in less than three months after each quarter on the Prime Minister s Office website. The main users of these reports are the Prime Minister s Office itself and the relevant sectoral MDAs. 50

67 Indicator PI-8 Transparency of intergovernmental fiscal relations Scores B Brief Explanation and cardinal data used (i) The horizontal allocation of almost all transfers is guided by certain measures, but because they are conditional transfers, sectoral policies also need to be taken into account (B). (ii) Reliable information is issued to LGAs three months before the start of the fiscal year, which is a short period for significant budget changes to be made (C). (iii) Fiscal information that is consistent with the central government fiscal reporting is collected on quarterly basis (B). Oversight of aggregate fiscal risk from other public sector entities (PI-9) 7.17 Extent of central government monitoring of AGAs and PES. AGAs and PES (collectively called parastatals) are required to submit quarterly financial statements and audited year-end statements to MoFEA. These reports, however, are not yet standardized and are frequently received with delays. The 2006/07 NAO s report mentions that there was a large increase in guarantees provided to parastatals, which results in risk o f increased expenditures should the parastatals not be able to repay the loans, thus pointing to a weakness in financial control and risk assessment Extent of central government monitoring of sub-national governments fiscal position. LGAs usually are not allowed to generate fiscal liabilities for the central government through borrowing, unless with special authorization from MoFEA, which is normally not granted. Moreover, their budget execution reports are prepared and published regularly by the Office of the Prime Minister, as mentioned above. Indicator PI- 9 Oversight of aggregate fiscal risk from other public sector entities Scores B Brief Explanation and cardinal data used (i) Most AGAsPEs submit fiscal reports to the central government, but these are normally delayed and their consolidated overview is incomplete (C). (ii) The LGAs do not have borrowing power, unless approved by MoFEA, which is not used, and their budget execution reports are available on quarterly basis (A). Public Access to Fiscal Information (PI-10) 7.19 The government has improved the public access to fiscal information through the dissemination of its reports on the national websites and in government gazettes and local newspapers. These include: 1. Annual budget documentation, as prescribed in PI-6 and as submitted to the National Assembly. 2. In-year budget execution reports (quarterly one page fiscal table with narratives), but not budgets of MDAs either in total or in detail. 3. Year-end financial statements, as they are completed. 4. External audit reports, as they are completed. 51

68 5. Contract awards, published bi-weekly, as reported to PPRA, which may or may not be complete Missing from the list is: Resources available to primary service providers, such as schools and health centers. Indicator PI-10. Public access to key fiscal information. Scores I B l Brief Explanation and cardinal data used The government makes available to the public five of six types of information, but two of them are not complete. POLICY-BASED BUDGETING Orderliness and participation in the annual budget formulation process (PI-1 1) 7.21 Existence of and adherence to a fixed budget calendar. There is a clear budget preparation calendar encompassing macro-fiscal studies, MTEF planning exercises, annual budget policy analysis, budget preparation circular issuance, and budget discussions between MDAs and MoFEA. However, the calendar is always implemented in a manner so that the government budget is presented to the National Assembly in mid-june, just days before the start of the fiscal year. This late budget submission has become an old tradition Claritykomprehensiveness of and political involvement in the guidance on the preparation of budget submissions (budget circular or equivalent). A very detailed and comprehensive budget preparation circular called the budget preparation guidelines is issued to the MDAs and separately to the LGAs, which apart from the budget preparation forms include: economic policy directions, major points of the NSGPR, three-year budget preparation forms, and so forth. The political involvement is secured, normally before communicating budget ceilings to the MDAs, in an attempt to make the budget submissions as affordable and as realistic as possible Timely budget approval by the legislature or similarly mandated body (within the last three years). Following the Budget Speech by the Minister of MoFEA around mid-june, the National Assembly has a roll call vote to approve the budget aggregates, called Finance Bill for revenue and Appropriation Bill for the expenditure of each MDA, which authorize government to implement the budget. Any combination within the total ceiling of an MDA s budget remains at the discretion of the executive branch, which is determined between the MoFEA and the MDA in the course of the budget execution. The late submission of the budget to legislature does not provide sufficient time to the National Assembly to meaningfully debate the government budget. The MDAs, as they discuss with MoFEA and finalize their budget within the ceilings during May, also attend the National Assembly s Sectoral Committees and explain their budgets to them before the budget is formally presented by the government in mid-june. 52

69 Indicator PI-1 1 Orderliness and participation in the annual budget formulation process Scores B Brief Explanation and cardinal data used (i) A comprehensive budget calendar exists, but always ends up with late submission of the budget just before the end of fiscal year, though MDAs need more time to submit their budgets (C). (ii) A comprehensive budget circular and budget preparation guidelines are issued, and the center of government has time to make necessary adjustments (B). (iii) The Legislative normally approves the budget before the start of the fiscal year, but all authority for changes within the approved total level of a MDA s budget remain within the power of the executive branch (B). Multi-year perspective in fiscal planning, expenditure policy and budgeting (PI-12) 7.24 Preparation of multi-year fiscal forecasts and functional allocations. The government budget system includes a three year rolling MTEF, but this has key limitations on the outer years expenditure projections. The budget preparation guidelines, the principal decision making process for the framework, is insufficiently strategic. The budget ceilings in the budget guidelines have little to do with the previous year s MTEF. The formal part of the MTEF continues vigorously, but because of short-term emerging needs, the rolling plans need to be adjusted widely to the annual budget realities. Moreover, in the absence of a functional classification of expenditures, the MTEF normally follows an administrative classification, which is less relevant to an MTEF Scope and frequency of debt sustainability analysis. A debt sustainability analysis (DSA) was carried out in April 2004, including both domestic and external debt. A new DSA was carried out in 2006 but has not yet been made public, although it is available to development partners involved in its preparation. It has been decided to conduct a full DSA only every two years as debt has been brought under control and generally does not change much from year to year. Additionally, the IMF, in collaboration with the World Bank, has carried out two DSAs in 2005 and The 2007 DSA covers domestic and external debt Existence of sector strategies with multi-year costing of recurrent and investment expenditures. Fully costed sector (or sub-sector) strategies cover most sectors but tend to be inconsistent with aggregate fiscal forecasts. The sectoral strategic plans are not being followed, sometimes to the extent that the implemented budget may not be totally relevant to the sectoral policies, and they may not represent the real priorities of the government. Part of the reason behind this is lack of a realistic connection between the sector strategies and domestic resources, coupled with the reliance on donor funds, which by themselves are not clearly known for longer periods, especially on program and project funds. On the other hand, the resultant drastic reductions to reach affordable budget ceilings together with the late involvement of senior policy-makers in the process make it difficult for stakeholders to see any transparent application of clear sector priorities Linkages between investment budgets and forward expenditure estimates. Separate recurrent and development budgets are prepared under MoFEA s coordination, but there is limited integration of recurrent and development expenditure proposals in the planning process. There are even two different budget classifications and volumes for recurrent and development budgets. Moreover, the 53

70 separation of these two budgets is based on their financing source rather than the nature of their operations. For example, considerable amounts of recurrent expenditure are classified as development expenditures simply because they are financed from external sources. Linkages between investments in different sectors are not being analyzed. There is no overall public investment programming process, but instead, political priorities (not always as reflected in sectoral expenditure programs) andor donor preferences are the main drivers behind the more sizeable investment projects. The links between investments and the recurrent cost implications of these investments are weak. Scores PI- 12 rn ulti-year perspective in fiscal planning, expenditure policy and budgeting (i) Forecasts of fiscal aggregates, on the basis of main categories of administrative and economic classification for recurrent expenditures and project or sector-based for development expenditures, are prepared for two years in addition to budget year 0. C (ii) DSA for external and domestic debt has been undertaken in at least two out of the last three years (B). (iii) Sector strategies exist, but they are inconsistent with aggregate fiscal forecast (C). (iv) Linkages among investment budgets and sector strategies and recurrent budgets are weak (D). PREDICTABILITY AND CONTROL IN BUDGET EXECUTION Transparency of taxpayer obligations and liabilities (PI-13) 7.28 Clarity and comprehensiveness of tax liabilities. Tanzania has relatively new and up-todate income tax (2004) and VAT (1997, and as amended in 2006) laws. A number of high profile dispute resolution cases in recent years indicate that this legislation is being enforced. The East African Customs Management Act (2005) is relatively comprehensive. Revenue administration procedures are clearly documented and uniformly implemented. Overall, discretionary powers have been reduced and the clarity of taxation liability has improved. Still outstanding is legislation regarding the development of the taxation procedures court. This involves harmonizing various taxation law regulations and, in doing so, reassigning the power from the other Acts. Standardization focuses on dates for the submission of tax returns, penalties for non-submission, and objections and appeals procedures. Common administrative procedures are, in some cases, applied differently across the income tax and VAT laws Taxpayer access to information on tax liabilities and administrative procedures. TRA has developed some comprehensive taxpayer education material, in both English and Kiswahili, and has an active taxpayer education program across the country involving communications in newspapers, the radio, and billboard advertising. TRA, however, does not yet have full-fledged tax information centers, phone-in call centers, and an advance rulings regime. Due to lack of a comprehensive and integrated tax administration system, the taxpayers may not know precisely what their liabilities are, especially since the self-assessment system is only beginning to operate. The positive side is that some information is available on the TRA website, including the taxation laws and the latest associated communications. 54

71 7.30 Existence and functioning of a tax appeals mechanisms. Tanzania has an independent disputes resolution system funded separately by government. The Tax Revenue Board and Tribunal has been established, although this is not used in all cases. There are still residual cases with the court system since the establishment of the appeals Board and Tribunal under the Tax Revenue Appeals Act of However, it is estimated that the number of such cases is small and reflects the nature of the work rather than a systemic and serious backlog. Indicator PI-13 Transparency of Taxpayer Obligations and Liabilities B Brief Explanation and cardinal data used (i) There are relatively new and comprehensive income tax and VAT laws (B). (ii) Taxpayer access to information on tax liabilities and administrative procedures. There is good taxpayer education systems but no call center or central information system (B). (iii) There is an independent disputes resolution system funded separately by government, though it needs further capacity enhancement (B). Effectiveness of measures for taxpayer registration and tax assessment (PI-14) Controls in the taxpayer registration system. Taxpayers have a single identification number. A block management system is in place, and currently, a large scale review of business is being implemented in a number of blocks in Dar es Salaam. Inconsistency still exists, as linkages within TRA are weak due to limited connectivity between the integrated tax administration system for VAT and income tax) and customs. Linkages with external systems such as the Tanzanian Bureau of Standards and the Ports Authority are nonexistent. There have not been any surveys of potential tax payers Effectiveness of penalties for non-compliance with registration and declaration obligations. Clear penalties are provided for under the income tax and VAT laws; penalties are too low, and negotiating an increase is a lengthy and challenging process. Penalties under the income tax and VAT laws not harmonized; therefore, different penalties for similar offenses, such as non submission of returns, may apply. There are no guidelines for penalties, which can lead to inconsistencies. For example, there is no guidance for the application of different penalties to first time or repeat offenders and for different types of offenses Planning and monitoring of tax audit and fraud investigation programs.a comprehensive and documented audit plan has been developed and implemented in the Large Taxpayers Department (LTD), which accounts for 70 percent of revenue but less than 5 percent of taxpayers. The audit plan is being operated effectively, with good risk profiling and quality assurance. The Domestic Revenue Department (DRD), which accounts for more than 90 percent of the taxpayer population, has an audit plan and manual, but implementation has not been possible due to capacity limitations. 55

72 Indicator Scores Brief Explanation and cardinal data used PI-14 Effectiveness of measures for taxpayer registration and tax assessment C+ (i) Taxpayers are registered in a database with an identification system, but there are no links to other systems (C). (ii) Penalties exist, but they are insufficiently specific and in some cases too low to have an impact on compliance (C). (iii) There is a good tax audit operation in the LTD (the main tax collection department), but a mixed performance in DRD (B). Effectiveness in collection of tax payments (PI-15) 7.34 Gross tax arrears collection ratio: the percentage of tax arrears at the beginning of a fiscal year collected during that fiscal year, expressed as an average of the last two fiscal years. The debt collection ratio in 2004/05 and 2005/06 was 33 percent according to LTD data and 71 percent according to LTD and DRD data. This indicates an improvement, but tax arrears remain significant. The improvement of the debt collection ratio in those two years suggests that some debts have been written off, and it is important to note the process used to select which debts are determined to be uncollectible, but unfortunately, it is not possible to get data to verify this. The process for writing off debt requires a Board resolution, which necessitates the submission of a list stating whether the taxpayer can t be traced or has gone out of business Effectiveness of transfer of tax collections to the Treasury by the revenue administration. The individual accounts for taxpayers are posted, and transfers are made to the Treasury main account, called the Paymaster General s account, twice a week. The LTD taxpayers pay directly into the treasury main account Frequency of complete accounts reconciliation between tax assessments, collections, arrears records and receipts by the Treasury. The authority to undertake assessments, collections, arrears and transfers is delegated from MoFEA to TRA, as a department of MoFEA. Tu s functions are to assess, collect and account for tax revenue. Assessments are raised and logged in the ledger account, as are collections when they are made. Arrears are the difference between the assessment and the collections. Therefore there are regular and routine reconciliations. Indicator PI-15 Effectiveness in collection of tax payments Brief Explanation and cardinal data used (i) Collection ratio for gross tax arrears in 2004/05 and 2005/06 was 33% (LTD data) and 71% (LTD and DRD data) (C). (ii) Revenue collected is transferred to the government main account in the BOT within a week (B). (iii) Reconciliation between tax assessments, collections, arrears records and receipts by the Treasury takes place as a matter of routine (A). 56

73 Predictability in the Availability of Funds for Commitment of Expenditures (PI-16) 7.37 Extent to which cash flows are forecast and monitored. After approval of the budget, cash flow projections and plans are prepared by the MDAs on a monthly basis and submitted to the MoFEA Budget Department. MDA cash outflow requirements are thereafter forwarded to the MoFEA Cash Management Committee, chaired by the Permanent Secretary, which determines the corresponding monthly ceilings for each MDA as well as transfers to the LGAs. These ceilings are mainly determined based on overall government's cash availability for the month, with a view to the MDAs' previous monthly implementation reports, their work plans and procurement plans The cash management system is still nascent and has not been fully implemented. The main challenge is that monthly projections are not yet produced by all the spending agencies. This makes it impossible to compile the full cash flow picture and allocate resources accordingly. In addition, the cash flow projections received by MoFEA tend to be frontloaded, making them difficult to implement under a cash budgeting framework. A draft report on the implementation of the cash management system indicates that the lack of an established administrative structure and expertise to handle the system rollout are the underlying causes for the lack of a coherent cash management system. These may largely be regarded as growing pains of this new process. Nevertheless, strengthening the cash management unit needs to be 'a priority for FY 08/09. In addition, the integration of the cash flow forecasts with the relevant module of the EMS should be pursued in order to enhance the systems' coherence. Currently, the cash management for procurement plans is not captured in the IFMS, which uses a pro rata allocation (divide by twelve) of the budget provisions as the monthly budget plan. These factors affect the resource allocation process within spending agencies. Since monthly releases tend not to follow a cash flow plan, a time consuming resource allocation process takes place each month within the spending units, causing delays in the deployment of resources.' Reliability and horizon of periodic in-year information to MDAs on ceilings for expenditure commitment. The existing cash management system, though very useful for maintaining overall macro-fiscal balance and controlling total government expenditures, by its nature creates uncertainty for the MDAs. MoFEA is aware that the monthly cash release system has undesirable consequences in terms of predictability of funding for the MDAs and thus affects their effectiveness, including service delivery. Without these monthly ceilings and their associated fund releases, commitment of funds is not possible, and processing bulk purchases for MDAs is difficult, though according to MoFEA staff, the CMC takes into account the urgent needs for bulk purchases in each month. A strong and predictable quarterly fund allocation for commitments, including monthly limitations for actual payments, may make the cash flow projections a more helpful tool for both maintaining macro-fiscal balance and securing predictability for commitment of funds by the MDAs Frequency and transparency of adjustments to budget allocations, which are decided above the level of management of MDAs. Budget adjustments take place throughout the year, based on MDA requests and MoFEA approval and/or due to lack of sufficient cash, which forces the MDAs to reallocate funds within their approved budgets. The adjustments are consolidated and submitted to the National Assembly at any interval for its information. Because the adjustments are within a given MDA budget, they do not require approval by the National Assembly. In the event that an MDA budget is increased, a supplementary budget submission to legislature is required. A contingency is also made for payroll adjustments and annual pay increases as well as for other needs that may arise in the course of the budget year. MDA budget reallocations are made routinely, as the monthly cash allocations necessitate such changes, and this mainly remains at the discretion of the MDAs. By its '' MOFEA, The Implementation of the Cash Management System, May

74 nature this is not fully transparent because the MDAs re-allocate their appropriation within the cash available to them month by month. Indicator PI-16 Predictability in the Availability of Funds for Commitment of Expenditures C+ Brief Explanation and cardinal data used (i) An annual cash flow forecast for government budget and each MDA is prepared, but substantially revised on a monthly basis (B). (ii) MDAs are provided with reliable information for their commitments, but only for each month and with relatively short notice (C). (iii) Significant in-year budget adjustments are frequent, but undertaken with some transparency (C). Recording and management of each cash balances, debt and debt guarantees (PI-17) 7.41 Quality of debt data recording and reporting. Debt is recorded and managed by the AGD using the CS-DRMS (a debt management system developed by the Commonwealth Secretariat), which provides information on servicing and repayment of the principal for foreign debt and a portion of domestic debt, the information of the latter being complemented by the BOT, as it manages government treasury bills. The debt database is updated on a monthly basis, including data received from the BOT. While some discrepancies exist, they are not large and are mainly due to the different timing of updating of the information. Reconciliation of the debt database is done on a monthly basis, using the system s reports and the creditors records. Monthly debt profiles (external and domestic), a quarterly newsletter, an annual debt report, and annual statements of public debt are prepared by the AGD Consolidation of the Government s cash balances. Apart from the main Government bank account held at the BOT, a number of government bank accounts are held in the commercial banks under different arrangements. This reduces the government s liquidity position while cash available in a number of bank accounts is not accessible to the MoFEA for its monthly MDA and LGA cash allocations in the budget execution. Due to different legal and/or managerial arrangements, until recently no attempt has been made to consolidate the balances of these accounts in the BOT, thereby allowing accumulation of idle cash in different government bank accounts. Some of these accounts hold donor funds. The MoFEA Minister announced in his 2008/09 budget speech that in order to improve MoFEA s cash position, the Government intends to transfer balances of these bank accounts from the commercial banks to the Government main account at the MOT, and that the government has begun discussions with donors. Apart from changing the rules governing government s own bank accounts in the commercial banks, such action, would also. require agreement of individual donors that are holding mostly project-related bank accounts in the commercial banks Contracting loans and issuing guarantees. The National Debt Management Strategy adopted in 2002, as well the Loans, Grants and Guarantees Act of 2003, outline comprehensive procedures for contracting and guaranteeing loans. Since 2004/05, the government has a policy of not providing guarantees for any external borrowing, but rather of considering selective guarantees for domestic borrowing, mostly by public enterprises An implementation plan for public debt management has been developed in line with the Strategy and the Act mentioned above. The National Debt Management Committee advises the MoFEA Minister on the contracting of debt based on an evaluation against set criteria, including 58

75 viability and suitability. Only the Minister has the authority to contract new debt. At the same time, the 2006/07 NAO report recommended that debt management be improved to ensure accurate information at all times and that communication be improved between all those involved. It is not clear whether issuing of guarantees are made within limits for total debt and total guarantees. Indicator Scores Brief Explanation and cardinal data used PI-17 Recording and management of each cash balances, debt and debt guarantees C (i) Debt data is of a relatively high quality but minor reconciliation problems occur, and timely statistical reports are produced (B). (ii) The balances of several government bank accounts in commercial banks are not consolidated, though there is a plan to do so in the future (D). (iii) Contracting of loans and issuing guarantees is approved by the Minister of MoFEA in line with rules, but it is not clear if ceilings apply (C). Effectiveness of payroll controls ( 1-18) 7.45 Integration and reconciliation between personnel records and payroll data. Personnel records kept at the President s Office-Public Service Management (PO-PSM) and payroll data kept at MoFEA are totally integrated and use the same software: a Lawson human resources software package called Human Capital Management Information System (HCMIS). The system is operated by the MoFEA Computer Center, which is connected to the PO-PSM. MDAs undertake monthly reconciliation between payroll and personnel records and send them to the MoFEA and PO-PSM, where relevant personnel records and payroll data are entered into the system prior to payment of salaries Timeliness of changes to personnel records and the payroll. Every month, MoFEA and PO-PMS receive approximately 10,000 MDA and LGA requests to change the payroll and personnel records. There is limited capacity to verify these requests in the two weeks currently allocated for monthly payroll data entry. Approval of these requests is generally fast, based on trust that the officers in the MDAs have checked and verified the requests before making their submissions. Although HCMIS is designed to assist the officers in authenticating most requests, each request cannot be checked through the system. Because salary payments are time-sensitive, the data processing both in the MDAs and the two central agencies is quite timely, but its accuracy is not assured, due, in part, to low quality of work in the MDAs and shortage of time for all parties involved. As mentioned below, there are reasons to be concerned about this process Internal controls of changes to personnel records and the payroll. The payroll system data is currently accessible to several users (operators and systems analysts), which poses challenges in protecting the payroll information from being lost, read, changed (either maliciously or accidentally), or modified by those not authorized to do it. Some internal controls are incorporated into the system, but these are not sufficient to prevent entries by unauthorized operators. The main risk area, however, originates in the MDAs, where operators are close to beneficiaries and thus more easily inclined to manipulate data than operators based in the central agencies. Auditing reports have frequently mentioned deficiencies in the MDAs and LGAs, indicating that central internal controls are not sufficient or even fully practical Existence of payroll audits to identify control weaknesses and/or ghost workers. Payroll audits are conducted for a sample of MDAs and LGAs on a quarterly basis. Usually, officers from 59

76 PO-PSM constitute teams that visit selected MDAs and LGAs. A key audit challenge is that the payroll data in the current payroll system are not accurate and the human resources information is incomplete. Therefore, the selection of the MDAs to be audited may be triggered by various events, including unusual payroll changes requests from a MDA, a static payroll that never changes over the year (ghost workers), or even someone receiving both pension and salary for some time. In such cases, the audits are not sufficiently funded to cover an entire MDA, such as the Ministry of Health or Ministry of Education, but rather focus on specific problem institution, such as a school or health facility As a result of these audits, it was recently observed that records are not being kept up to date by the MDAs, and MoFEA issued a circular to the MDAs to do so. In 2000, the government computerized the payroll by collecting data from employees. Since then, payroll data has been changed through payroll amendment requests submitted by MDAs to MoFEA and PO-PSM. Since 1994, no survey has been conducted for government employees, now totaling about 350,000 (including about 200,000 teachers). A complete employee census is planned for 2009 in order to overhaul the payrowpersonne1 database. Brief Explanation and cardinal data used PI-18 Effectiveness of (i) Personnel records and payroll data are stored in the same database and are filly integrated and reconciled (A). C+ (ii) Changes to personnel records and payroll data take place on a monthly basis, but its quality is not assured (C). (iii) Some internal controls exist to data entry for changes on personnel records and payroll data, but are not adequate to ensure the integrity of data, which originate in the MDAs (C). (iv) Quarterly payroll audits with limited coverage have been undertaken during the last three years, but no survey has been conducted since 1994 (C). Competition, value for money and controls in procurement (PI-19) 7.50 A new Public Procurement Act (PPA) was enacted in 2004 and, together with its associated regulations, became operational on May 1, The Public Procurement Regulatory Authority (PPRA) was established and is now fully staffed. The PPRA is responsible for the oversight of public procurement and overall capacity building in the procuring agencies. The Authority has developed a system for checking and monitoring procurement activities in the MDAs, LGAs, AGAs, and PES. Recent work of the PPRA has included the training of approximately 1200 staff for procuring agencies. Some government agencies have not yet established their Procurement Boards or Procurement Management Units During 2007/08 the PPRA carried out procurement audits in 70 procuring agencies. However, 30 further audits for the financial year 2007/08 have yet to be completed. The audits objective was to determine whether the procedures, processes and documentations for procurement, contracting and disposal of public assets by tender were in accordance with the provisions of the PPA regulations. Further, they were to establish whether the standard documents prepared by PPRA and those procurements carried out achieved the expected economy and efficiency, and whether the implementation of contracts conform to the terms therein. The audits indicated an average level of 60

77 compliance of 43 percent computed from the 13 established compliance indicators. The average level of compliance was 45 percent for MDAs and 41 percent for LGAs. The Tanzania Ports Authority attained a maximum average compliance of 73 percent while the Tanzania Library Services attained a minimum average compliance of 7 per~ent.'~ 7.52 Use of open competition. According to the NAO's 2006/07 report, 66 percent of tenders under the open tendering process in that year were advertised. However, the report pointed to a weakness in the publication of contracts awarded, as a majority of them had not been communicated properly to the public. The PPRA's own periodic audit reports, which are also published, show similar results. Presently, the PPRA's bi-weekly publication and website contain a wealth of information on the procurement plans and bids by government agencies, as well as contract awards, though the full coverage of the latter has to be verified by further procurement audits Competitive procurement. The procurement act allows the use of less competitive procurement methods with justification. Although PPRA's professional views may be sought, the procuring agency is responsible for interpreting the law and choosing such methods, the quality of which may differ from one agency to another Existence and operation of a procurement complaints mechanism. The legislative framework has a very comprehensive complaints mechanism. A three level complaint mechanism has been established consisting of the procuring agency, the PPRA, and the Public Procurement Appeals Authority (PPAA). The model for complaints handling is in line with international practices and gives aggrieved bidders adequate means to protect their interests. It has, however, been observed that although the PPRA is timely in processing these complaints, the number of complaints is declining. While this may be interpreted as a good development, it is also argued that some local contractors avoid filing complaint cases, fearing the risk of being excluded from future bids. i Indicator Brief Explanation and cardinal data used PI-19 Competition, value for money and controls in procurement B (i) 66 percent of tenders under open tendering process were advertised in fiscal year 2006/2007 (B). (ii) Using less competitive procurement is allowed with justification. Although PPM's view may be sought, observing the procurement act remains within the power of procuring agencies (B). (iii) A comprehensive complaints mechanism operates, but for unknown reasons the number of complaints has declined (B). Effectiveness of internal controls for non-salary expenditure (PI-20) 7.55 Effectiveness of expenditure commitment controls. As part of the functionality of the IFMS, a commitment control system was introduced in 2001 under which a Local Purchase Order (LPO) is required for the purchase of goods and services. The system is relatively effective, restricts the production of an LPO to financial codes with adequate funds, and is issued only when the resources are released to the MDAs. On the other hand, the system has provided MoFEA with a rebudgeting power during the year, creating uncertainty among the MDAs and LGAs. In fact, the accumulation of payment arrears in recent years, as noted in IP-4 above, indicates that some unpaid Further information, including indicators used for audit compliance is available on the PPR4 website. 61

78 commitments at the end of fiscal year may be related to commitments, such as utilities or price escalations embodied in some contracts that cannot be prevented by the IFMS Comprehensiveness, relevance and understanding of other internal control rules/procedures. There is a comprehensive set of controls, although in some instances concern is expressed that they can excessive. These include payment for casual labor, minor goods, and services as well as collection and handling of minor non-tax service charges and administrative fees, safekeeping, store management, and similar items covering a long list that is generally understood by MDA staff Compliance with rules for processing and recording transactions. Levels of compliance vary. According to the NAO s 2006/07 report, of the 70 MDAs audited, eight had outstanding bank reconciliation items such that the discrepancy between receipts and bank statements pointed to the potential for revenue misappropriation, or there was vouched expenditure where a payment voucher existed but was not supported by other documentation, such as LPO, invoice, delivery note, or expenditure statement. The report further noted that a large improvement in reducing the amount of improperly vouched expenditures had taken place, decreasing such cases significantly from previous years. Indicator PI-20 Effectiveness of internal controls for nonsalary expenditure Scores C+ Brief Explanation and cardinal data used (i) The use of IFMS with a commitment control feature is relatively effective, but cannot provide a full guarantee at all times (B). (ii) Other internal rules and procedures cover major sets of controls that are generally understood, but in some areas these are excessive and compliance rate varies (C). (iii) Rules are observed in a majority of transactions, but there is some evidence of misconduct (C). Effectiveness of internal audit (PI-21) 7.58 Coverage and quality of the internal audit function. According to the NAO s 2006/07 report, most MDAs have established Internal Audit Units. Most are understaffed and ineffective. Consequently, they are able to undertake only audits of narrow scope with minimal follow-up on findings. A rough estimate is that 4 percent of MDAs do not have an internal audit committee. Where they are established, they lack charters or do not meet quarterly as they are expected to do. The NAO s report emphasizes the need to increase capacity (qualified staff and regular training) and to establish audit committees properly. Establishing a central internal audit unit at the AGD in September 2006 was a step toward enhancing the effectiveness of the internal audit function. This unit is now operative and has undertaken some audits and identified activities to be undertaken Frequency and distribution of reports. Internal audit reports are issued for most MDAs, but these are not routinely relied upon, in part due to the weak capacity in most MDAs, as highlighted in the NAO s report. Reports have been standardized with some improvement in quality. The NAO, however, still reports that a considerable number of cases of previous years audit issues remain unresolved. 62

79 7.60 Extent of management response to internal audit findings. Internal audit findings followup is limited, although Audit Committees in the MDAs are tasked to do so. The central internal audit unit is working to change the style of auditing, which includes agreeing on a timetable with the client for the follow-up of queries or recommendations, with the objective of improving the quality and timing of follow-up. Due to the large numbers of MDAs, it has not been possible to monitor the follow-up of all audit queries. Indicator PI-21 Effectiveness of internal audit Brief Explanation and cardinal data used (i) Internal audit fhction exists in most MDAs, and it is estimated that 20 percent of the staff time is allocated to systembased reviews and high risk areas (C). D+ (ii) Reports are issued for most MDAs but these are not submitted either to the AGD or the NAO (C). (iii) To some degree actions are taken by management on major issues but often with delay (D). ACCOUNTING, RECORDING AND REPORTING Timeliness and Regularity of Accounts Reconciliation (3'1-22) Regularity of reconciliation of bank accounts. Transactions of the central government's bank accounts (main treasury account held at the BOT and other accounts with smaller amounts held in commercial banks) are reconciled with the government accounting records at the end of each month, but with substantial delays, and its quality is questionable. Although with some delay, the AGD produces lists of unreconciled items and submits them to the MDAs to identify matching pairs and enter them into an adjustment model for reconciliation, the process is unreliable and at times incomplete. The bank reconciliations are reviewed by the AGD and subsequently made available to a resident NAO staff member in MoFEA. Delays in the monthly reconciliation and the size of the unreconciled transactions make this process less effective, but these issues are taken more seriously for the year-end final accounts as required by law. However, less effective reconciliation between government accounts and their corresponding bank transactions lead to lack of proper bank reconciliation in the annual accounts as well. The 2006/07 NAO report disclosed such unreconciled items, noting that payments were recorded in bank statements but not recorded in cash books (government accounting records) Regularity of reconciliation and clearance of suspense accounts and advances. Reconciliation of suspense accounts and clearance of advances is also a long process, and again it appears that the year-end reconciliation and clearance is taken more seriously than the monthly process. Suspense accounts are normally reconciled within two months of the end of the period, which in principle is monthly because of the bank accounts reconciliation cycle of government transactions. But the size of unreconciled items is not known, nor does the reconciliation exercise happen in a comprehensive and timely manner. Advances, such as travel advance and imprests, are small in size, but their clearing time varies. In fact, these factors may be contributing to the size of unreconciled bank account and accounting reconciliation mentioned above. 63

80 ~ accounts Indicator Scores Brief Explanation and cardinal data used * PI-22 Timeliness and Regularity of Accounts Reconciliation C (i) Bank reconciliations for Treasury-managed accounts are undertaken monthly within 4 weeks, but the size of unreconciled items seems to be large (C). (ii) Suspense accounts are reconciled and advances are cleared in monthly periods, but the size of unreconciled uncleared items is not known. A significant number of have uncleared balances brought forward. (C). Availability of Information on Resources Received by Service Delivery Units (PI-23) 7.63 Collection and processing of information to demonstrate the resources that were actually received (in cash and kind) by the most common front-line service delivery units (focus on primary schools and primary health clinics) in relation to the overall resources made available to the sector@), irrespective of which level of government is responsible for the operation and funding of those units. In last three years, MoFEA has carried out a number of expenditure tracking surveys on LGAs and their Direct Service Delivery Units, most recently in MoFEA has not finalized any of the reports and cleared them for publication. As a result the coverage, methodology and presentation of the findings of these reports are unknown. In 2004 a donor-supported PETS was undertaken in the education sector, but its findings and results were not published because of disagreement between the government and the donors who helped carry out the exercise. Presently, a PETS is being undertaken for the water sector with the help from the World Bank, and a new PETS for the education sector is under discussion for 2009, also with the support of the World Bank. Indicator Scores Brief Explanation and cardinal data used PI-23 Availability of Information on Resources Received by Service Delivery Units Information on resources received by front line service delivery units is mostly lacking. Special surveys were undertaken within the last three years, but their results and methodologies used have not been published. Quality and timeliness of in-year budget reports (PI-24) 7.64 Monthly in-year budget execution reports, called flash reports, are produced by the AGD, mainly for internal use. The Policy Analysis Department also publishes reports on the MoFEA website on a quarterly basis for revenue, expenditures, deficit and financing (a fiscal table) Scope of reports in terms of coverage and compatibility with budget estimates. The IFMS has the capacity to produce timely central government expenditure reports (including transfers to the LGAs) with any coverage that may be needed. However, the actual coverage is limited to the vote level (total expenditures of a MDA) with no information on its sub votes, economic classification, or operations. These flash reports are itemized by original and revised budgets, funds made available by MoFEA to each MDA, and commitments and payments. The flash reports have limited circulation, and their usage, though very helpful, is mainly for total cash management purposes, as there is no further detailed information within an MDA. This mainly stems from the appropriation structure of Tanzania (total budget of each MDA), by which the executive branch has 64

81 the full authority to introduce changes and reallocations within any MDA s total budget. Flash reports do not cover the revenue side of the budget. The quarterly reports, on the other hand, include all fiscal data as mentioned above Timeliness of the issue of reports. Since the State House and the Ministry of Defence are using the IFMS on a stand-alone basis outside the network, they provide data to the AGD on backup tapes that are restored in the AGD s main server. As a result, the consolidated expenditure flash reports are not produced until around two weeks after the end of the month. Quarterly budget execution reports are published on MoFEA s website, usually within three months after the end of each quarter Quality of information. Because flash reports are prepared before reconciliation of government accounts with corresponding bank transactions, by their nature they cannot provide a full picture, but mostly serve as a timely indication of commitments and payments and the status of the budget execution in its total magnitudes. The weakness of the quarterly reports, on the other hand, is that the data is classified in very broad categories. That is mainly because MoFEA needs to collect data from different sources on grants, revenue, domestic and external financing, fiscavmonetary data gaps, and other items. Indicator PI-24 Quality and timeliness of inyear budget reports. Scores B Brief Explanation and cardinal data used (i) Scope of flash reports is compatible with the budget estimates for both commitments and payments, but in very broad categories. Revenue is also classified very broadly (B). (ii) Data are provided monthly on flash reporting for expenditures and quarterly for all fiscal components, including revenue, expenditure deficit, and financing (A). (iii) Information is accurate in flash reports, but unknown in overall quarterly fiscal reporting due to multiple sources of data (C). Quality and Timeliness of Annual Financial Statements (PI- 25) 7.68 Completeness of the financial statements. The financial statements for the central government transactions (including its transfers to the LGA) include revenue, expenditures, and bank balances, but not complete assets and liabilities. The 2007/08 financial statements include a sort of financial liabilities, thanks to changing the accounting standards (see below), but the assets side is still not reported Timeliness of submission of the financial statements. All MDAs are required to submit their final accounts to MoFEA and NAO within 3 months after the end of the fiscal year. The consolidated government financial statements are then prepared by the AGD and submitted to the NAO within an additional month. For the last three fiscal years, these were produced and submitted to the NAO within the statutory four month period, with only two weeks delay for the accounts of the 2007/08 due to changing the accounting standards Accounting standards used. Until the end of fiscal year 2006/07, accounts were prepared using local cash accounting standards. Beginning 2007/08 the government has adopted the 65

82 internationally recognized IPSAS cash accounting standards, but a full reporting of assets and liabilities required under such standards is yet to be established. Indicator PI-25 Quality and Timeliness of Annual Financial Statements B Brief Explanation and cardinal data used (i) Central government final accounts include revenue, expenditure and bank balances, and since , some data on liabilities but not on assets (C). (ii) Financial statements are submitted for external audit within 6 months of the end of the fiscal year (A). (iii) IPSAS cash accounting standards are used, but still assets and liabilities are not filly reported (C). EXTERNAL SCRUTINY AND AUDIT Scope, nature and follow-up of external audit (PI-26) 7.71 Scope and nature of audit. The last three years NAO annual reports cover all central government MDAs and LGAs. By law, all AGAs and PES are supposed to submit their final accounts and financial statements to the NAO, but some fail to do so in a timely manner. As a result, the NAO audits whatever reports are received from the AGAs and PES within 3 to 4 months of their accounting period, issues its report in a reasonable time span, and mentions the same limitation in the audit report. The MDA and LGA audits comprise primarily transaction level testing, and the audit report is primarily of a financial nature (whether accounts have been properly kept, rules and procedures followed, resources expended for the purposes appropriated, and records maintained) with very little attention to performance audit. The audit is carried out on a test basis, therefore the audit findings are confined to the extent that records, documents and information requested for the purpose of audit are were made available to the NAO However, the audit report mentions various aspects of the PFM processes, such as compliance with the Public Procurement Act, internal controls, internal audit functions and audit committees. Further, the audit report analyzes major findings of the accounts under auditing, mainly in the form of organizational cross-cutting issues, but at the same time explains certain examples from the MDAs to substantiate its findings The audits broadly adhere to appropriate auditing standards (INTOSAI) and the international standards on auditing issued by the International Federation of Accountants. However, there are still some deficiencies in terms of meeting international standards; full compliance is expected to be achieved after The passage of the new act is considered an important factor in improving further independence of the NAO as it addresses staffing issues, budget allocations and the coverage of the external audit task. At present, however, comparing audit reports in the last three years in terms of style, nature and quality of the fiduciary work does not indicate a distinct change to be pointed out, and all recommended or initiated changes are expected to bear fruit in the future Timeliness of submission of audit reports to legislature. Audit reports of the last three years, including government financial statements, have been submitted to the legislature within 8-9 months of the end of the period covered. The audit reports are made available to the public through NAO s website. 66

83 7.75 Evidence of follow-up on audit recommendations. Follow-up on addressing the external audit recommendations by the executive branch has been generally weak. Following donors request, some responses were prepared by the MDAs and LGAs on the NAO report and were submitted to NAO, the results of which should be observed in the future audit reports by indicating that how the audit findings were addressed or cleared. The NAO indicates that these responses first are expected to be discussed with the PAC in January 2009, when the audit report is scheduled for discussions (see further on this long time lag in indicator 28 below). Until now, very limited evidence has been observed in addressing or clearing external audit recommendations. A new external audit law passed the National Assembly and was made public in a September 2008 gazette, which is expected to bring about some wide-ranging improvements in the external auditing task, including further independence of the NAO, and response of the executive government to the NAO s findings. The NAO is drafting the enabling regulations of the new law, so it will take some time to assess the impact of the new law. Scores PI-26 Scope, nature and follow-up of external audit (i) The audit report covers all MDAs and LGAs, but part of the AGAs and PES. The nature of the audit remains centered on transaction level testing, but reports identify some significant issues (C). C+ (ii) In the last three years, the audit report, including consolidated financial statements of government, was presented to the legislature within 8-9 months after end of period covered (B). (iii) MDAs and MFOEA submitted formal responses to the NAO on only the audit report of the 2005/06. These are due to be discussed by the PAC in 2009 after a long time gap. No evidence can be demonstrated for a systematic follow up on audit findings by the MDAs (C). Legislative scrutiny of the annual budget 7.76 Scope of the legislature s scrutiny. The National Assembly is provided with information in regard to macro fiscal policy mostly in the form of previous calendar years economic performance (though not as budget documentation series) and detailed revenue and expenditure data. This provides a reasonably good opportunity for the legislature s scrutiny of the budget. But as explained below, the time for such analysis and scrutiny is very short, and in practice it becomes a formal action rather than one of substance. In an attempt to overcome this, the Sectoral Committees of the National Assembly and the MDAs officials meet before the budget presentation and the MDAs explain their draft budgets to the committee members before the government budget is formally presented. There is no approval or disapproval process at this stage; it is a sort of briefing on the MDAs draft budget, with no information on fiscal policies and aggregates of the budget s resource envelope. When the government budget is formally tabled, normally two weeks before the start of the budget year, approval is given to the appropriation bill, which is a very brief document containing the total budget of each MDA, within the short time available to the National Assembly Extent to which the legislature s procedures are well-established and respected. In addition to the PAC, the Local Authority Accounts Committee, and the Parastatal Committee, there are 11 sectoral committees associated with the review of the budget prior to its submission to the 67

84 National Assembly. There are some established internal procedures for committee meetings, but since the nature of debate is mostly in the form of briefings only, their effectiveness is limited Adequacy of time for the legislature to respond to budget proposals with both the detailed estimates and, where applicable, for proposals on macro-fiscal aggregates earlier in the budget preparation cycle (time allowed in practice for all stages combined). In practice the legislature has only two weeks or less to approve the government s proposed budget. The ability of the National Assembly to question or influence inter-sectoral allocations and to ensure that they follow sectoral policies is therefore very limited. The technical capacity of the sectoral committees is also limited Rules for in-year amendments to the budget without ex-ante approval by the legislature. Rules exist for in-year amendments within the appropriation structure and total budget of a MDA. The MDA must submit a request to the MoFEA by completing a standard form. Such requests are discouraged within the first half of the fiscal year, as it is recognized as poor planning, though they do still arise. The requests are submitted and approved as and when needed. These virements are consolidated into a Reallocation Warrants, which are submitted to the National Assembly for its expost information, normally during the second half of the fiscal year. No specific limits are set on the extent of the virements. However, if the total appropriation of a MDA needs to be increased, approval by the National Assembly is required in the form of a supplementary budget. Indicator PI-27 Legislative scrutiny of the annual budget law Brief Explanation and cardinal data used (i) Legislative scrutiny of the annual budget (details of expenditure and revenue) takes place within a short period of time, but MDAs Mief the National Assembly s Sectoral Committees on their estimated expenditures in advance and before the budget is submitted (C). C (ii) The legislative committees have established rules for debating the government budget. (B) (iii) Legislature s time for a meaningful debate of official government budget is very (iv) Clear rules exist for in-year amendments, but they allow unlimited and extensive re-allocation within an MDA budget (C). Legislative Scrutiny of external audit reports (PI-28) 7.80 Timeliness of examination of audit reports by the legislature (for reports received within the last three years). The scrutiny of audit reports takes a very long time in the PAC, simply because they begin the exercise 1 1 months after they receive the NAO report. In none of the last three years has the PAC issued its reports within 12 months of the NAO s reports submission date Hearings on key findings undertaken by the legislature. For the preparation of the PAC report, in depth hearings take place with a selection of responsible officers from the audited entities, usually at a pace of one or two entities per day. The hearings commence with entities in receipt of audit queries or qualified reports. However, those with clean reports are also selected at random and 68

85 asked what improvements they want to undertake in their MDA, which often reveals some weaknesses despite a clean audit report Issuance of recommended actions by the legislature and implementation by the executive. A substantial weakness within the accountability process is the lack of response from the MoFEA to the recommendations of the PAC. Until now, there is no evidence to show that recommendations are acted upon by the executive branch. Indicator PI-28 Legislative scrutiny of external audit reports Brief Explanation and cardinal data used (i) Examination of audit reports begins 11 month after their receipt and takes another 4 months to complete (D). C (ii) In depth hearings take place with responsible officers from the audited entities (B). (iii) Actions are recommended by the PAC, but they are not acted upon by the executive (C). Predictability of Direct Budget Support 0-1) DONOR PRACTICES 7.83 Annual deviation of actual budget support from the forecast provided by the donor agencies at least six weeks prior to the government submitting its budget proposals to the legislature. Provision of information on direct budget support by donors is quite timely, and in the last three years all donor agencies engaged in direct budget support have provided information several months before the submission of the government budget to legislature. There has been no deviation between announced amounts and paid amounts In-year timeliness of donor disbursement of direct budget support. All donors have agreed to provide their annual direct budget support in the first quarter o f the fiscal year and in one installment to provide a further cash facility to the government early in the fiscal year; they have done so in the last three years.20 D-1 Predictability of Direct Budget Support. A (i) In none of the last three years has direct budget support outcome fallen short of forecast (A). (ii) All direct budget support is paid in the first quarter in one installment to help boost (enhance) the government cash position in the beginning of the fiscal year (A). 2o The direct budget support for fiscal year 2008/09, exceptionally, has been released in the second fiscal quarter due to the external audit issues in the BOT. 69

86 Financial Information Provided by Donors for Budgeting and Reporting on Project and Program 7.85 Completeness and timeliness of budget estimates by donors for project support. Most donors provide estimates of their support for programs and projects, and baskets in a timely manner and in line with the government budget cycle to help budgeting and secure government counterpart funds, where applicable. This process is specifically linked to the MTEF exercise, but the estimates need annual adjustment, as the MTEF itself is regarded as a rolling plan. It is also possible that a new project, which was not known to the parties at the time of the budget preparation, may be initiated in the course of the fiscal year; similarly, a grant may be agreed, or the negotiations of a loan may be concluded, during the fiscal year Frequency and coverage of reporting by donors on actual donor flows for project support. MoFEA has noted a number of difficulties related to the financial information on projects and basket support funds. Moreover, it is difficult for MOEFA to compile and aggregate data from different projects from MDAs and LGAs. According to MoFEA officials, some donors do not provide quarterly reports within two months of the end of the quarter in which the disbursements were made, which is the main concern of this indicator. On the other hand, since most donor project and basket funds use national budget execution procedures, MDAs and LGAs are required to provide timely information on using these funds, which also faces some problems. Unlike direct budget support, project and basket funds lack a serious data reporting gap. A low budget implementation rate in donor project and basket funds activities in part indicates the lack of proper reporting. A draft World Bank policy note of September 2008 concludes that in 2006/07 (Footnote 7), the implementation rate for project and basket supported activities were as low as 62 and 51.6 percent respectively, which is according to the study largely associated with under-reporting.2' *' According to this policy note, in some instances, poor budget execution performance is largely a problem of under-reporting rather than an issue of non-expenditure. For instance, while donor-funded infrastructure projects are reported in the budget books release, allocation and spending data are not integrated in IFMS for 84 percent of the projects. This results automatically in a very low level of recorded spending for the development budget of Ministry for Infrastructure Development. This is naturally considered a major problem by both MOFEA and Ministry for Infrastructure Development, as more than TSh. 100 billion is wrongly left unaccounted for according to IFMS. This problem of reporting is directly linked to the disbursement procedures used for donor projects: the dummy exchequer system. A recent study has concluded that this system is not functioning well and that very little data is captured through the dummy exchequer. Discussions with the Ministry for Infrastructure Development suggest that this is partly correct. The dummy exchequer system appears to be a sound system from a technical perspective, but the main actors in charge of its implementation, in particular Tanzania Roads Authority and the Ministry for Infrastructure Development, don't appear to have the incentives to apply it adequately. This specific issue will be the subject of a separate report financed by Japan International Cooperation Agency on the exchequer system in the road sector. Notwithstanding the poor implementation of the dummy exchequer system, it does report on a quarterly basis on the physical and financial execution performance of these projects. However, financial execution information is limited to approximately 25% of the projects. Therefore, it is difficult to reconcile the list of projects completed with those planned for in the budget. 70

87 Indicator Donor 2. Financial information provided by donors for budgeting and reporting on project and program aid. Scores C+ Brief Explanation and cardinal data used (i) At least half of donors provide complete budget estimates for disbursement of project and program aid in line with government budget calendar, with a breakdown that could be transformed to the government budget classification, which is very broad for accommodating classification of any project or program (B). (ii) Most donors provide quarterly disbursement reports within two months of end of quarter for at least 50 percent of externally-financed project estimates in the budget (C). Proportion of aid that is managed by use of national procedures 0-3) 7.87 The use of national procedures means that the banking, authorization, procurement, accounting, audit, disbursement and reporting arrangements for donor funds are the same as those used for government funds. All direct budget support and some sector support will by definition use national procedures in all respects. Other types of donor funding (such as project and basket funds and other specific funds) use some or no elements of the national procedures In response to the government s strong request to channel support through the national systems, some donors have converted significant shares of their annual aid to budget support or basket funding for sector-specific activities. In 2006/07, such support accounted for 68 percent of the total assistance (56 percent budget support and 11 percent basket funding), compared with just 3 1 percent in 200 1/02. Consequently, reporting on donor-funded activities using national systems and procedures has significantly improved during the past few years, although there are still considerable amounts of funds that bypass government systems In 2006, a survey monitoring the implementation of the agreements made under the Paris Declaration on using government systems was undertaken. This found that aid using the country PFM systems was on average 66 percent. However, given the increased share of the direct budget support in recent years, it is estimated that at present between 75 and 90 percent of external aid funds are using national procedures, though at times with some minimal additional steps required by certain donors based on their own requirements. Table 3.5: Aid flows using national systems: 2005/06 (YO) National process Aid using budget systems Aid using financial reporting systems Aid using audit systems Aid using procurement systems Percentage of total aid 76% 61% 61% 61% managed using national procedures 71

88 8. PFM REFORM PROCESS RECENT AND ONGOING REFORMS 8.1 Tanzania has, over the last several years, initiated several reform measures in a number of PFM components, some of which are continuing. These mainly include: macro-fiscal analysis, central payment and recording system, forward looking expenditure forecasting, tax policy and administration, external auditing, procurement, and, to a certain degree, internai auditing and internal controls. In February 2008, the government adopted a new PFM reform plan called Public Financial Management Reform Programme (PFMRPIII), which calls for further reforms in the areas of policy analysis and development, external resources management, budget management, treasury management and accounting, procurement, information technology services, investment management, administrative support services, external audit services, and program leadership coordination, monitoring and evaluation. 8.2 An evaluation of the degree of success of the PFMRP I1 was conducted in 2006, and some modalities of the reform process were changed. The PFMRP I11 has now broadened its scope to include the BOT, the National Assembly, MDAs and LGAs. The new reform plan needs to be implemented by first addressing the basics and then ensuring that they are firmly in place before expanding the reforms. For example, accurate, transparent and improved cash forecasting, enhanced control over all public investments, improved credibility of MTEF, a reformed budget calendar and comprehensive budget classifications should be addressed as soon as possible. Moreover, in the plan there is little narrative to understand why certain outputs and activities have been included, and there is little sense of prioritization among outputs. The calendar to achieve these outputs (3 years) may also appear ambitious. 8.3 Additionally, and perhaps as significantly, the reform plan addresses what appears to have been an important obstacle to reform implementation over the last years: poor capacity within MoFEA and the PFMRP Secretariat to coordinate or steer the reform process. The new plan envisages a much stronger coordination structure than in the past. But coordination, leadership and overall program management may nonetheless remain a challenge for many reasons that cannot be detailed in this report. INSTITUTIONAL FACTORS SUPPORTING REFORM PLANNING AND IMPLEMENTATION 8.4 While there has been ownership of some specific reforms and piecemeal actions have been proceeding, government leadership in the coordination of PFM reforms, in particular of the PFMRPs as the coordinating vehicle for reforms, has been problematic in the past. It is believed that some basic reforms in accounting and reporting have been lagging compared with reforms in, say, macrofiscal analysis or tax reforms. The budget preparation process clearly needs to address the issues of budget classifications, timely preparation and submission of the budget, and related concerns. The dissemination of the government vision in reform strategy documents has been an area of strength, with the production of an annual progress report on PFM reforms, although there is scope for improving the analysis, and the absence of a forward looking action plan is a matter of concern. While the allocation of funds to PFM reforms is substantial, the procurement of resources by government for this purpose is rather weak, as government has been unable to spend the allocated funds. It is assumed that some of these problems relate to the ownership of the reforms. 72

89 8.5 The reform design process is generally led by the President s Office and the MoFEA. The involvement of line ministries is limited unless the reforms concern the particular sector in which case they take the lead role. The PFMW series represents an example of a coordination structure that supports a set of reforms and serves as a focal point within government for coordination of donor support for PFM reforms. The successful establishment of the coordination structure has been problematic and prolonged, and a number of challenges remain, though the institutional framework has been put in place through the establishment of a coordination unit, a steering committee and a management committee. It was particularly challenging to clarify the roles and responsibilities for implementing the reforms, though the formalization of the institutional structures can be viewed as a success. The main challenge remains the buy-in of the most relevant and directly responsible units for the proposed reforms. 8.6 In many areas, the reform process has been driven by technical assistance and pressure from development partners. This represents the development of partnerships whereby the donors and government have identified key areas for reform and donors have agreed to provide supporting assistance. However, it also means that there is a heavy reliance on technical assistance to undertake the reforms, and an ongoing challenge has been to build domestic capacity through recruiting and retaining of the personnel involved. 73

90 PART 111: PEFAR CAPACITY BUILDING PROGRAM 74

91 9. BACKGROUND AND OBJECTIVES OF THE MISSION 9.1 For nearly a year, a PEFAR Capacity Building team led by the World Bank has been assisting the Government of Tanzania (GOT) in strengthening its performance-based planning and budgeting, budget-mkukuta link, Strategic Plans (SP) and the Medium Term Expenditure Framework (MTEF). This initiative is part of the ongoing collaborative efforts of the GOT and the Development Partners (DP) to strengthen public financial management (PFM) in Tanzania. The broad capacity building program s objective is to strengthen the budget preparation process to more effectively align resource allocation to the priorities of MKUKUTA, thereby enhancing expenditure efficiency as well as growth and social impacts of the budgetary policies. 9.2 With support from the World Bank and CIDA, the core PEFAR Capacity Building team, with its GOT counterpart, carried out a preparatory mission between February 22 and March 20, 2008 (cf. Attachment 6: CB Mission1 Report). The mission identified the following priority areas for intervention during the first phase of the capacity building program: 0 To strengthen the macroeconomic model (MACMOD) to enhance its capacity to simulate growth and poverty reduction policies and to strengthen the Macro Policy Framework for the Planning and Budgeting. This is key for a reliable budget frame that constitutes a critical input for preparation of the Budget Guidelines; 0 0 To strengthen the MTEF in order to facilitate and optimize the budgetary resource allocation system, while taking into account strategic priority choices as well as basic sector needs; To strengthening sector Strategic Plans (SP) and sector MTEF with a programmatic approach to enhance the readability and strategic focus of the MTEF as well as ensuring consistency between SP/MKUKUTA/MTEF and results/outcome. This is also important in order to facilitate dialogue between the MoFEA and the sector ministries in the interactive inter-sectoral resource allocation process. 9.3 This second mission took place in Tanzania from September 4 to December 17,2008 to kickstart the capacity building work program. The mission was supported by the Swiss Development Cooperation (SDC) and the World Bank. The specific objectives were to assist the GOT team in: 0 Upgrading the macro model (MACMOD) with the construction of an endogenous growth simulation module; and preparing inputs to the Macro framework for the 2009/10-11/12 Plan and Budget Guidelines (PBG); 0 0 Preparing a budget background paper that explains the macro framework and the strategic and sector-specific challenges that determine the medium-term budget resource allocations (MTEF 2009/10-11/12) and the 2009 annual budget; Proposing upgrades to the MTEF that could improve the inter-sectoral resource allocation process with the Strategic Budget Allocation System (SBAS). 75

92 9.4 The mission was led by Charles N cho-oguie (lead consultant), and included Fulbert Tchana (consultant), Leonidas Luteganya (consultant), Paolo Zacchia, Florence Charlier, Emmanuel Mungunasi (all from the World Bank), and Rose Aiko (SDC). This report provides a summary of the main accomplishments of the mission and outlines a road map for the capacity building work program for the remainder of

93 10. MAIN ACCOMPLISHMENTS OF THE MISSION How this work was conducted STRENGTHENING THE MACROECONOMIC MODEL - MACMOD 10.1 Roughly two-thirds of the mission was devoted to modeling work to revise and upgrade the MACMOD as well as to prepare the macro framework paper for the 2009/10 Plan and Budget Guidelines. Through a series of technical meetings, workshops, and on-site training, we upgraded the model and strengthened both human resources and capacities. These tasks were undertaken by a core macro group organized and led by the Policy Analysis Department (PAD) of the MoFEA with some facilitation from the DP PEFAR Capacity Building team. The core macro modeling group included key resource persons from the President s Office - Planning Commission, the Bank of Tanzania (BOT), the National Bureau of Statistics (NBS), the Tanzania Revenue Authority (TRA), and the regular staff in the PAD of the MoFEA The capacity building work was organized in three stages (Attachment 1): First, a series of regular technical meetings and on-site training sessions were held on a daily basis in the Conference Room of the PAD (September 15-27) (cf. Work Plan, Annex 1). This series of technical meetings and training were restricted to the core macro modeling group, which is composed of 10 members. The group reviewed MACMOD and discussed and implemented the first set of upgrades, notably the Endogenous growth module. Second, a two-week workshop was held at the POTREA COURTYARD Hotel (October 6-16) for the larger Macro Modeling Group (-25 members). The workshop was used to appropriate, review the data, and organize the policy simulations work for the input to the macro framework paper. The workshop was extended to the entire Macro Group, including more participants from the Central Bank and the University of Dar es Salaam. Third, regular meetings/training sessions were organized in the Conference room of the PAD in the MoFEA to prepare simulations for the macro framework while continuing to strengthen the model and enhance its appropriation. Accomplishment 10.3 As a result of the series of technical workshops described above, the team accomplished the following tasks: A complete review of the original version of MACMOD. This was the first step in the capacity strengthening process and was meant to fully explain the original MACMOD to the core group so as to enhance its appropriation and also to initiate future upgrades. These sessions involved: (i) revisiting the theoretical underpinnings of the MACMOD (the IS- LM-BOP/AD-AS framework) and briefly retracing the history of the development of MACMOD; (ii) reviewing the main equations and closure of the IS-LM-BOP model; (iii) analyzing the specifications, econometric estimates, and key parameters of the model; (iv) reviewing the database and the computer implementation of the model on Excel (the INPUT, 77

94 CORE MODEL and Output files). This review provided the core group with a sufficient understanding of MACMOD and how to perform simulation exercises with the model. Discussion of model upgrades. The objectives were to revisit the theoretical underpinnings and the model specification in the context,of current needs of the GOT in terms of macro policy analysis and MKUKUTAMTEF-based planning. A further goal was to provide rationale for proposed model upgrades. These series of technical meetings were used to: (i) assess the key limitations of MACMOD and to argue for a disaggregated supply approach in order to enhance growth and poverty policy simulation capabilities of the model; (ii) specify the key equations of the aggregate supply block (an endogenous growth module); (iii) reconcile the supply-driven growth module with the rest of the IS-LM-BOP/AD-AS framework, including required changes in model closures. Econometric analysis and estimation of model parameters. This session focused on how key model parameters are estimated, including: (i) preliminary econometric analysis of key model equations by Econometric-Views using both panel data and Tanzanian economic series; (ii) estimation of other key model parameters by calibration methods; (iii) preliminary draft of a note on the econometric analysis of the model. Programming MACMOD-TZ on Excel. This included reorganizing and programming the key components of the model, namely the Input module (MACMOD-TZ-IN), the new core module (MACMODTZ-MODEL), and the output module (MACMOD-TZ-OUT); (ii) programming a navigation tool (the SART module) to facilitate data input as well as policy simulations and results analysis. The Macro includes an overall control module (START) that allows users to navigate from one module to another, as well as navigation tools within each module. Policy simulation and analysis. This set of activities was designed to facilitate appropriation and the exploitation of the model by the core group. The group was trained to: (i) design alternative scenarios (consistent $et of exogenous as well as policy variables); (ii) perform simulations with the model and analyze the results, which subsequently constituted inputs to the macro policy framework paper. Achievements 10.4 As a result of these activities, the core macro modeling group was strengthened in two complementary ways: A vastly improved and user-friendly macro simulation tool (see Attachment 2). The team is now equipped with a model that has enhanced capacity to simulate growth policies, including budgetary policies, sector strategies, investment, as well as exogenous shocks. The model is also organized to facilitate scenario analysis and risk assessment. Enhanced human capacities for macro policy simulation and analysis. The series of workshops and the sustained involvement of the core group have contributed to strengthening capacity and bringing the macro group together for substantive technical work leading to the preparation of the macro policy framework paper. As a reflection of this, the core group has been able to organize and simulate alternative scenarios for the macro framework 2009/10-11/12, and is now in the process of revising the baseline scenario to take into account new information about the world economic outlook as well as new data from the economic survey (see. Section 2.2). 78

95 PREPARING THE MACRO FRAMEWORK 2009/10-11/12 FOR THE PLAN BUDGET GUIDELINES How this work was conducted 10.5 At the request of the Commissioner for Policy Analysis, the work program was slightly modified to cater to immediate needs related to the preparation of the Macro Policy Framework Paper (MPFP) for the 2009/10 Plan Budget Guidelines (PBG). The modeling work was reorganized in order to provide inputs to the PBG mid-way through the mission (first input by the end of October 2008). To this end, the workshops sessions on econometric analysis were shortened to accelerate the programming of the model and its exploitation for policy simulations. 0 The two-week workshop with the larger Macro Group provided a venue for organizing and intensifying the policy simulation work for the MPFP and the PBG. During the workshop, the team was divided into specialized groups (Production; Government, External Sector, and Money). The sub-groups took charge of their respective sub-modules (GDP, Gov-Fin, Trade/BOP, and Money). They prepared the database for each sub-module, updating estimates for 2008; and they prepared input to the model, including specifying the main exogenous and policy assumptions for in their respective areas. Then, a core group was designated to prepare draft zero of the MPFP, including an analysis of the most recent developments and medium-term outlook and targets, as well as simulated macro and budget frames (GDP growth and public revenue and expenditure profile). Thereafter, in-house technical meetings were held to revise and upgrade the draft, including performing scenario analyzes. Accomplishment 10.6 The following tasks were executed by the Group in order to prepare the first draft of the MPFP for the 2009/10 PBG: Review of recent economic developments: This included: (i) a survey of real sector development for 2007 and 2008 and the analysis of recent (2008) macroeconomic developments (Chapter 1 of the draft PFP); (ii) assessment of the world economic outlook and the implications for Tanzania. 0 Design and simulation of a baseline scenario : This includes: (ii) Specifying key macroeconomic and fiscal policy targets (growth, inflation, exchange rates, revenue and primary deficit targets), and making well-informed conjecture on exogenous world developments (world markets trends for Tanzania s export, FDI and ODA); (ii) simulating a baseline scenario using MACMOD-TZ; (iii) performing risk analysis to develop alternative growth scenarios based upon the likely magnitude and impact of the world economic recession, persistent energy problems, and government fiscal strategies. Draft the MPFP: The team prepared the first draft of the MPFP, including chapters on recent economic developments (Chapter l), outlook for (Chapter 2), and analysis of the baseline scenario and possible alternative scenarios, as the world recession was proving more severe than originally anticipated (Chapter 3). Achievement 10.7 The above activities, combined with the modeling exercise, yielded the following outputs: A preliminary draft of the MPFP for input to the 2009/10 PBG (see Attachment 3). The draft is still being strengthened as new information regarding the extent and noticeable impact 79

96 of the crisis becomes available, and also on the latest estimates of actual macroeconomic performance in 2008 (GDP growth and fiscal performance). Strengthened human capacities. The macro team has grown considerably in its capacity to generate a consistent set of macro framework tables (Real Sector, Gov, Trade and BOP, Money) and to simulate policies and undertake scenario analysis in a timely fashion. PREPARING THE BUDGET BACKGROUND AND MEDIUM TERM FRAMEWORK 10.8 More than one third of the mission was devoted to assisting the MoFEA Budget Department in preparing the first edition of a Budget Background and Medium Term Framework Paper (BBMTF). The paper provides a coherent review of the macroeconomic background as well as the strategic focus of the budget; it also analyzes key trends in budget executions and explains how the medium-term expenditure allocations are aligned to the strategic priorities of MKUKUTA and related sector needs. This document is at the cornerstone of the GOT S effort to make the budget more readable and transparent to stakeholders, as well as align the MTEF and the budget on the priorities of MKUKUTA and sector Strategic Plans. The GOT intends to produce the BBMTF every year to accompany and explain the traditional budget books. How this work was conducted 10.9 The preparatory work for the BBMTF was led by the Commissioner for the Budget. A task team was designated, which mainly consists of technical staff from the Budget Analysis Department and was supported by the DP PEFAR capacity building team. Regular in-house technical meetings. As in the macro modeling exercise, most of the preparatory work that led to the first and full draft of the document was undertaken in the form of regular technical meetings within MoFEA s facilities. These meetings were organized to discuss the framework and content of the document, provide practical training on how to extract budget data, and prepare key tables for analyzing budget execution and allocations trends using pivot tables in Microsoft Excel. Technical workshops. Several technical workshops were organized outside the MOFEA in order to draft the document or to reviewhpgrade the existing draft in an iterative fashion until a complete and satisfactory draft was obtained. Several of these meetings were chaired by the Commissioner for the Budget. Internal review within the GOT. The full draft of the BBMTF was circulated to the rest of the senior staff and management of the MOFEA as well as other central government institutions for an internal review process organized by the Commissioner for the Budget. Cross-support and full collaboration with the DP s Rapid Budget Analysis Group. The preparation of the BBMTF has benefited considerably from a close collaboration with the PEFAR Rapid Budget Analysis work that was carried out by the World Bank-led DP task team. The collaborative work includes exchange of data, discussion of methodology, and exploitation of the results of the PEFAR analytical work on budget allocation and execution. Accomplishment The Task team undertook the following activities in order to prepare the full draft of the BBMTF: 80

97 Training for budget data manipulation and budget trends analysis. This task involved strengthening team member s capacities to (i) manipulate budget data (extraction, pivoting and tabulation techniques), (ii) prepare key tables and analyze budget execution and budget allocations by clusters, sectors and economic classification, as well as allocation and execution within each sector. Drafting the report. This task included (i) preparing the outline of the document, (ii) organizing the team by sub-groups to take charge of drafting the various chapters and sections. Reviewing, discussing, and strengthening the draft. This was done in an interactive and iterative fashion via the series of meetings and technical workshops mentioned above. Validating the draft. At the conclusion of the work, the draft document was submitted to the Commissioner for Budget for validation. The task team received many comments from the internal review process and revised the final document accordingly. Achievements The coordinated activities above yielded important tangible results, including: A strong analytical budget policy framework document (see Attachment 4). The BBMTF ties together the macroeconomic framework and the challenges in the execution of the budget, especially development budget, to realize MKUKUTA objectives and MDG goals as well as the medium-term expenditure trends. The GOT is determined to institutionalize the document and to prepare it as a cover document to the traditional budget books. An organized database for annual budget trends analysis, including key tables of expenditure trends by clusters, by sectors,, economic classification, as well as budget allocations and expenditures tables within each sector. A stronger budget analysis department. In the process of accomplishing the above tasks, the budget team considerably enhanced its capacity to carry-out in-house analytical work, including an expenditure review and MTEF allocation analysis. How this work was conducted STRENGTHENING THE MTEF While the previous components of the work program were undertaken primarily by the respective task teams organized by the GOT, the exploratory work to upgrade the MTEF and SBAS was conducted mainly by the core PEFAR capacity building team. This is partly because the GOT task teams were already stretched thin and consumed by other ongoing assignments, namely the macro modeling exercise, the preparation of the MPFP for the PBG, and the production of the BBMTF. But it is also by design, that, given the complexity of the MTEF system, the mission has decided to adopt a more cautious approach for strengthening it. The strategy is to first develop key upgrades on an experimental basis, to test their pertinence and usefulness and also assess the scope of the programming work needed to upgrade the SBAS, both Macro and Micro versions. It is expected that with proven results from the experimental exercise, the capacity building work will be enticed to embark on a full-scale revision of the current MTEF and its computer system, the SBAS. Accomplishment 0 First, a simple Excel-based MTEF optimization model was designed by the team (Attachment 4). This simple spreadsheet model serves as a transparent tool for implementing 81

98 the 3-step strategic allocation process, distinguishing a baseline budget covering basic necessities, a pipeline budget covering ongoing investment programs, and a marginal budget allocated to new projects on the basis of performance, MKUKUTA priorities and the available budget space. A new SBAS Macro module is developed. This work included: (i) discussion of the 3-steps allocation approach and identification, by government functions/sector/votes, of baseline expenditure, pipeline spending, and new allocations, based on priorities, merits, and the available budget space; (ii) programming of the 3-step approach in SBAS Macro; (iii) allocations analysis with the enhanced MTEF model. Achievements The experimental exercise proved largely successful and provides a stepping stone for a full scale upgrade of the MTEF: e e The three-step approach was tested on the 2008 and 2009 budgets andprovides an effective and more transparent means of streamlining the resource allocation process in SBAS-Macro. It also provides an effective tool for projecting the MTEF over the outer years (n+l, n+2). This is because the baseline budget can be projected with simple and explicit assumptions on the progression of the wage bill and the cost of basic expenditure on goods and services. Likewise, the pipeline budget provides an effective mechanism for keeping track of multi-year investment programs and consistently supporting them over the planning period. The Revised SBAS-Macro is operational (Attachment 5). It provides a more effective tool for optimizing resource allocations on the basis of budget space remaining once basic necessities and pipeline program costs have been covered. BUILDING A PROGRAMMATIC DIMENSION IN THE SP AND THE MTEF An experimental framework was developed for the health sector during the preceding mission (Attachment 6). The remaining tasks were (i) to revise the suggested program-based framework and reconcile it with the new priority action plan developed by the sector as part of the revised sector strategy, and (ii) to revise the sector MTEF accordingly However, the Commissioner for Budget requested that this component of mission activities be moved to the next phase of capacity building work program. This is partly because the budget team was already consumed with other tasks, and also because of considerable delay in the implementation of the functional budget classification system, which the GOT regards as a prerequisite to the program-budgeting system. Therefore, and at the request of the PAD, the time and resources freed from the health sector program-budgeting work were reallocated to intensify the modeling work and prepare rapid inputs to the MPFP and the 2009/10 PBG. 82

99 11. ACTION PLAN FOR THE REMAINDER OF THE YEAR (2009) A tentative action plan was discussed with the PAD with regard to the macro modeling work and the Budget Department for the BBMTF and the MTEF work. The following were identified and agreed as key areas of immediate attention for the remainder of STRENGTHENING MACMOD-TZ AND THE MACRO POLICY FRAMEWORK PAPER 2009/10-11/12 Revising the baseline macroeconomic framework for the PBG 11.2 The world economic outlook has continued to worsen since the first draft of the macro framework was prepared last fall. There are signs that the world financial crisis and the ensuing recession may have impacted Tanzania s economy much more deeply than anticipated in terms of FDI and ODA flows as well as tourism and exports. The sensitivity analysis performed last fall indicated that under such a scenario, GDP growth may be reduced from 7 percent to 6.5 percent or even lower in There are also indications that the fiscal revenue performance may be weaker than planned, partly the reflection of a slow-down in economic activities In light of these developments, the core macro modeling group has resumed work to revise the exogenous assumptions, redo the simulations and revise the baseline scenario for the budget frame. This work is being supported by the PEFAR capacity building team. The revised framework was prepared in April 2009, based on the new data on production sectors that was released by the National Bureau of Statistics in early April. Revising the econometric analysis of MACMOD-TZ 11.4 As of now, the main equations and model parameters have been estimated using a panel approach with data for SSA countries, and on the basis of Tanzania s old economic series. During the next mission, a workshop will be organized with the core macro team to redo the econometric analysis on the basis of the new SNA-93 economic series. Developing a simple poverty and social development indicators projection module 11.5 A simple poverty module will be developed to enable the model to project key poverty and social indicators, including the income poverty incidence as well as education and health sector MDGs. Econometric techniques will be used to predict such indicators on the basis of growth pattern (per capita growth rate and distribution of growth), public spending in education, health, and infrastructure. This technique is similar to the approach used in the MAMS model (Marquette for MDGs Simulation). This task could possibly be accomplished during the next capacity building mission. Developing a labor market and full poverty and MDG modules During the second half of the year (July-December), the modeling work will concentrate on strengthening the capacities of MACMOD-TZ to perform Poverty and Social Impacts Analysis (PSIA). To this end, a labor market module will be developed that will allow the model to project 83

100 trends in employment and wages consistent with sector growth. A framework has already been developed by the core macro team, which will provide a starting point for the labor market module Also, expanding from the previous work, a full Poverty/MDG module will be developed to project poverty incidence and all key MDGs. This will ensure that the model is well equipped to assist in poverty simulations for ongoing work for preparing the new edition of MKUKUTA. Intensifying the training component of the capacity building work program 11.8 During this first phase, priority has been given to achieving models, documents, and other specific outputs. As a consequence, the training component has not received adequate attention. In all subsequent missions, emphasis will be placed on training the larger macro team in order to facilitate appropriation and exploitation of the model. All aspects will be covered, especially the econometric analysis as well as the modeling of growth and poverty reduction policies. At the central level (SBAS-Macro) MTEF UPGRADES AND SECTOR PROGRAM BUDGETING 11.9 The first step is to discuss the proposed upgrades of the MTEF allocation system as well as the revised SBAS project with the Commissioner for Budget. The second is to agree on a list of baseline expenditure items as well as the nature of the typical pipeline programs. The Priority Investment Program (PIP), which is being developed by the GOT, will form the major input to this exercise. Then, a full-scale revision of SBAS, both macro and micro, could be implemented as a means for a consistent 3-year MTEF allocations system. Sector Strategic Plan and Program Budget During the second half of the year (July-December), the capacity building team can resume work to build a program dimension in sector strategic plan and MTEF, which would also help sectors MDAs to perform scenario analysis based on projected budget ceilings SBAS-Micro will be revised accordingly, allowing for program codes as well as codes for identifying expenditure by degree of contingency (basic necessity, pipeline, and new programs). The new approach will be fully implemented for designated priority sectors. Budget Background and Medium Term Framework The Commissioner for Budget is planning to resume work on preparing the BBMTF-2009/10-11/12 in the second half of the year (July-September). In order to assist the capacity building work, the PEFAR team will prepare an Excel program for extracting budget data and preparing key tables for rapid expenditure analysis for the BBMTF The capacity building work will put emphasis on the training component in order to strengthen capacity for analysis of budget execution and strategic budget alignment. This will help the task team as it prepares the 2009/10-11/12 edition of the BBMTF. 84

101 ANNEX 1: PERFORMANCE INDICATORS SUMMARY TABLE hdicator?i-1 Aggregate expenditure outcome :ompared with original approved budget Score B Brief Explanation and cardinal data used In last three years, only in 2007/08 did the actual total expenditure deviate from budgeted expenditure by more than 10 percent. PI-2 Composition of expenditure outcome :ompared with original approved budget PI-3 Aggregate revenue outcome :ompared with original approved budget PI-4 Stock and Monitoring of Expenditure Payment Arrears PI-5 Classification of the budget PI-6 Comprehensiveness of information included in budget documentation C A C+ C B In 2005/06, variance in expenditure composition exceeded overall deviation in primary expenditure by more 10 percentage points. In the last three years, revenue collection was consistently above the approved budget. (i) The stock of expenditure payment arrears constitutes 2-10 % of total expenditure, and there is no evidence that they have been reduced significantly in the last two years (C). (ii) Data on the stock of arrears is consolidated annually, but they do not have an age profile (B). The budget documentation and execution is based on GFS-based administrative and economic classifications. No GFS-based hnctional classification is part of the budget documentation and reporting system. the required information. PI-7 Extent of unreported government operations PI-8 Transparency of inter-governmental fiscal relations PI- 9 Oversight of aggregate fiscal risk from other public sector entities PI-10. Public access to key fiscal information c+ B B (i) Extra budgetary spending is estimated to be between 5-10 percent of total government expenditure (C). (ii) Complete income/expenditure information is included in the fiscal reports for all loan-financed operations and at least 50 percent of grant-financed operations (B). (i) The horizontal allocation of almost all transfers is guided by certain measures, but because they are conditional transfers, sectoral policies also need to be taken into account (B). (ii) Reliable information to LGAs is issued three months before the start of the fiscal year, which is a short period for significant budget changes to be made (C). (iii) Fiscal information that is consistent with the central government fiscal reporting is collected on quarterly basis (B). (i) Most AGAsPEs submits fiscal reports to the central government, but these are normally delayed and their consolidated overview is incomplete (C). (ii) The LGAs do not have borrowing power, unless approved by the MoFEA, which is not used, and their budget execution reports are available on quarterly basis (A). The government makes available to the public 5 out of from 6 types of information, but two of them are not complete.

102 ) 1 ndicator '1-11 Orderliness and participation in the innual budget formulation process Brief Explanation and cardinal data used (i) A comprehensive budget calendar exists, but always ends up with late submission of the budget just before the end of fiscal year, though MDAs need more time to submit their budgets (C). B (ii) A comprehensive budget circular and budget preparation guidelines are issued, and the center of government has time to make necessary adjustments (B). (iii) The Legislature normally approves the budget before the start of the fiscal year, but all authority for changes within the approved total level of an MDA's budget remain within the power of the executive branch (B). PI-12 Multi-year perspective in fiscal Ilanning, expenditure policy and Iudgeting (i) Forecast of fiscal aggregates, on the basis of main categories of administrative and economic classification for recurrent expenditures and project or sector-based for development expenditures, are prepared for two years in addition to budget year (C). (ii) DSA for external and domestic debt has been undertaken in at least two out of the last three years (B). (iii) Sector strategies exist, but they are inconsistent with aggregate fiscal forecast (C). (iv) Linkages between investment budgets and sector strategies and recurrent budgets are weak (D). PI-13 Transparency of Taxpayer 3bligations and Liabilities (i) There are relatively new and comprehensive income tax and VAT laws (B). PI-14 Effectiveness of measures for taxpayer registration and tax assessment PI-15 Effectiveness in collection of tax payments C+ B (ii) Taxpayer access to information on tax liabilities and administrative procedures. There are good taxpayer education systems but no call center or central information system (B). (iii) There is an independent disputes resolution system funded separately by government, though it needs further capacity enhancement (B). (i) Taxpayers are registered in a database with an identification system but there are no links to other systems (C). (ii) Penalties exist, but they are insufficiently specific and in some cases too low to have an impact on compliance (C). (iii) There is a good tax audit operation in the LTD (the main tax collection department), but a mixed performance in DRD (B). (i) Collection ratio for gross tax arrears in 2004/05 and 2005/06 was 33 percent according to LTD data and 7 1 percent according to LTD and DRD data (C). (ii) Revenue collected is transferred to the main account in the Bank of Tanzania within a week (B). 86

103 [ndicator Brief Explanation and cardinal data used (iii) Reconciliation between tax assessments, collections, arrears records and receipts by the Treasury takes place as a matter of routine (A). PI-16 Predictability in the Availability of Funds for Commitment of Expenditures (i) An annual cash flow forecast for government budget and each MDA is prepared, but substantially revised on a monthly basis (B). C+ (ii) MDAs are provided with reliable information for their commitments but only for each month and with relatively short notice (C). (iii) Significant in-year budget adjustments are frequent, but undertaken with some transparency(c). PI-17 Recording and management of each cash balance, debt, and debt guarantees (i) Debt data is of a relatively high quality, but minor reconciliation problems occur, and timely statistical reports are produced (B). C (ii) The balances of several government bank accounts in commercial banks are not consolidated, though there is a plan to do so in the future (D). (iii) Contracting of loans and issuing guarantees is approved by the Minister of MoFEA in line with rules, but it is not clear if ceilings apply (C). P1-18 Effectiveness of payroll controls (i) Personnel records and payroll data are stored in the same database and are fully integrated and reconciled (A). (ii) Changes to personnel records and payroll data take place on a monthly basis, but quality is not assured (C). C+ (iii) Some internal controls exist for data entry for changes on personnel records and payroll data, but are not adequate to ensure the integrity of data, which originates in the MDAs (C). (iv) Quarterly payroll audits with limited coverage have been undertaken during the last three years, but no survey has been conducted since 1994 (C). PI-19 Competition, value for money and controls in procurement (i) 66 percent of tenders under open tendering process were advertised in fiscal year 2006/2007 (B). B (ii) Using less competitive procurement is allowed with justification. Although PPRA s view may be sought, observing the procurement act remains within the power of procuring agencies (B). (iii) A comprehensive complaints mechanism operates, but for unknown reasons the number of complaints has declined (B). PI-20 Effectiveness of internal controls for non-salary expenditure (i) The use of IFMS with a commitment control feature is relatively effective, but cannot provide a full guarantee at all times (B). 87

104 [ndicator Brief Explanation and cardinal data used C+ (ii) Other internal rules and procedures cover major sets of controls that are generally understood, but in some areas, these are excessive and compliance rate vary (C). (iii) Rules are observed in majority of transactions, but there is some evidence of misconduct (C). PI-21 Effectiveness of internal audit (i) Internal audit function exists in most MDAs, and it is estimated that 20 percent of the staff time is allocated to system-based reviews and high risk areas (C). D+ (ii) Reports are issued for most MDAs, but these are not submitted either to the AGD or the NAO (C). (iii) To some degree, actions are taken by management on major issues, but often with delay (D). PI-22 Timeliness and Regularity of Accounts Reconciliation (i) Bank reconciliations for Treasury-managed accounts are undertaken monthly within 4 weeks, but the size of unreconciled items seems to be large (C). (ii) Suspense accounts are reconciled and advances are cleared in monthly periods, but the size of unreconcileduncleared items is not known. A significant number of accounts have uncleared balances brought forward. (C). PI-23 Availability of Information on Resources Received by Service Delivery Units PI-24 Quality and timeliness of in-year budget reports. Information on resources received by front line service delivery units is mostly lacking. Special surveys were undertaken within the last three years, but their results and methodologies used have not been published. (i) Scope of flash reports is compatible with the budget estimates for both commitments and payments, but in very broad categories. Revenue is also classified very broadly (B). (ii) Data are provided monthly on flash reporting for expenditures and quarterly for all fiscal components, including revenue, expenditure deficit, and financing (A). (iii) Information is accurate in flash reports, but unknown in overall quarterly fiscal reporting due to multiple sources of data (C). PI-25 Quality and Timeliness of Annual Financial Statements (i) Central government final accounts include revenue, expenditure and bank balances, and since 2007/08 some data on liabilities but not on assets (C). (ii) Financial statements are submitted for external audit within 6 months of the end of the fiscal year (A). (iii) IPSAS cash accounting standards are used, but still assets and liabilities are not fully reported (C). 88

105 [ndicator PI-26 Scope, nature and follow-up of :xternal audit Brief Explanation and cardinal data used (i) The audit report covers all MDAs and LGAs, but part of the AGAs and PES. The nature of the audit remains centered on transaction level testing, but reports identify some significant issues C+ (ii) In the last three years, the audit report, including consolidated financial statements of government, was presented to the legislature within 8-9 months after end of period covered (B). (iii) MDAs and MFOEA submitted formal responses to the NAO on only the audit report of the 2005/06. These are due to be discussed by the PAC in 2009 after a long time gap. No evidence can be demonstrated for a systematic follow up on audit findings by the MDAs (C). PI-27 Legislative scrutiny of the annual wdget law (i) Legislative scrutiny of the annual budget (details of expenditure and revenue) takes place within a short period of time, but MDAs brief the National Assembly s Sectoral Committees on their estimated expenditures in advance and before the budget is submitted (C). C (ii) The legislative committees have established rules for debating the government budget (B). (iii) Legislature s time for a meaningful debate of official government budget is very limited (D). (iv) Clear rules exist for in-year amendments, but they allow unlimited and extensive re-allocation within the budget of an MDA 0. PI-28 Legislative scrutiny of external audit reports (i) Examination of audit reports begins after 1 1 month of their receipt, and takes another 4 months to complete (D). Donor 1. Predictability of Direct Budget support C A (ii) In depth hearings take place with responsible officers from the audited entities (B). (iii) Actions are recommended by the PAC, but they are not acted upon by the executive (C). (i) In none of the last three years has direct budget support outcome fallen short of forecast (A). (ii) All direct budget support is paid in the first quarter in one installment to help boost (enhance) the government cash position in the beginning of the fiscal year (A). lonor 2. Financial information provided iy donors for budgeting and reporting on iroject and program aid C+ (i) At least half of donors provide complete budget estimates for disbursement of project and program aid in line with government budget calendar, with a breakdown that could be transformed to the government budget classification, which is very broad for accommodating classification of any project or program (B). 89

106 Indicator Score Brief Explanation and cardinal data used Donor 3. Proportion of aid that is managed using national procedures B (ii) Most donors provide quarterly disbursement reports within two months of end of quarter for at least 50 percent of extemallyfinanced project estimates in the budget (C). 90

107 ANNEX 2 : ASSESSMENT OF PEFAR 08 CAPACITY BUILDING WORK Macro Framework Paper (Views from the Macro Modeling Group and Budget Brief Team) 1. The Macroeconomic framework paper is prepared every year as one of the necessary background papers for the Plan and Budget Guidelines (PBG). The Macro framework paper for 2009/ /12 is one of the products of the PEFAR 0 8 capacity building, which also contributed to and informed the Budget Guidelines Committee (BGC) on the available resources both from domestic and foreign sources. Based on the projected GDP level, the resource envelope is determined using macro modeling. During the preparation of the 2009/10 PBGs, the upgraded MACMOD-TZ was used in projecting the resource envelope based on the projected levels of real GDP growth. Using the MACMO-TZ, the government was able to project robust levels of real GDP growth, which in turn was used to project more reliable levels of domestic revenue, aid and expenditure. In the 2009/10 PBG, projected real GDP for 2009/10 is 7.0 percent and the revenue to GDP ratio is 16.8 percent, which is more reliable than the 17.7 percent projected in the previous year. The increased reliability of these key macroeconomic variables signifies the improvement achieved with the upgrade of the MACMOD-TZ and subsequent training of a dedicated team coordinated by the Ministry of Finance and Economic Affairs (MoFEA). Budget Brief Paper 2. Despite the government efforts to improve the budget preparation process, some challenges have also emerged, including the increased need and demand for more communication with stakeholders on the background of the budget. The stakeholders have continued to demand more information on the guiding principles, macroeconomic underpinning, strategic objectives, and medium-term targets of the budget in a simple, integrated, and compelling manner. The government has risen to this challenge in a number of ways. First, the Budget Guidelines have been streamlined and enhanced to more clearly reflect strategic objectives and choices, and the implications for key sectors in terms of medium-term resource allocation. Second, a summary budget framework document is published (Budget Digest) and submitted to the parliament accompanying other budget documents on budget speech day. This note explains in a clear and concise fashion the main macroeconomic assumptions and fiscal underpinnings of the annual budget and the medium-term expenditure framework. 3. The Budget Brief Paper (titled Budget Background and Medium Term Framework, BBMTF) constitutes a milestone in the government s sustained effort to address stakeholders demand for enhanced.communication and transparency of the budgeting process. The document is the first edition of an annual report on the budget and the MTEF, designed to serve as a narrative cover for the traditional budget books. The BBMTF is an integrated document that knits together the reviews of the macroeconomic framework, the achievements and strategic challenges in the implementation of MKUKUTA, and the analysis of budget execution, expenditure trends, and medium-term focus with the strategic priorities of MKUKUTA as a backdrop. The BBMTF was prepared by the Budget Department of the MoFEA. This work has received technical support from the Development Partners in the framework of Capacity Building. 4. The BBMTF highlights critical issues in the budget preparation and implementationlexecution, including allowances, better balance between MKUKUTA clusters, low maintenance budget in social sectors, slow pace in D by D policy, under-execution of the 91

108 development budget, need for prioritized medium-term public investment planning, and many other issues. Many of the issues identified in the BBMTF have received recognition of the BGC and have also been included in the PBGs in order to be addressed in the 2009/10 budget. In addition, BBMTF presents analysis of the medium-term expenditure by strategic objectives, by MKUKUTA clusters, major sectors, economic nature, and key sectors. The paper also provides an analysis of expenditure by current and capital components. This information has significantly enhanced the required levels of clarity and transparency in the budget and gives an indication of budget consistency with policy objectives. 5. In the future, the BBMTF could serve as a unique and comprehensive background document for different review processes such as PEFAR, budget review, and sector reviews. This will help streamline the large number of sometimes repetitive background papers used for these processes. In this regard, the government intends to use the BBMTF to better inform and considerably facilitate the dialogue with domestic stakeholders as well as its Development Partners, thus reducing the burden of related preparatory work on government staffs. The government intends to publish the BBMTF for wider circulation to stakeholders, including CSOs, academia and general public. And a streamlined calendar for its dissemination will be established for future publications. MACMOD-TZ 6. It should be noted that in 2007, the National Bureau of Statistics revised the National Accounts of Tanzania Mainland with the goal of enhancing the quality of National Accounts (NA) estimates to better portray the economic activities in the country and ensure international comparability in accordance with the United Nations System of National Accounts 1993 (SNA 1993). The revision is based on the 2001 results of the Household Budget Survey, and thus the benchmark for the newhevised GDP series is at constant 2001 prices as opposed to the former outdated 1992 prices. 7. Given the revision of the NA data, it was imperative to update the MACMOD with the new data series, which starts in The task was undertaken by a core macro group (CMG) organized and led by the Policy Analysis Department (PAD), Ministry of Finance and Economic Affairs and included key resource persons from President s Office, Planning Commission, the Bank of Tanzania (BOT), the National Bureau of Statistics (NBS), and the Tanzania Revenue Authority (TRA). 8. The main tasks undertaken by the Team under the guidance of Professor Ncho were to review the existing version of the MACMOD and update the database with the new national income series and any other new data available. The upgraded model was used to assist in the preparation of the first draft of the macro policy framework for the plan and budget for the period 2009/10-1 1/12. Other tasks undertaken so far include developing a navigation tool to make the new framework userfriendly and facilitating the utilization of the model for policy analysis and planning. 9. Apart from producing the Macro framework paper, the upgraded MACMOD-TZ has been a useful tool for simulating different scenarios of the impact of the current economic crisis. This is necessary input into the dialogue and discussions on the crisis. The Commission, which was formed by the President (and chaired by the Governor of the Bank of Tanzania) to oversee/evaluate the impact of the crisis, will benefit from the inputs produced by the MACMOD-TZ team. The MACMOD team comprised about ten energetic staff from different relevant institutions (as indicated above), is working hard to simulate different scenarios with regard to the crisis as well as updating the Macro framework paper and the budget framework as new information comes on board. The modeling process is always a work in progress, and therefore there is some work still pending, including further update of the model in the work of Global Financial Crisis; revising the model 92