Tanzania - Public Expenditure Review (Vol. 1 of 2) : Main report

Report No. 16578-TA Tanzania Public Expenditure Review (In Two Volumes) Volume 1: Main Report August 11, 1997 Resident Mission in Tanzania Country Department 4: Tanzania and Uganda Africa Region Document of the World Bank GOVERNMENT FISCAL YEAR July 1 - June 30 CURRENCY EQUIVALENTS Currency Unit: Tanzania Shilling (T Sh) Official Rate: US$1.00 = T Sh 581 (1996 Ave.) ACRONYMS AND ABBREVIATIONS ADB African Development Bank AMAP Aid Management and Accountability Project OCAG Office of Controller and Auditor General BOT Bank of Tanzania ODA Official Development Assistance C&CL Contractual and Contingent Liabilities O&M Operation and Maintenance CFS Consolidated Fund Services PC Planning Commission CG Consultative Group PCIS Personnel Control and Information System CPS Central Payments System PE Personal Emoluments CS-DRMS Commonwealth Secretariat Data Management System PER Public Expenditure Review CSRP Civil Service Reform Program PIP Public Investment Program CSRS Civil Service Reform Secretariat PMG Paymaster General DAWASA Dar es Salaam Water and Sanitation Authority PMO Office of the Prime Minister EAP Executive Agencies Project PMIU Payroll Management Information Unit EU European Union PSRC Parastatal Sector Reform Commission ERP Economic Recovery Program RASs Regional Administration Secretariat GADP Government Accounting Development Project RPFB Rolling Plan and Forward Budget GDP Gross Domestic Product RS Regional Supply GNP Gross National Product S&G Subventions and Grants GOT Government of Tanzania SGR Strategic Grain Reserve HESAWA Health, Sanitation and Water SIDA Swedish International Development Authority IDA International Development Association SPA Special Program of Assistance for Africa IMF International Monetary Fund SSA Sub-Sahara Africa IPC Investment Promotion Center STD Sexually-Transmitted Disease LART Loans and Advances Realization Trust TA Technical Assistance LGRP Local Government Reform Program TANESCO Tanzania Electric Supply Company LIC Low Income Countries TFC Tanzania Fertilizer Corporation MEFMI Macroeconomic and Financial Management TFR Total Fertility Rate Institute (formerly ESAIRDAM) MOA Ministry of Agriculture TRA Tanzania Revenue Authority MIC Middle Income Countries UK-ODA UK Overseas Development Administration MIT Ministry of Industries and Trade UNDP United Nations Development Program MOCD Ministry of Community Development, Women's UPE Universal Primary Education Affairs and Children MOF Ministry of Finance URT United Republic of Tanzania MOL Ministry of Labor and Youth Development VAT Value-Added Tax MS Ministerial Supply WDR World Development Report MTEP Medium Term Expenditure Program NBC National Bank of Commerce NGO Non-Governmental Organization NORAD Norwegian Agency for Development Cooperation OC Other Charges Vice President: Callisto E. Madavo Director: James W. Adams Technical Manager: Roger Grawe Task Team Leader Sandra Hadler TANZANIA PUBLIC EXPENDITURE REVIEW VOLUME 1 CONTENTS PREFACE SUMMARY OF KEY RECOMMENDATIONS EXECUTIVE SUMMARY ..................................................................................................... i-xxi 1. INTRODUCTION .....................................................................................................................1 2. AN ALYSIS OF PUBLIC EXPENDITU RES ....................................................................3 A. MACROECONOMIC FRAMEWORK..............................................3 B. COMPOSITION OF PUBLIC EXPENDITURES......................................9 Recurrent Versus Development Expenditures: Issues....................14 Breakdown of Recurrent Expenditures .......................15 Debt Service ....................................... .....15 Subventions and Grants ....................................16 Contractual and Contingent Liabilities ............. ..... ........17 Personal Emoluments and Other Charges ............... .........20 C. CONCLUSIONS ...........................................................22 3. THE SIZE AND AFFORDABILITY OF GOVERNMENT...........................................23 A. INTRODUCTION............................................ ...........23 B. INTERNATIONAL COMPARISONS ............................................24 C. CIVIL SERVICE PAY ......................................................25 D. COMPOSITION OF THE WORKFORCE .........................................25 E. AFFORDABILITY ............................................... ..........31 F. PROSPECTS FOR 1997/98 AND BEYOND.................... ...................32 G. CONCLUSIONS AND RECOMMENDATIONS ................................ .....33 4. ISSUES IN STRATEGIC ALLOCATION ......................................................................37 A. SECTOR ALLOCATION ISSUES...................................... 37 B. BUDGETARY PRIORITIES IN THE SOCIAL SECTORs..............................41 C. BUDGETARY PRIORITIES IN THE ECONOMIC SECTORS ................... .........59 D. BUDGETARY PRIORITIES IN PUBLIC ADMINISTRATION ......................61 CONTENTS (cont'd) 5. STRENGTHENING PUBLIC EXPENDITURE MANAGEMENT..............................65 A. SCOPE OF THE PUBLIC EXPENDITURE MANAGEMENT ........... ................65 B. PUBLIC EXPENDITURE MANAGEMENT IN TANZANIA ...................... .......68 C. ORGANIZATIONAL IMPLICATIONS AND DONOR SUPPORT .........................80 6. BUDGET IMPLEMENTATION ISSUES ........................................................................87 A. CASH MANAGEMENT .....................................................87 B. IMPLEMENTATION ISSUES RELATED TO THE WAGE BILL ................... ......92 C. OTHER BUDGET CONTROL AND IMPLEMENTATION ISSUES.......................96 Debt Management ....................................................96 Disbursement Systems ................................................97 TABLES: Table 2.1 Central Government Operations, FY86-FY97 ...................... ..... 4 Table 2.2 Expenditure Deviations from Budget, FY97.................................................................. 8 Table 2.3 Breakdown of Total Expenditures by Economic Category, FY86-97................ 10 Table 2.4 Economic Composition of Ministerial Supply Expenditures, 1992-97 ....................... 16 Table 2.5 Major Subventions and Grants, 1992-97 ......................................... 16 Table 2.6 Analysis of Indirect Transfers ..................................................... 18 Table 3.1 Relative Size and Remuneration of the Government Workforce. International Comparisons, (MRE)............................................. 25 Table 3.2 Tanzania: General Government Workforce, December 1996.......................28 Table 3.3 Tanzania - Pay and Employment Reform...............................32 Table 4.1 Budget Variation By Vote ...............................................38 Table 4.2 Deviation From Budget For Selected Votes, FY 95-96 ..........................39 Table 4.3 Real Government Budgetary Allocations to the Social Sectors 1993/1994 - 1996-97: Recurrent and Development Budgets ................................ 42 Table 4.4 Public Recurrent Expenditures in the Education and Health Sectors, 1989/90 ........... 43 Table 4.5 Public Expenditures on Health & Education compared with Outcomes in Tanzania and Neighboring Countries (1993)... .................................. 44 Table 4.6 Recurrent and Development Budget Allocations By Education Program, 1995/96- 1996/97 .................... .............................. 46 Table 4.7 Education Recurrent Expenditures and Budgeted Allocations by Sub-Sector, 1992/93- 1996/97 ................................................. 47 Table 4.8 Average Public Recurrent Expenditure Per Student by Education Level, 1992/93- 1996/97 ....... ....................................................................................... 48 Table 4.9 Shares of Recurrent Allocations by Input for Primary, Secondary and Teachers Education, 1994/95 - 1996/97 ............................................ 49 Table 4.10 Infant Mortality per 1,000 Live Births 1992................................................................ 51 Table 4.11 Health Recurrent Expenditure and Budget Allocation by Sub-Sector, 1992/93- 1996/97 .................... . .................... ........ 53 CONTENTS (cont'd) Table 4.12 Shares of Recurrent Allocations by Input and Facility Level for Curative Services 1994/95-1996/97 ............................................... 53 Table 4.13 Recurrent and Development Budget Allocations by Health Sector Program 1995/96 and 1996/97 ........................................................ 54 Table 4.14 Water Sector: Recurrent and Development Budget Allocation by Administrative Level 1993/94-1996/97 ....................................... .......... 58 Table 4.15 Share of Ministerial Expenditures in the Economics Sectors, FY 95-97.....................59 Table 4.16 Share of Ministerial Expenditures on Public Administration, FY 95-97 ...........62 Table 5.1 Main Interventions Related of Planning and Resources Management......................... 81 FIGURES: Figure 2.1 Government Budget: Total Expenditure, Domestic Revenue, Recurrent and Development Shares, FY90-97 ....................7............. ...............7 Figure 2.2 Structure of the Government Budget, 1996/97................................ 11 Figure 2.3 Government Expenditure Components, FY90-FY97 .......... ............... 12 Figure 2.4 Government Recurrent Budget Expenditure Components, FY90-97 ...... ........ 13 Figure 3.1 Sectoral Comparison of Pay Levels in Tanzania........... ................. 27 Figure 4.1 Comparison of Government Budget By Ministry, 1996/97....................40 Figure 4.2 Health and Education Sector Shares (% of GDP) ............................... 44 Figure 4.3 Average Public Expenditure Per Public/Student Education Level, 1992/93 - 1996/97 ........................................................ 48 Figure 4.4 International Comparison Expenditures on Public Administration FY 95-97.............62 BOXES: Box 3.1 Potential Reduction in FY98 Workforce ....................... ..........29 Box 5.1 Summary Data on the 1996/97 Development Budget........................77 Box 6.1 Uganda's Experience with Cash Management ........................ .....90 MAP NO. IBRD 27941  PREFACE This Public Expenditure Review (PER) for Tanzania reviews progress made in public expenditure management since the first PER in 1989. The report recognizes both areas where there has been progress and where there has not; it also acknowledges that expectations of institutional progress, were overly ambitious, especially within the time frame defined. The report concludes: one, that if the public expenditure management system is to support Government objectives of accelerating growth with a view to reducing poverty, major improvements are needed in the system at all levels as well as in continuing to redefine the role of Government and two, that with the demonstrated firm commitment of the Third Phase Government to fiscal discipline, there is a greater prospect of the long- recognized institutional and system changes being consistently implemented than at any time in the previous decade. Accordingly, and notwithstanding the outcomes of earlier budget strengthening efforts, the environment is now ripe for an integrated and comprehensive program to revamp the public expenditure system. The report is organized in two volumes. Volume I contains the Executive Summary and agenda for future work and the main report consisting of six chapters: Chapter 1 provides an introduction to the PER; Chapter 2 assesses overall public expenditures; Chapter 3 looks at the size and affordability of Government; Chapter 4 addresses priority sectoral allocations; Chapter 5 addresses the overall strengthening of the public expenditure management system; and Chapter 6 outlines measures to improve budget implementation. Volume II is the Statistical Annex. This report is a joint effort of the Government of Tanzania and the World Bank, with support, from UKODA. The donor mission visited Tanzania from November 11-29, 1996. The mission consisted of Sandra Hadler (mission leader and principal author, World Ban, Aister Moon World W ), David Wood World Rak) Andrew Lawson (consultant, OD), *epben stwer (consultant, ODA) Elizabeth Kariki (consultant, ODA) May Serumaga (consultant, ODA). Andrew FoIner (World Bank) and Nicholas Kessy (Consultant, prepared the Social Sector Data The mission was guided by an Inter-Agency Steering Committee chaired by A. G. Mgonja (Treasury) and a Technical Advisory Group chaired by Mr P. Lyima (udget Commissioner). In addition, several people from the World Bank provided guidance, ineluding: Jim Adams (Country Director) Roger Grawe, (Technical Manager) Mike Stevens (Peer Reviewer), Ronald Brigish and Charles Gffin. Camir Makoye (Resident Mission, Tanzania) provided able research assistance, Finally. Huba Mannoro (Resident Aission. Tanzania), Lorraine James and Roboid Covington (Macroeconomics2) were responsible for the word: processing and physical production of the report. Summary of Key Recommendations Macroeconomic Stability: Reduce the imbalance between expenditures and revenues, continue efforts to improve tax policy and to strengthen revenue (including customs) administration. In the interim, the cash budget system - with improvements - should be maintained to ensure continued fiscal discipline. 4k Role of Government- Government's programs and activities remain severely underfunded. This calls for a continuing priority to redefining the role of Government - within the available resource envelope. Simultaneously, the structure of Government deserves to be revisited. Civil Service Reform: Continue ongoing reform components and, to improve the performan=e of the civil service, adopt a medium-term pay reform strategy targeting selective pay enhancement Priority consideration should be given in FY98 and beyond to downsizing the military and general Government workforce based on efficiency criteria. Roling Plan and Forward Budget (RPFB): The adoption of the RPFB has led to a distinct improvement in planning and budgeting. Nevertheless, strengthening of the medium-term fram is necessary While the principles and objectives of the RPFB remain appropriate, its effectiveness can be strengthened by fin political support for the RPFB (including changing the sequence of budget discussions); strengthening the macro policy capability in Treasury; greater involvement of the sector ministries; and better integration of the Recurrent and Development Budgets (including aid fuds) Integration of the Recurrent and Development Budgets: Government has decided to begin the integration of the budgets with the FY98 budget, This is welcomed as a critical opportunity to improve the planning and coordination of expenditures and to realize a radical rationalization of the Development Budget. Integration will require strong political support, which should be sought also as part of a broader Cabinet commitment to pursuing the RPFB and CSRP in a coordinated and sequenced manner. Composition of Expenditures: Ensure adequate funding of other charges and development expenditures - control the share of debt service and the wage bill in aggregate expenditures and introduce systems to improve their predictability. Simultaneously, continue efforts to reduce expenditures on non-essential components of the other charges and to increase the efficiency of outlays. Development Budget: While the planned development budget share of the total budget is not high, the development budget is seriously over-extended. The 1400 some projects in the budget cannot be managed, nor can the recurrent expenditures to realize returns on the investments be afforded, A radical rationalization of the budget is recommended: namely, to limit the composition of the budget to those priority projects for which the Government allocates counterpart fands. In PY9, Government allocations to the development budget barely funded the 125 super care projects. Prior to the introduction of the FY98 budget, agreement should be reached with the relevant donors on the future ofnon-priority projects. Allocation of Publk Resources to the Social and Infrastructure Sectors: It is proposed that key programs in these sectors be allocated "protected expenditure status" under the cash budget system and that any additionality of funds (inter alia from debt relief) should be allocated to these sect4m. Aid Management: To improve the Government's capacity to manage and control externally- funded expenditures adopt comprehensive sector strategies for donors to fud jointly with Government, instead of continuing the current management-intensive, project by project donor by donor approach. This strategy would also address inclusion of aid flows in the budget. Local Government: Government is proceeding apace with its program to streamline the regional administration and simultaneously, strengthen the local governments, who shoulder the responsibility for the delivery of key social services. It is essential that adequate resources be provided to ensure the timely and successful implementation of this important reform program. EXECUTIVE SUMMARY "I firmly believe that the main problem in Government is the lack of discipline. Not respecting ceilings as given in the votes of expenditure is part of lack of discipline, and adversely affects services...." President Mkapa in his inaugural address to the National Assembly in Dar es Salaam on November 30, 1995. Introduction 1. This report addresses one of the central development challenges facing Tanzania's policymakers: how to improve budget outcomes while remaining within the tight constraints indicated by fiscal stabilization. The report finds that improvements in both budget structure and systems are essential. Specifically, the PER suggests that there is an urgent need for the Government of Tanzania to: * realign the balance of expenditures towards other charges and the development program; * improve the efficiency and effectiveness of its expenditures; * strengthen the overall expenditure management system; * and simultaneously, to improve budget implementation. Proposals are made on how to realize the needed improvements. While not a focus of this report, improvements in revenue collection are recognized as key to improving fiscal outcomes. 2. The report notes the following areas where there has been significant reform: * Significant progress has been realized in redefining the role of Government in the economy. Economic management is now more market-oriented and, although slow to commence, the privatization program is well under way; * A comprehensive civil service reform program is also under way: there has been a significant retrenchment program and some reform of pay structures. Ministry-by-ministry functional and efficiency reviews are still going on; * Reform of the planning and budget system has included the adoption of a three-year Rolling Plan and Forward Budget (RPFB). Government has begun the integration of the Recurrent and Development Budgets. Parallel efforts to strengthen accounting and audit systems are now making significant headway; ni Executive Summary * The recent establishment of the Tanzania Revenue Authority has led to a significant improvement in revenue collection and the prospect is for continued improvement; * The present Government has demonstrated a firm commitment to fiscal discipline, matching expenditures with the available resources; this has led to a significant improvement in the macro aggregates. 3. The report recognises that the measures implemented so far have not yet achieved the transition to a sustainable pattern of public expenditures, focused on the effective and efficient delivery of core public services. This does not mean that past prescriptions were inappropriate, although the difficulties of implementation may have been underestimated. 4. The PER takes the view that many of the necessary pieces are in place, and the requirement is to adjust and build on them. Given the interconnections between different elements of the public expenditure management system, it is also argued that the needed reforms must complement each other. The report points to some difficult decisions ahead, particularly with respect to the size of Government. It concludes that Government should, as a first priority in the budget area, formulate in FY98 an integrated public expenditure management strengthening program, which could then provide the framework for donor support. The attached matrix summarizes key components of such a program. Macro Stability 5. Tanzania's macroeconomic situation continues to strengthen: the performance of investment, savings, and exports have all improved. Nevertheless, the situation remains fragile: the savings-investment gap remains large, official exports finance only about one- half of imports, and investment flows remain at low levels, leaving Tanzania dependent on foreign aid to finance much of its imports, development expenditures and some recurrent expenditures. Continuation of the current tight fiscal and monetary policies together with continuing efforts to bring inflation down will be essential through the medium-term. Furthermore, appropriate market-based incentives for raising the domestic savings rate and improving the efficiency of investment are necessary if Tanzania is to raise its trend rate of growth and reduce its dependency on foreign aid. 6 While there have been solid, structural improvements in the economy in recent years, Tanzania's economic performance between FY93 and FY96 deteriorated, in large part, because of fiscal imbalance. Through FY92, the Government had maintained a tight fiscal stance and, as a result, generated a fiscal surplus after grants. In FY93, the fiscal situation deteriorated dramatically and the fiscal deficit (after grants) jumped to 6.4 percent of GDP. In the absence of strong political commitment to fiscal discipline, return to the pre-FY93 situation proved elusive until early 1996, when the Third Phase Government demonstrated a new commitment to achieving macroeconomic stability. This commitment has entailed a serious effort both to improve the domestic revenue effort, and a determination to keep aggregate expenditure within the limits of the macro Executive Summary iii program and to adjust expenditure as required by the performance of domestic revenue. As a result, the macroeconomic objectives of the FY97 program are being achieved. 7. The significance of the Government's commitment to fiscal discipline needs to be acknowledged. Firstly, fiscal discipline at the aggregate level is the most important factor in achieving macroeconomic stability, which development experience worldwide has confirmed to be a fundamental precondition for sustained economic growth and poverty reduction. Realism and prudence in the management of the macroeconomic framework for the budget is a necessary condition for virtually any other reform in public expenditure to be addressed in the present review. In particular, establishment of a credible macroeconomic base is a necessary first step in the move towards an effective framework for medium-term expenditure planning, improving strategic allocation within the budget and enhancing incentives for efficient resource use. Composition of Public Expenditures 8. Notwithstanding the Government commitment to fiscal discipline, the Tanzanian budget is not yet robust. After priority claims on total expenditures of debt service and wages, little remains for the running of Government (Other Charges) or the development program. Specifically, the share of debt service and wages in total expenditures has grown from under one-half in FY86-92 to two-thirds of total expenditures in FY96, while simultaneously the share of Other Charges and development disbursements declined from one-half to less than one-third. In addressing this imbalance, the Government of Tanzania faces a formidable challenge. This will necessitate difficult decisions regarding, inter alia: * the strategic allocation of funds in line with Government policy priorities; * measures to increase the efficiency and effectiveness of expenditures; * and, in particular, the size and affordability of Government. 9. Total debt service payments have risen steadily in the last three years as a proportion of Government's total recurrent expenditure, from about 29 per cent in FY86- 92 to 34 per cent in FY96. The impact of the recent Paris Club agreement and lower interest rates on domestic debt service will relieve some of the pressure on non-debt expenditure, both recurrent and development'. The concern from the budget standpoint, however, is not only with the level of debt service but the estimation of service payments. Tanzania does not have an effectively functioning debt management system and this has contributed to volatility in service estimates. A Presidential Debt Task Force was established in late 1996, and as a result, efforts are being made to reconcile the debt stock data and to operationalize the debt management system by end 1997. A review of Tanzania's debt sustainability shows that while debt distress continues to plague both the fiscal and external positions, debt indicators will improve steadily through the medium term, assuming strict adherence to the ESAF and Paris Club agreements, and future borrowing only on highly concessional terms. iv Executive Summary 10. Of greater concern, however, is the increase in recent years in the share of the wage bill - from approximately 17% in FY86-92 to 32% in FY96, notwithstanding the significant retrenchment program. The problem of the share of the wage bill, however, goes beyond the size of government to the underlying structure of the budget system and its implementation. Since the problem of overcommitment on the wage bill has been recognized for some years, the report explores in some detail the factors which obstruct avenues for correcting the problem. It is widely recognized that the present imbalance between wages and other charges2 is seriously inefficient and imposes a cut in real service delivery far more severe than would be required by a balanced reduction in line with aggregate expenditure control. 11. In the past, the sectoral "hard budget" constraints were not enforced and, since wages were a "given" in a sector/agency budget, there was little incentive to seek a more appropriate balance between wages and other charges. The commitment now to fiscal discipline, provides such an incentive. Also, the FY98 Budget Guidelines provide for the first time an opportunity to realize an appropriate balance between wage and non-wage expenditures, by providing for savings from wages to be retained within the agency. Potentially, this is an important first step to both rationalizing aggregate resource allocation and improving the efficiency of resource use. Its impact will need to be closely monitored in the course of FY98. 12. In addition to controlling the growth in the wage bill to free up resources for essential components of Other Charges, there also remains scope for improving the allocation of expenditures within the Other Charges category by reallocating from non- essential to essential inputs. In the education sector for example, a large proportion of the sector's budget continues to be spent on food and housing for students and overseas studies, rather than on textbooks and school maintenance. Similar allocative issues are evident in other sectors. In total, the PER concludes that such savings (including increased cost sharing) could amount to amount to 5% of discretionary expenditures. The Government also attempted to rationalize expenditures systematically when it set up the Presidential "Cost-Cutting Commission" in mid-1993 and charged it with the responsibility of identifying areas for savings across a wide range of Government activities. The Report of the Commission provided a wide range of proposals for cost cutting, including for instance, the reduction in the number of Ministries from 23 to 15. Many of the recommendations, however, were not implemented. The PER recommends that a review of the potential savings from non-essential expenditures be undertaken and implemented in FY98 and, in particular that the recommendations of the Cost Cutting Commission be revisited. The importance of such a review is accentuated by the changing role of Government in the economy, which would indicate prima facie a need for a smaller administration. The PER also finds that the share of the budget allocated to 2 For instance in FY97, 91 per cent of recurrent allocations to primary education are for wages. Executive Summary v administration in Tanzania is high relative to neighboring countries, suggesting another area of potential cost savings. Role of Government 13. The prime objective of the Government is to accelerate growth while simultaneously reducing poverty. The Government's intention is to accomplish this by means of providing an appropriate macroenvironment for private sector development and to focus public expenditures on priority economic, social and administrative infrastructure. Translating these functions of Government into public expenditure priorities has proven elusive. The 1989 PER found that the inability to define priorities and focus expenditures had resulted in chronic underfunding of most programs. The 1994 PER anticipated that the Rolling Plan and Forward Budget (RPFB) - the newly introduced medium-term expenditure framework - would facilitate the focusing of Government expenditures on priority areas. This has not occurred. In the context of the preparations of the Fifth RPFB, progress in priority delineation may now be realized by undertaking a reassessment of the Operation and Efficiency Reviews (O & E, Phase II) by a strengthened efficiency review team in CSRS/CSD to ensure both that affordability criteria govern the core program and standards of service delivery and that non-core activities are eliminated. Size and Affordability of Government 14. Over the past decade there have been numerous commissions, reviews and studies of Government performance in Tanzania. These investigations were designed to address weaknesses in Government policy making and execution that have been manifested in inappropriate economic policies, failing social services, deteriorating infrastructure, poor civil service performance and increasing public sector corruption. The conclusions of the studies are consistent, namely: * the present scope of Government activities and the range of Government services are no longer affordable; also, that they are increasingly inappropriate given the changing role of the Government in the economy; * the role and functions of Government need to be redefined within Tanzania's resource envelope; * core programs are seriously under-resourced in terms of non-wage operating costs; * the wage bill share is too high; * civil service wage levels are low. 15. The key issue facing Government here is how to control the wage bill share and simultaneously deepen the pay reform such that the incentive system can be improved. While there is no one correct size for a civil service, affordability is key. One approach noted in the PER is that of the civil service as a percentage of the population: namely, Tanzania's civil service over the past three decades grew at 4.5 per cent per annum while population grew at 3 per cent per annum. If the civil service had expanded at the same vi Executive Summary rate as population, then its size today would have been 225,000 - rather than 275,000. Since the Government's role and functions are contracting, this further suggests over- manning. Serious consideration should be given to expanding the retrenchment program on efficiency criteria, so as to realize an affordable wage bill, payment of a minimum living wage, increased resources for non-wage costs and improved services. 16. The present reforms are not achieving these objectives. Notwithstanding the retrenchment of about 25 per cent of the Government workforce since FY92, the wage bill share of total expenditure has virtually doubled - reflecting in large part, Government's efforts to restore real wages to their early 1980 levels. The magnitude of Government's dilemma is illustrated by the fact that 94 per cent of the general Government workforce currently earns less than T Sh 65,000/ month ($105), the estimated minimum living wage in Dar es Salaam in 1996 and that almost one quarter of the workforce currently earns less than $60/month. Further, since the 1980s, Government pay levels have lost ground to both the parastatal and the private sectors, and are now seriously uncompetitive: in 1996, average civil service earnings over a range of benchmark jobs were 34 per cent of average private sector earnings. 17. The impact of pay policy in Tanzania since the mid-seventies has been a decline in the professional and technical capacity in the civil service. Many middle-level staff now lack the necessary skills. At current pay levels and given the differentials between public and private sector pay, there is little prospect of being able to increase professional and technical employment in the Government service. The civil service has not only lost professional and technical staff but has been unable to recruit successors, resulting in an aging workforce and increasing reliance on contract-based employment (often donor funded) to carry out key management responsibilities. The increasing loss of civil service capacity and the failure to develop future managers is not sustainable and has important implications for the quality of civil service performance. 18. The PER concludes that improvements in the efficiency of Government administration in Tanzania will depend on further changes to the structure of Government pay and employment. Namely: * further reductions in the size of the civilian (including education and health employees in local government and non-commercial parastatal employees) and military Government workforce based on efficiency criteria, continued restrictions on recruitment, and improved payroll integrity, are required to release resources for pay reform; * a medium-term pay reform strategy should be adopted as a key component of the civil service reform program. The strategy should take into account the competitiveness of the Government pay structure and capacity building requirements in the form of targeted pay increases, especially for professional and technical groups. Affordable annual pay reform targets should be set. A wage simulation model used in the PER demonstrates that it is only with Executive Summary vii continuing efficiency-based redundancies that substantial progress in deepening pay reform could be realized, while simultaneously containing the share of the wage bill; * it is also proposed that on a priority basis in FY98, a review be undertaken of the options for privatizing, commercializing or closing the 85 non- commercial parastatals, that employ approximately 45,000 staff, in addition to the general government workforce. A major rationalization is needed of these parastatals and will have a clear payoff. 19. Further, wage bill ceilings for FY99 and beyond should be based on the RPFB and budget framework. The implications of wage bill ceilings and pay reform targets for the overall size of the Government workforce should be derived and reconciled with the workforce implications from the O&E program for central ministries and local Government, and efficiency reviews of the workforce in health, education, and law and order sectors. In this way affordability criteria should determine the Government's core program and standards of service delivery. Strengthening of Public Expenditure Management 20. There is now wide agreement within Government and amongst Donors, that it is necessary to take a broad view of public expenditure management - one that is concerned with all stages in the budget cycle, including implementation as well as formulation of the budget. Specifically, that the effectiveness of the system depends as much on budget implementation (expenditure control, accounting, monitoring and audit) as on planning and budget preparation. There needs to be a comprehensive view of public expenditures - including recurrent and development expenditures, both on-budget and off-budget items (the latter still include a large proportion of aid flows and significant flows to parastatals) and revenues. It is important too, to consider personnel as well as financial resource management. 21. The PER distinguishes between the three main levels at which the public expenditure management system operates: * macro: the control of budget aggregates in a way that supports overall economic management; * strategic: the allocation of public resources in ways that reflect public policy priorities; * operational: the efficiency and effectiveness of resource use in implementation. 22. In reviewing public expenditures, there is a tendency to focus on allocative shares (reflecting the competitive process of the annual budget and competition for development funds etc), but macro and operational dimensions are equally important. Moreover, there are tensions between the different levels. Specifically, competition for strategic viii Executive Summary allocations invariably creates pressure to adopt a budget that is too ambitious to be sustainable in terms of macroeconomic management. The pressure to overcommit Government resources has direct implications at the operational level: because there are insufficient resources adequately to finance all the activities that Government attempts, implementing agencies find that nominal budgets are less than the full cost of an effective service, and, often, that funds released are less than those budgeted. 23. In summary, it is not sufficient to specify a level and pattern of public expenditures that, for Tanzania's present economic circumstances, reconciles resource availability with consistent policy priorities. There is a deeper need to develop a public expenditure management system that can continually reconcile macro, strategic, and operational objectives, adjusting to changing circumstances from one budget cycle to the next (and also during budget implementation within each fiscal year) and which avoids getting stuck in suboptimal (inefficient) patterns of resource allocation and use. The report argues that this entails: * building a national consensus about objectives, resource levels, and the strategic priorities for government so as to limit the extent of overcommitment of resources and political decisions disrupting expenditure plans early in the budget year; * discipline in implementation - this is both about control at macro level so as to serve the objectives of fiscal stabilization, and about predictability of resources at implementing agency level, so that the limited resources that are available can be allocated to priority services and used efficiently; * a comprehensive approach to public expenditure management (encompassing aid, development, recurrent, implicit flows, deployment of personnel, effective implementation systems etc.) - both to ensure discipline and to contain shocks; * linking short- and medium-term perspectives, so that short-term imperatives do not dominate. Central to these objectives are four main elements of public expenditure strengthening: * the Rolling Plan and Forward Budget; * the integration of recurrent and development budgets * the development budget; * issues concerned with program costing, planning and management at the sector level. 24. The Rolling Plan and Forward Budget represents a distinct improvement on previous approaches to planning and budgeting in Tanzania but, in the fourth year of the RPFB system, there is a widespread feeling that the RPFB has not met expectations. There are several reasons why the RPFB has not succeeded in becoming a functional medium-term expenditure framework. Budget projections were over-optimistic, thus Executive Summary ix evading the problem of attaining a realistic initial budget. The Government's political commitment to the budget figures it announced was weak: thus, there were frequently large reallocations early in the budget year. General expenditure discipline was poor, and budgets were also thrown off by expenditure arrears carried forward. Weaknesses in estimating, monitoring and accounting were another source of uncertainties. Nonetheless, the principles and objectives of the RPFB approach remain appropriate, and given the present government's seriousness of purpose, the RPFB can still be an effective instrument for expenditure planning The principal requirements are: * Firm political support for the RPFB; * Changes to the timetable and sequence of budgetary discussions so as to advance the review and approval of the medium-term program by Cabinet ahead of the budget; * Strengthening of the technical foundations of the RPFB and, in particular, of the macro-analytical and policy capability of the Treasury; * Greater involvement, and a more responsible role for, sectoral ministries; * Better integration of recurrent and development budgeting (including aid funds). 25. Integration of the Recurrent and Development Budgets: The Ministry of Finance assumed early in 1997 the Planning Commission's responsibility for preparation and administration of the development budget. This provides a critical opportunity not only to realize integration in the planning of recurrent and development expenditures but also to achieve a rapid and radical rationalization of the portfolio of development projects. In the longer term, the Government should reconsider whether maintaining a separate development budget continues to be appropriate. In the short term, the priority must be to ensure that, both at sector level as well as in the center, expenditure programming takes place on the basis of a comprehensive view of all resources applied to the sector. This objective can be served by: * issuing budget guidelines that focus on the total ministry allocation of domestic resources, rather than beginning with an arbitrary division of available funds into recurrent and development; * ensuring that submission and review of estimates deal jointly with recurrent and development proposals; * undertaking an immediate and radical rationalization of projects: * close liaison with donors, both to ensure that government's information about donor projects is complete and consistent, and to ensure that donors observe government priorities in project funding (including withdrawing and abstaining from projects government designates as non-priority). 26. The Development Budget: The existence of separate recurrent and development budgets is widespread, but its appropriateness is increasingly questioned. Traditionally the development budget was conceived as an investment budget focusing on capital x Executive Summary projects. In principle there are some advantages in this. Capital projects are typically one-off activities with a multi-year time scale, and in practice much of the development budget was financed by donors, and most donor funds were directed towards such projects, so there were certain advantages in having a separate administrative channel for them. But the existence of a separate development budget can have serious disadvantages too. It tends to build in an upward pressure on government expenditure commitments: this happens because of the natural tendency of the responsible agency to try to expand the resources it controls, exacerbated by the fact that it is not directly responsible for the eventual recurrent costs of the projects it sponsors. As public expenditures come more under pressure, there is a tendency, as in the recurrent budget, for resources to be spread too thinly across too many projects. At the same time, the development budget comes to be seen as an alternative source for essentially recurrent funding. The separation of responsibilities for recurrent and development budgets makes coherent sector planning more difficult. As projects proliferate there is a danger that the planning agency becomes preoccupied with administrative detail, losing sight of strategic issues. 27. These weaknesses have been very apparent in Tanzania. The previous PERs acknowledged that the large number of projects is unmanageable and that the recurrent expenditures needed to realize returns on the investments cannot be afforded. In recent years, adequate Government funding has not been available - even for the approximately 125 "super core" projects. 28. The present situation of minimal Government funding of the development budget highlights the need for a rapid and thorough rationalization of projects. Criteria for inclusion of projects in the budget are, inter alia: * the project is in the Government's priority sectors of education, health, water, economic infrastructure and environment; * the project has adequate documentation, inter alia justifying its financial and economic viability; * has not been under implementation for more than, say 10 years; * is not recurrent in nature; * could not be undertaken in the private sector; * projects with donor funding would only be included in the budget if there is a firm commitment for disbursements prior to completion of the budget; * projects that do not report on donor funds would be excluded from the Government budget. Based on these criteria, it should be possible to reduce the number of projects from over 1,400 to a manageable 300 or so projects over the next two years. Undertaking such a rationalization, however, will require strong political support, as part of a broader Cabinet commitment to pursuing the RPFB and CSRP in a coordinated and carefully sequenced fashion. Close coordination with the donors in the closure or divestiture of existing projects will be needed. Executive Summary xi 29. Public Expenditure Management at Sector Level: At the sector level, the requirements for better use of public expenditures are two-fold: one, there is a need to carry through a restructuring of public expenditure in the sector so as to match government activities to the level of resources that is available on a sustainable basis; two, there need to be improvements in the efficiency with which available resources are used. While there have been significant changes in Government policies concerning its sector activities, these have not, in most cases, been adequately reflected in budgets and expenditure programming. 30. Part of the problem is that implementing agencies have continued to live in hope of an increase in real funding levels that would allow existing activities to be fully funded, and have persisted with short-term tactics of spreading available resources across all activities, rather than make a radical adjustment that would involve curtailing some activities in order to fund other properly. But even where the need for a restructuring of activities and expenditures is recognized, the existing framework of public expenditure management inhibits reform. Sector ministries have not been given a clear, or politically credible, indication of the esource framework on which their medium term expenditure planning should be based. Although there have been exhortations to make savings and efficiency improvements, the incentive to do so is lacking when line agencies' perception is that any savings they make are likely to be appropriated by the center and reduce the sector's future resources even further. The present institutional framework inhibits fundamental expenditure reform in other ways too: a comprehensive sector-wide view of requirements is inhibited by the fragmentation of expenditure planning between recurrent and development budgets, and the further complication that aid funding is often determined on a piecemeal, project by project basis. Reviews of government's role in each sector (notably the CSRP-coordinated organization and efficiency reviews) have not been given clear resource parameters to start from. The established line-item incremental approach to budgeting does not support the non-marginal adjustments that are needed; it focuses on the costs of inputs and does not facilitate analysis of the costs associated with different outputs. It is difficult to discover the current costs of different programs, let alone what their costs would be if adequately funded. Thus it is difficult to present policy makers with clear expenditure options and trade-offs. 31. Efforts to restructure and improve the efficiency of public expenditures at sector level need to proceed along the following lines: * Line ministries need to base their programs on explicit medium-term budget projections that are politically endorsed as reflecting Government's sector priorities; * Sector ministries should be given more responsibility for preparing their sections of the RPFB, which should include explicit links between policies and sector expenditure proposals based on affordability, together with time- xii Executive Summary bound sector action programs that would provide a framework for monitoring implementation; * Continuing moves to increase the efficiency of resource use by giving more autonomy to implementing agencies, and increasing the scope for competition between public and private providers; * That options in each sector are reviewed on a sector-wide basis. There also needs to be sector-level coordination with donors to ensure that project interventions are consistent with a sustainable sector expenditure strategy; * This PER has not focused on the reform and strengthening of local government; however, the balance of responsibilities as between central and local government bodies will be a critical issue for review in most sectors. Improving Budget Implementation 32. The effective implementation of the budget requires timely and reliable data. Current systems make it extremely difficult to estimate, monitor and account for expenditures (and revenues) and are thus, undermining implementation of the budget. Areas in particular need of strengthening are debt service, the wage bill, aid flows, disbursement systems, the monitoring of aggregate expenditures and revenues, and user charges. On-going initiatives to improve systems in many of these areas are assessed and it is concluded that significant improvements are beginning now to impact on budget implementation, in particular with the Central Payments System and Sub-Treasuries. Recommendations, however, call for further resources and an acceleration of implementation, and in particular, for the payroll and aid management systems. Aid Management 33. Despite Government's objective of reducing dependence on aid, aid continues to underpin public expenditures. The relationship between Tanzania and its aid partners has often been suboptimal on both sides, as is well described in the Report of the Group of Independent Advisers on Development Cooperation Issues between Tanzania and its Donors (the Helleiner report). Aid flows, including technical assistance, and project management structures have often bypassed regular government structures in attempts to circumvent the weaknesses of public administration and expenditure management. However, donors' independent actions have both contributed to and perpetuated those weaknesses, inhibiting strategic national and sector planning and expenditure programming. This is a situation that needs to be corrected in the near future. 34. The approach to public expenditure management set out in this PER requires Government and donors to collaborate in ensuring that: * Government takes the lead in planning, prioritization and expenditure programming leadership of the institutional reform process; Executive Summary xiii * a coordinated CSRP/RPFB exercise provides the framework for articulating national priorities and programs, both domestically and in relations with the donor community; * a sector-wide approach to planning and aid coordination is increasingly adopted, with donors collaborating to support coherent sector programs that address both the recurrent expenditure and the investment requirements of each sector as a whole; * Government is more proactive in liaison with donors both during budget preparation and during budget implementation to ensure that aid flows are anticipated and captured; * aid agencies take seriously their commitment to allow Government to take the lead , and respect the priorities Government adopts; * both Government and aid agencies as far as possible avoid project management arrangement that bypass existing organizational structures; * and a coordinated budget calendar is developed so that necessary Government dialogue with the international financial agencies is integrated with the national expenditure programming cycle, and not superimposed on it. Strategic Allocation Issues 35. Overall, the PER finds that the strategic allocation of funds across sectors is largely consistent with Government's priorities and, in recent years, that there has been progress in increasing allocations to high priority activities. The key issue in strategic allocation is the development of realizable, medium-term sector development programs, consistent with the macro objectives, and the enforcement of the hard budget constraints at the sector and program level. 36. Until this fiscal year, non-compliance (deviations) from the sector allocations has been a serious concern. At the beginning of FY97, the resulting overhang of domestic arrears exceeded T Sh 50 billion. The placement of accounting officers within agencies and the on-going implementation of the Central Payments Systems is expected to significantly reduce this source of budgetary deviation. The PER also notes other measures that will help improve compliance with sector budget allocations. 37. Given Tanzania's constrained budget outlook over the next several years, there is little prospect for increasing real allocations to the priority sectors or programs. Thus measures to improve the efficiency of expenditures are vital - especially in the social sectors, where Government is facing impossibly harsh decisions. Nonetheless, this report has identified activities that can be cut immediately or deferred. Combined with the potential for increased cost sharing, these amount to approximately 5 per cent of discretionary expenditures. The report also finds that Tanzania allocates a higher proportion of expenditures to general administration than in other countries in the region and proposes that efforts be made over the next two to three years to reduce the share of general administration. xiv Executive Summary 38. Reflecting the significance of the social sector to Government's poverty alleviation strategy, the report focuses on social sector expenditures. In terms of the likely future budget resource scenario, the PER concludes that it will be unlikely in the near term to be able to do more than preserve the real per capita expenditure level to education and health. Comparisons with neighboring countries, however, show that in terms of public expenditure as a share of GDP, Tanzania's allocations are relatively high. The focus of the analysis is therefore on raising the effectiveness and efficiency of expenditures within these sectors, with due consideration to equity and service quality. 39. The poverty focus of expenditure reallocation: The priority allocation of public resources to primary education, basic health and economic infrastructure should in itself lead to an improvement in the living standards of the Tanzanian population. However, it is possible and desirable to direct additional expenditures towards the poorer segments of the population. Since the bulk of local government expenditures are on primary education and basic health, one way of improving the delivery of social services to the poor is via the efforts to strengthen the local governments and to channel at least some of the additional resources through them to the poor. While local government issues have not been addressed by the PER, the mission is in agreement with the Government's program to streamline the regional administrations and strengthen the local governments, who bear responsibility for delivery of key social services. Adequate funding of the program is essential for its successful and timely implementation. It is recommended that local government issues constitute a focus of the 1998 PER. PER Recommendations: Main Interventions related to Planning and Resource Management Issue Recommendation Timing Lead Agency' c Macro Stability -Increase revenues from approximately Improve tax administration; reduce exemptions; broaden tax On-going TRA 15% to 17% of GDP in FY98 base; increase compliance; -Cash Management System -Internalize the discipline so as to maintain fiscal balance; - Simultaneously, improve operation of system through: Immediate Treasury/BOT - Improving forecasts of key expenditure components; - Improve accounting for aid flows; - Broadening composition of cash management committee; -Introduce "protected expenditure" status for priority programs. Role of Government Government's programs and activities O&E Phase II reviews to be reassessed by a strengthened FY98 CSD remain severely underfunded efficiency review team in CSRC/CSD to ensure that non- core activities are eliminated and to ensure that affordability criteria govern the core program, strengthen capacity- building and incentives so as to enhance standards of service delivery. Size and Affordability of Government -Implement O&E phase II recommendations FY98 Government agencies -Control wage bill share of expenditures -Extend O&E reviews to local government, including Begin FY98 teachers and health workers -Further reduction in size of general Government including FY98 and military and national service) based on efficiency criteria; beyond 1 Government will bring to the donor community activities for which support will be sought. Issue Recommendation Timing Lead Agency" -Establish a retrenchment fund; Begin FY98 -Undertake review of potential for commercializing, FY98 CSD privatizing or closing the 85 public institutions -Early review of privatization/commercialization options of On-going CSD/ODA 37 identified agencies under Executive Agency Program -Introduce improved payroll controls and audits together FY98 CSD/ODA with new payroll system; -Continue restrictions on civil service recruitment On-going -Review promotions policy FY98 CSD//CSC/Treasury -Civil service wages are low -Adopt medium-term pay reform policy with annual targets FY98 CSD Strengthen RPFB -Firm political commitment; } ) -Strengthen foundations of RPFB to increase realism of } } medium-term frame; } } - Strengthen technical functions of the RPFB (especially } Immediate } Treasury macro capacity in Treasury); } } -Greater involvement (realism) and more responsibility for } } sectors in RPFB; } ) -Better integration of recurrent and development budgets; } } Integration of Recurrent and Development Budgets Will provide critical opportunity to -Ensure that at macro and sectoral level, expenditure FY98 } improve the planning and coordination of programming takes place on basis of comprehensive } expenditures and rationalize the review of resources available to sector. } Development Budget. -FY99: issue Budget Guidelines that focus on ministerial FY99 }Treasury allocations rather than arbitrary division between recurrent } and development expenditures. } - increase strategic and analytic capability of Treasury Immediate } budget division } - Integrate the Budget Books FY2000 } Issue Recommendation Timing Lead Agency' Rationalize Development Budget -Undertake radical rationalization of projects (to By FY99 Treasury/Ministries/ Number of projects unmanageably large approximately 300) in priority sectors and for which Regions/Donors and underfunded; recurrent expenditures to Government funding is available; realize returns on investments not -Liase with donors to ensure that donor disbursements are Immediate Treasury /Donors affordable. Donor project funds not contained in the budget; captured in budget. -Establish task force to address divestiture/liquidation of FY98 Treasury discontinued projects Aid Management Aid flows and project management Consistent with the Helleiner Report: Immediate structures bypass Govemnment systems - Government to elicit donor support for agreed sector programs, addressing both recurrent expenditures and investment requirements as a whole; -Government to be more proactive in liasing with donors during budget preparation and implementation to ensure that aid flows are anticipated and captured. Composition of Public Expenditures Crowding out of Other Charges and Development Expenditures - by debt service -Long-term: Will improve with time: Paris Club, a possible HIPC and lower domestic interest rates; - Short-term: Complete data reconciliation and implement End 1997 Treasury/BOT with Donor - by wages: CS-DRMS to help increase reliability of budget estimates; support - Improve integrity and control over wage bill (see above on affordability) Other Charges: crowded out by priority - Protect priority expenditures; } ) allocations to debt service, wages and - Eliminate expenditures on low priority activities which } } C&CL account for some 5% of discretionary expenditures; }FY98 }Budget Issue Recommendation Timing Lead Agency' -Revisit recommendations of the Cost Cutting Commission; }Steering Committee - Focus on increased efficiency in use of resources across } sectors. } International Subventions: Membership Review and rationalize membership FY98 in organizations has increased in recent years to 150. Domestic Subventions: Budgetary Undertake review of potential for privatization, FY98 CSD support to the 85 public institutions commercialization or liquidation of these institutions Executive Agency Project: Early review of the commercialization/privatization options FY98-99 Government/ODA of the 37 identified agencies. Contractual and Contingent Liabilities (C&CL): } Direct and indirect budgetary support -Tighten discipline in financial operations of parastatals and }Immediate Treasury/NBC/TRA/PSRC increase transparency } -Suspension of utilities services for non-payment; } -Enforce payment of tax liabilities Contingent Liabilities -Following completion of stock taking exercise, adopt FY98 Treasury policy stance for future. Contingencies -Establish transparent framework for contingencies and set FY98 Treasury a ceiling Strategic Allocation Issues Non-compliance with Sectoral Budget -Strengthen realism of macro framework Immediate Treasury/Planning/BOT constraints: Average deviation by vote -Strengthen medium-term sector expenditure budgets between outturn and budget in FY95-96 (capital and recurrent) and their link to sector objectives and Treasury/Sector Agencies was over 50%. Such non-compliance policies; I Issue Recommendation Timing Lead Agency' undermines budget process -Hold budget managers accountable for actions -Continue practice of monitoring commitments through stationing of accounts officers; -Accelerate coverage of the CPS to remaining agencies and FY98 AGD/SIDA to commitments. Education: Improve quality of services At a minimum: within constrained budget framework and - Maintain real per capita expenditures at the FY97 level. } ) in line with sectoral objectives Any additionality of funds to the budget, targeted first to } ) raising per student primary allocations; } ) - To ensure continued increases in enrollments at all levels, } ) target increase in pupil: teacher ratios, especially in } ) secondary and tertiary education; ) ) - In line with the Master Plan for Education, increase )On-going )MOE/MOSTHE/Treasury efficiency in use of staff and physical infrastructures; ) ) increase cost recovery on catering and boarding costs to ) ) drive down per student expenditures at secondary and } ) tertiary levels and simultaneously expand enrollments } } within limits of existing facilities; enabling resource shift } } in favor of primary education, where quality } } improvements are necessary Health: Improve quality of services within At a minimum: constrained budget framework in line with - Maintain real per capita expenditures on health at FY97 } } sector objectives levels; }On-going }MOH /Treasury - Continue intrasectoral allocations in favor of preventive } } and primary services; reduce subsidies to overseas } } treatments, and continue efforts to extend and deepen cost } } sharing arrangements; - Complete health sector strategy and integrate with RPFB: }1997 }MOH Issue Recommendation Timing Lead Agency' this to include measures to restructure expenditures and } raise efficiency, detailing changes entailed and provide performance targets. Public Administration Tanzania's share of expenditures on To free up resources for the social sectors, Government FY98-00 public administration is higher than in should set a target of reducing the allocation to public neighboring countries administration over the next 3 years Budget Implementation Cash Management System (See Macro Stability above) Wage Bill -Obsolete payroll system and inadequate -Accelerate implementation of the Personnel Control } } personnel information system Information System; }FY98 }CSD/ODA -Accelerate completion of new payroll software system; } } -Undertake comprehensive payroll audit; } CSD/OCAG -Promotions process contributes to wage -Review statutory promotions process that is merit-based CSD/CSC/Treasury bill uncertainty and supported by the budget. Debt Management (See above Composition of Public Expenditures) Disbursement System Need to regain control over expenditures Recent introduction of Central Payments System and Sub- and improve monitoring and accounting Treasuries has significantly improved control, FY98 AGD/SIDA/ODA for funds accountability and timeliness. Extending CPS to all agencies (including Military) and Sub-Treasuries nationwide is a priority; Issue Recommendation Timing Lead Agency' - Simultaneously, accelerate adoption of second phase of CPS including commitments-based system, integrated to FY98 AGD other books of account is recommended Reporting and Compliance on Revenue To improve compliance: AGD and Expenditures -Develop an internal audit strengthening program; FY98 -Apply penalties/sanctions to officers failing to comply with Immediate regulations; -Introduce and enforce revised Code of Ethics and Conduct FY98 for Civil Servants -Introduce standardized reporting system for user charges; FY98 AGD -Undertake tracking study of Public Expenditure in primary FY98 Treasury/MOE/MOH education and health for recent years in a sample of districts with objective of establishing effective expenditure system Local Government The PER has not addressed the issue, but is in agreement with Government's program to streamline the regional administration and strengthen the local governments, who bear responsibility for delivery of key social services. It is essential that adequate resources be provided to ensure the timely and successful implementation of this reform program. It is recommended that Local Government be a focus of the 1998 PER.  1. INTRODUCTION 1.1 This chapter provides the context for the study and reviews the progress made in public expenditure management since the first Public Expenditure Review for Tanzania was undertaken in 1989. The report examines the nature of the problems facing the policymakers in Tanzania today and indicates the urgent need for the Government of Tanzania to realign the balance of expenditures towards other charges and the development budget; to increase the efficiency and effectiveness of its expenditures; to strengthen the overall expenditure management system; and to improve budget implementation. 1.2 The report is organized in two volumes. Volume I contains the Executive Summary and agenda for future work and the main report consisting of six chapters. Chapter 2 examines the composition of public expenditures; Chapter 3 looks at the size and affordability of Government; Chapter 4 addresses priority sectoral allocation issues; Chapter 5 addresses the overall strengthening of the public expenditure management system; and Chapter 6 outlines measures to improve budget implementation. Volume II is a Statistical Annex. 1.3 The first Public Expenditure Review (PER) for Tanzania was undertaken in 1989 by the World Bank jointly with the Government of Tanzania. As the first such report, the PER covered a wide range of issues relating to public sector management. It thoroughly documented the mismatch between the role Government was attempting and the resources available to it and the weaknesses of Tanzania's public sector management. Specifically, the report highlighted the need to redefine the role of Government in the economy; the need to avoid focusing primarily on the development budget and capital projects and make the efficiency and effectiveness of recurrent expenditures a central concern; the link between public expenditure rationalization and civil service reform; the need to revamp the heavily distorted tax structure; and, the need to strengthen aid management as part of a transparent and comprehensive approach to the management of public expenditures. The PER identified institutional changes in planning and budgeting systems that would facilitate rationalisation and better management of expenditures. A key recommendation was to strengthen the link between planning and budgeting by moving towards a system of rolling three-year plans addressing the entire budget (as opposed to the traditional 5-year fixed, capital-focused plan). The second PER in 1994, endorsed the same approach and focused on the practicalities of carrying forward the implementation of the recommendations, especially in relation to rationalization of the development budget, linking budgetary reform to a systematic overhaul of the civil service and accounting for aid. The report, in addition, focused on social sector budgetary allocations. 1.4 The findings of this PER echo those of the previous two PERs. While there has been significant progress, the fundamental issues are essentially the same. There is still a 2 Chapter I serious mismatch between resources and commitments and many of the symptoms of unsatisfactory management of public expenditures are still very visible, continuing the low efficiency and effectiveness of expenditures. This PER therefore pays particular attention to the problems of improving budget implementation and strengthening management systems, trying to understand and to address the institutional factors that have caused unsatisfactory patterns of public expenditure to persist in spite of serious efforts at reform. 1.5 The report notes the following areas where there has been significant reform: - The role of Government in the economy has been significantly redefined. Economic management is now much more market-oriented and, although slow to commence, the privatization program is well under way; - A comprehensive civil service reform program is also under way. There has been a significant retrenchment program and some reform of pay structures. Ministry-by-ministry functional reviews are still going on; - Reform of the planning and budget system has included the adoption of a three-year Rolling Plan and Forward Budget (RPFB). Currently, Government is addressing the integration of the Recurrent and Development Budgets. Parallel efforts to strengthen accounting and audit systems are now making significant headway; - The recent establishment of the Tanzania Revenue Authority has contributed to a significant strengthening of revenue collections and the outlook is for continued improvements in collection; - There have been some valuable improvements in management and service delivery at sectoral level, for example in the agriculture sector; - The present Government has demonstrated a firm commitment to fiscal discipline since the elections of late 1995, which it intends to carry forward into a more radical restructuring of public expenditures than there has been before. 1.6 The present PER recognizes that the measures implemented so far have not yet achieved the transition to a sustainable pattern of public expenditures focused on the effective and efficient delivery of core public services. This does not mean that past prescriptions were inappropriate, although the difficulties of implementation may have been underestimated. Given the interconnections between different elements of the public expenditure management system, necessary reforms must complement each other. Many of the necessary pieces are in place, and the requirement is not to abandon past efforts but to adjust and build on them. 2. ANALYSIS OF PUBLIC EXPENDITURES 2.1 This chapter analyses the Tanzanian budget in the context of the current macro- environment. It indicates that, while Tanzania's Government sector is relatively small compared to neighboring countries, the mismatch between revenues and expenditures combined with the prevailing aid climate, mean that no increase in the Government sector can be supported through the medium term. Given also the poor performance of Government and service delivery, improvement in the quantity and quality of basic Government services is essential. Within these constraints, improvement in budget performance can be realized by: * a restructuring of public expenditures-between debt service, wages, and other charges and the rationalization of the Development Budget; * improving the efficiency and effectiveness of expenditures; * and, simultaneously strengthening the budget management systems and their implementation. It will be argued that this requires change by all players in the system: at the political level, continued commitment to fiscal discipline and willingness to make hard decisions regarding the size and affordability of Government, will be necessary; at the central agencies and sectoral level there will be a need to redefine programs and policies against a deliverable hard budget constraint with the objective of strategic forward movement. How this should be done, both across and within sectors, is the subject of discussion of the following chapters. A. MACROECONOMIC FRAMEWORK 2.2 In Tanzania, while the macroeconomic situation has improved consistently over the past decade, and in 1996 in particular, it nevertheless remains fragile.' The performance of investment, savings, and exports have all improved. The savings- investment gap remains large, official exports finance only about one-half of imports and investment flows remain at low levels, leaving Tanzania dependent on foreign aid to finance much of its imports, development expenditures, and some recurrent expenditures. Continuation of the current tight fiscal and monetary policies together with continuing efforts to bring inflation down are essential through the medium-term. Furthermore, appropriate market-based incentives for raising the domestic savings rate and improving the efficiency of investment are necessary if Tanzania is to raise its trend rate of growth and reduce its dependency on foreign aid. 2.3 Table 2.1 below, which presents data on the size of Central Government operations in Tanzania in recent years, reveals the magnitude of the problem facing the Tanzanian Government. It shows the growing mismatch between revenues and For a recent evaluation of the macroeconomic performance of Tanzania, see World Bank, "Tanzania Country Assistance Strategy", May 1997. 4 Chapter 2 expenditures after FY92 and the fragile fiscal situation that resulted. Through FY92, the Government maintained a tight fiscal stance and, as a result, generated a fiscal surplus after grants. In addition, in FY92 Government also generated positive savings of 1.1 percent of GDP on its current account. In FY93, the fiscal situation deteriorated dramatically when expenditures increased by 60 percent. The resulting fiscal deficit (after grants) in FY93 was 6.4 percent of GDP, and Government dissavings were at 5.9 percent of GDP. The decrease in revenue collection lead the Government to resort to bank borrowing (from the central bank and from the commercial banks), to finance its activities, contributing both to continuing high levels of inflation and to a credit squeeze on the private sector. Subsequent measures to enhance revenue collection and reduce the deficit failed to improve the macro situation. Hence, the weak performance of the Tanzanian economy from FY93 until FY96 is attributable principally to the weak fiscal position, compounded by the rising debt burden. Return to the pre-FY93 situation proved elusive in the absence of strong political commitment to fiscal discipline, until early 1996 when the Third Phase Government demonstrated a new commitment to achieving macroeconomic stability. Table 2.1: Central Government Operations, FY86-FY97 (percent of GDP) FY86 FY92 FY93 FY94 FY95 FY96 FY97 ESAF Program Total revenue 14.9 168 12.8 14 8 14.5 15.0 15.7 Total expenditure 23.1 189 23.8 22.7 21.4 19.9 18.2 Recurrent 19.2 15.7 18.7 18.1 17.6 15.8 14.6 Development ' 3 9 3.2 5.0 4.6 3.8 4.0 3.5 Deficit before grants 8.2 -2.1 11.0 -7.9 -6.9 -4.9 -2.4 Grants .. 3.2 4.5 4.7 2.4 2.5 3.0 Deficit after grants .. 11 -64 -3.2 -4.5 -2.4 0.7 Adjustment to cash .. -0.2 0.5 -2.9 -1.3 -09 0.0 Financing 6.0 -09 5.9 6 1 5.8 3.3 -0.7 Foreign (net) 1.6 22 23 2.9 1.0 -1.4 0.5 Domestic (net) 4.4 -3 1 3.6 2.5 3.8 3.7 -0.5 Bank 3 7 -3.5 3.4 0.9 2.5 3.1 -0.5 Non bank 0.7 0 4 0.3 1.6 1.4 0.6 0.0 Other 0.0 00 0.0 0.7 0.9 1.0 -0.8 Memo Item: Government Savings3 43 1 1 -5.9 -3.3 -3 1 -0.9 1.1 32.4 19.7 31.6 30.3 27.6 22.7 10.0 1 The increase in the size of the budgeted Development Expenditures beginning in FY93 reflects Government's efforts to record more fully in the Budget aid received from donors. 2 This is the difference between the Deficit after Grants and Financing, i.e. "float (the difference between checks issued and checks cleared). 3 Defined as the difference between revenues and recurrent expenditure (on a checks issued basis) Source- IMF, EBS/96/165 2.4 Weaknesses in the fiscal situation arose in 1992, when the Government made a partial attempt to reform the tax system following the presentation of the Tax Commission's Report in December 1991. In June 1992, consistent with the recommendations of the Tax Commission, various taxes were simplified, and a number of rates were lowered in the FY93 Budget. These were accompanied by some measures aimed at widening the tax base. It was expected that these measures would yield Chapter 2 5 additional revenue that would more than offset the anticipated decline in revenue due to the tax rate reductions. The revenue enhancing measures included the elimination of exemptions on customs duties and the improvement of duty collection with the planned introduction of preshipment inspection arrangements. In the event, however, there were major shortcomings in the implementation of measures to widen the tax base, and tax exemptions, far from being eliminated, increased even further, culminating in a suspension of donor balance of payments support in late 1994. 2.5 In addition to poor implementation of the tax policy measures, another major factor contributing to Tanzania's fiscal problems was the inordinate delay in the adoption of important base-broadening measures recommended by the Tax Commission. The two major recommendations in this area were the introduction of a value-added tax (VAT) and the monetization and incorporation of all monetary and non-monetary allowances paid to formal sector employees into the income tax base.' While in the June 1992 Budget Speech, the Government announced its intention to replace the existing Sales Tax with a VAT in January 1994 and the monetization and taxation of all allowances, the implementation of these measures has suffered long delays. The introduction of VAT is now scheduled for July 1, 19983 and the monetization of the major allowances only began on July 1, 1996. 2.6 Since early 1996, the Government of Tanzania has demonstrated renewed commitment to achievement of macroeconomic stability. This has entailed a serious effort both to improve the domestic revenue effort, and a determination to keep aggregate expenditure within the limits of the program and to adjust expenditure as required by the performance of domestic revenue. 2.7 The significance of this commitment of the Third Phase Government to fiscal discipline needs to be acknowledged. Firstly, fiscal discipline at the aggregate level is the most important factor in achieving macroeconomic stability, which development experience worldwide has confirmed to be a fundamental precondition for sustained economic growth and poverty reduction. Fiscal discipline has been a weak point in earlier phases of economic reform in Tanzania-and in other countries undergoing adjustment-and the substantial efforts to control aggregate expenditure since early 1996, recent though they may be, represent the most serious attempt to date to correct this weakness. Realism and prudence in the management of the macroeconomic framework for the budget is a necessary condition for virtually any other reform in public expenditure to be addressed in the present review. In particular, as discussed in more detail below (Chapter 5), establishment of a credible macroeconomic base is a necessary first step in the move towards an effective framework for medium term expenditure 2 Until FY97, the remuneration system for civil servants consisted of a meager basic salary supplemented by a host of non-taxable monetary and non-monetary allowances. These severely eroded the tax base, especially since the Government's practice of paying non-taxable allowances to its employees was also adopted by the private sector. In FY97, most major allowances were monetized. In FY98, the major remaining allowances (housing and housing maintenance) will be monetized. This timetable is now firm and necessary preparatory work is on schedule. 6 Chapter 2 planning, improving strategic allocation within the budget and enhancing incentives for efficient resource use. 2.8 For these reasons, it is valuable to examine in some detail in the present review how the recent improvement in fiscal discipline has been implemented and the conditions necessary for this performance to be sustained, while also making progress towards effective medium- term expenditure planning. This section provides a brief assessment of the cash management system, which has been the principal instrument of restoring aggregate expenditure control. Chapter 6 below on Budget Implementation issues considers in more detail the policy context for use of this instrument and the options available for maintaining the achievements on macroeconomic stability while simultaneously improving the strategic allocation of resources and the efficiency of public expenditure. 2.9 Table 2.1 above and Figure 2.1 below summarize recent fiscal performance. The overall fiscal performance of the four years up to FY97 can be seen as a gradual and painstaking restoration of the fiscal situation as it was in FY92, correcting the rapid deterioration in fiscal performance of the single year, FY93. Domestic revenue so far in FY97 is on track to achieve the program target, still 1 percent of GDP below the level achieved in FY92. Expenditure reduction has also contributed to fiscal adjustment, with total expenditure in FY97 well on track to achieve the revised program target of just over 18 percent of GDP. This will bring the deficit (before grants) back to a similar level to FY92. Similarly, the achievement of the program target will return the Government to the same level of recurrent savings as last achieved in FY92. 2.10 The impact of tighter fiscal control in 1996 has been impressive. The fall in inflation to around 15 percent in December 1996 from 27 percent a year earlier can be predominantly attributed to the impact of restored fiscal discipline. Indeed, the twelve month change in the CPI disguises the sharpness of the correction which has been achieved. Prior to the introduction of the cash management system, inflation during the five month period October to February 1996 averaged 5 percent per month, equivalent to an annual inflation rate of 80 percent. Average inflation during the subsequent seven months from March to September 1996 averaged a decline of 0.3 percent, equivalent to an annual rate of -3 percent. 2.11 As is to be expected in an economy with relatively little financial depth, fiscal deficits have a large and rapid impact on inflation but corrective measures can have similarly swift results as well. The damage to private sector confidence from the four year episode of fiscal lapse and gradual recovery is, however, longer lasting and can perhaps be seen in lags in adjustment of Treasury Bill rates. Chapter 2 7 Figure 2.1: Government Budget: Total Expenditure, Domestic Revenue, Recurrent and Development Shares, FY90-FY97 800 1 ---- 700- Total Expenditure 600-- Domestic Revenue 500 400- Tsh. Billion 300 Recurrent Budget 200- 10 Development Budget FY90 FY91 FY92 FY93 FY94 FY95 FY96 FY97 Likely Budget Outturn Estimates Period in Fiscal Years The lag in adjustment of expectations of reduced inflation has only added to the substantial costs of borrowing. For example, adjusting for recent inflation, a 90 day Treasury bill issued in March 1996 can be seen ex post to have cost the Government approximately a per annum real rate of 35 percent, well above the nominal T bill rate of 22 percent prevailing at the time. 2.12. The achievement of aggregate expenditure control since early 1996 has resulted from the introduction of a cash management system, known in Tanzania as a "cash budget" system. In practice, this has meant that aggregate expenditure ceilings have been revised from month to month, sometimes quite significantly, in the light of the Government's position at the Central Bank at the point of determining each monthly release. On some occasions the authorities have been guided by the average revenue in the previous three months, rather than simply the latest month's data, with the more 8 Chapter 2 conservative guideline usually prevailing. The office of the Accountant General and the Bank of Tanzania liase closely on progress in the Government's overall cash position, following which the Accountant General advises the Budget Division of the Ministry of Finance on the maximum allowable volume of releases which can be made in the following month. The maximum is determined as the level of expenditure consistent with the program targets, adjusted for the cash outturn on domestic revenue. The Budget Division then advises on the allocation of funds across votes, with priority to Law and Order, the social sectors and economic infrastructure. The institutional arrangements of the cash management framework are continuing to evolve. Chapter 6 presents some suggestions for strengthening the current system. 2.13 The macroeconomic benefits of the new system are very clear. However, budget managers have long operated without a hard budget constraint, in practice, and there will need to be a period during which their expectations readjust and simultaneously, the system is fine-tuned. In the short term, there have been costs in terms of disruption to the budget. This can be illustrated with reference to performance in the first half of 1996/7 (Table 2.2 below). The figures are illustrative only, partly because there are some significant problems with the consistency and coordination of within year monitoring data available to the authorities, an issue to be discussed further below (Chapter 6 on Budget Implementation Issues). Table 2.2: Expenditure Deviations from Budget, FY97 (T Sh billion) 1st half Target Actual % of Target Total Expenditure 358.5 309.7 86.3 Recurrent Expenditure o/w 293.5 284.9 97.1 Debt Service' 99.8 106.0 106.2 Wages 103.0 100.8 97.9 Other charges 90.5 78.1 86.3 Development Expenditure2 65.0 24.8 38.2 Memo: Domestic Revenue3 261.5 286.1 109.4 Aid 54.7 31.4 57.4 Including debt amortization and clearance of arrears 2Actual development expenditures are substantially understated as donor supported projects are not properly captured in the expenditure data. Including the BOT dividend. Source: Monetary and Fiscal Affairs Department, Ministry of Finance and IMF. 2.14 Reflecting Government's commitment to expenditure control, it can be seen (Table 2.2) that total expenditure during the first half of FY974 was approximately T Sh 50bn below the budget level. The restraint shown on expenditure in the first half of In taking data for the first half of the year, there is a danger of exaggerating divergences from budget plans which might be less acute in looking at the year as a whole. Annual data are also likely to be more consistent and are subject to greater scrutiny. However, the pattern of resource availability during the year is significant for the effectiveness of expenditure and annual data may conceal sustained interruptions of activity and reduction of planning horizons, leading to a reduction in the efficiency of resource use. Chapter 2 9 FY97 arose from the concerns that domestic revenue may not meet the annual target and because foreign aid (grants and loans) for the first half of FY97 was below estimates The restraint also reflects Government's concerns as to the need for contingencies and its commitment to control inflation. 2.15 Within the reduced total expenditure, adjustment was borne by other charges and 6 development expenditure6, as priority was given to wages and debt service. Domestic arrears demands - arising from resources committed earlier, mostly in the previous year, and only currently reaching the Ministry of Finance as claims - also represented an unpredictable and priority element of expenditure (see Chapter 6 below). 2.16 It is clear that this pattern of adjustment is likely to have a more severe impact on service delivery than implied by the reduction in aggregate expenditure, owing to the adverse movement in expenditure composition. The prospects of accelerating necessary adjustment in expenditure composition, especially the balance between wage and non wage expenditure and an appropriate allocation to development spending, are discussed further below. Chapter 6 looks at the options for improving the cash budget system of within year adjustment in order to retain the benefits of macroeconomic discipline while minimizing the costs in terms of reduced efficiency of public expenditure. B. COMPOSITION OF PUBLIC EXPENDITURES 2.17 Notwithstanding the significant commitment to fiscal discipline shown by the Third Phase Government since early 1996, the Tanzanian budget remains fragile. Although relative to other SSA countries, the Government sector is not large 19 percent in FY97 compared to the Low Income Country average of 28 percente, the prior claims on total expenditures of debt service and personal emoluments, leaves little for the running of Government (Other Charges) or the Development Budget. In addressing this imbalance the Government of Tanzania faces a formidable challenge. It will necessitate difficult decisions, in particular regarding the size and affordability of Government, the allocation of funds in line with Government policy priorities and, simultaneously increasing the efficiency and effectiveness of expenditures. This section examines the composition of spending by economic categories, i.e., debt service, wages and salaries, other charges, subsidies and transfers, etc. Chapter 4 examines key sectoral expenditure issues. 2.18 The main data sources for the PER are the Budget Estimates for various years. Each year's Estimates are published in four volumes: Revenues (Volume I); Consolidated Fund Services and Ministerial Supply (Volume II); Regional Supply (Volume III); and the Development Budget (Volume IV). The most recent Budget Estimates are for FY97, supplemented by the FY98 Budget Guidelines. Forecasts for FY98-99 are from the s This reflects both an overestimate in the budget of aid inflows and also the bypassing of the budget by donor funds. 6 A large part of the development budget releases to date reflect donor funding for retrenchment. 7 Government Financial statistics, IMF, 1995. 10 Chapter 2 Fourth Rolling Plan and Forward Budget (RPFB4), which covers the years 1996/97- 1998/99. Figure 2.2 shows the budget breakdown by revenue and expenditure. 2.19 In the Tanzanian budget structure, Recurrent expenditures are broken down into three broad categories: Consolidated Fund Services (primarily debt service), Ministerial Supply, and Regional Supply, which in FY92 accounted for 27 percent, 47 percent, and 9 percent, respectively, of total expenditures. The remainder is recurrent expenditure broken down by level of Government. In Tanzania, spending is undertaken by three levels of Governments: the Central Government, Regional Governments and Local Governments. Government Recurrent expenditure (net of debt servicing) is allocated to 41 "ministerial supply votes" or central spending agencies and to 20 "Regional supply votes" or Regional Governments. The Central Government funds local Governments through the Regional Administration9. 2.20 Of more significance, however, is the breakdown of total expenditures by economic category. Table 2.3 and Figures 2.3 and 2.4 below summarizes what has become a crucial concern in the Tanzanian budget - namely the crowding out of Other Charges and Development Expenditures by Debt Servicing and Personal Emoluments. While Debt Service and Wages together accounted for a stable 46 percent of total expenditures between FY86-92, their share increased to 66 percent in FY96 percent, while the share of Other Charges fell from around 37 percent in FY92 to 16 percent in FY96 and the Development Budget share, from around 16-18 percent to 10 percent in FY96. The fall in availability of Government counterpart funds is even more marked - declining from 11 percent in FY92 to less than 1 percent. Table 2.3: Breakdown of Total Expenditures by Economic Category, FY86-97 (percentage shares) FYS6-89 FY92 FY93 FY95 FY96 FY97 Estimate Likely (ESAF) FY97 Outturn 1st Half. Recurrent Exp. o/w 77.7 83.9 81.2 88.6 89.7 81.7 92.0 Debt Service 29.0 29.0 22.7 20.2 34.3 21.4 34.2 Wages 17.4 17.5 16.9 25.1 31.6 28.1 32.6 Other Charges 31.3' 37.0' 38.9 33.3 16.3 32.3 25.2 Development Exp. olw 22.3 16.1 18.8 11.4 10.3 18.3 8.0 Local 10.6 11.2 6.8 4.6 0.9 3.9 - Total Expenditures 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Domestic Revenue as percentage of Total 60.1 81.0 47.8 75.9 92.7 81.5 92.3 expenditure I Estimated as a residual and therefore is overstated and should be reviewed as indicative only. * The first RPFB was for the years 1993/94-1995/96. 9 Currently, there are 20 urban and 115 district (rural) councils. For detailed description of the structure and functions of local governments, see "Handbook on District Level Administration in Tanzania", Semboja and 0. Therkildsen, Ministry for Regional Administration and Local Government, 1991. See "Tanzania: Local Government: A System in Distress", World Bank (1992b). Prior to FY92, the grants to local governments were included under ministerial supply and channeled through the Office of the Prime Minister (PMO). During FY92, this arrangement changed and these grants are now included under regional supply and channeled through the Regional governments. The PMO, however, still remains responsible for the budgets and accounts for local government expenditures. Chapter 2 11 FIGURE 2.2: STRUCTURE OF THE GOVERNMENT BUDGET, 1996/971 Expenditure Revenue/Financing %of GDP 21 EVmlifne (35%) Fining NetFan:ig(0.5%) (0.0)% NetDcmesdi(-0.5%) a - ConsoldatedFundls DetRedemplian(3.4%) Sebvies (63%) Debtineest(2.8%) hItDuies andl (4.1%) n____CFS (Oier) (0.1%) Graos/ Subv. (0.9%) Domesfic Sales and Domesdc ElIseDulies SCc, Cat Ila. (1.9%) Renu (3.6%) fMiimw a (15.6% Sul (92%) Frsonal 6 Emdimrd(5.1%) Icme Ta (3.6%) 3 Ox Taxes O.C? (1.3%) andlFes (2.8%) Regionl Sply (2.0%) N -tnaxes (1.5%) 0 'Overall Budget Deficit (Before grants, checks issued) is 3.0% of GDP 2 Definition as per Table 6, ESAF Report No. EBS/96/165 of October 25, 1996 3 Other Tables indicate Other Charges (O.C.) to be 4.1% of GDP, which includes technically Grants & Subventions (0.9%), Contractual & Contingent Liabilities (1.9%) and Supply for Direct Spending 1.3%) Source: 1996/97 Budget Framework and 1996/97 Government Budget Books Volumes II, m and IV Figure 2.3: Government Expenditure Components, FY90 - FY97 700.0-A z z 600.0-z 0 00.0- 400.0 - Total Expenditure E dRecurrent Budget olw: 300.0-Z' I 1Debt Service 200 0- , Wages 100 0 .. Development Budget olw: . sp Local 0.0 -- I II FY90 FY92 FY94 FY96 Likely Outturn Period in Fiscal Years Figure 2.4: Government Recurrent Expenditure Components, FY90-FY97 700.0-/4 / 600.0- 500 0- .21 Recurrent Budget o/w 300.0f Debt Service l ii*ij I)~ ages - ~ 'I 1) Li ii lI~Ih I 11111 1 11111 IhHIIY .1 Lii Other ch ages FY90 FY91 FY92 FY93 FY94 FY95 FY96 FY97 Likely Budget Outturn Estimates Period in Fiscal Years 14 Chapter 2 Recurrent Versus Development Expenditures: Issues 2.21 The previous two PERs found that the share of the Development Budget relative to total expenditure in Tanzania was high - raising concerns about the future impact of development expenditures on recurrent expenditures, which were widely recognized to be seriously underfunded, especially on other charges. With the decline in the share of Development Expenditures to around 10 percent since the last PER and, given that a significant share of Development Expenditures are actually recurrent in nature (an estimated 25 percent), Tanzania's Development Budget share can no longer be considered high. Compared with recent country averages, (viz. Kenya, 19 percent, Malawi, 26 percent and Zimbabwe, 31 percent0), Tanzania's budgeted share for development expenditures is currently relatively low. Given the constraints on Tanzania's budget through medium term, an increase in the share of Development expenditures is unlikely and raises, therefore the serious issue of the composition of the Development Budget (see Chapter 5 below on rationalization of the Development Budget) and the efficiency (low productivity) of Tanzania's investment expenditures. This is particularly important since there are not enough recurrent funds available to operate and maintain these investments once they are completed. The effect of the Development Budget on future recurrent spending depends on both the magnitude as well as the composition of the Development Budget, since some investment expenditures have a relatively high recurrent impact (especially in the social sectors), while those in the economic service areas have a relatively low impact. 2.22 Before judging the relative shares of the Recurrent and Development budgets, it should be noted that in Tanzania donor development expenditures often bypass the budgetary system. in FY92, it was estimated that about 20 percent of project-related aid was in the Budget. While the coverage of the Development budget improved after FY92, more recently as Government development disbursements have fallen, so has the reporting of donor funds. One recent audit by a donor found that only 15% of its disbursements appeared in the budget. This is a long-standing problem in need of resolution. Chapters 5 and 6 present recommendations both for incorporating donor funds in the budget and improving Governments capacity to forecast aid flows. 2.23 The significant part of the Development Budget which consists of rehabilitation, necessitated by lack of routine maintenance also provides evidence of the underfunding in Tanzania's budgets. There are no firm estimates of the extent of the underfunding of O&M expenditures in the Recurrent budget, and the consequent increase in recurrent expenditure which will be needed to ensure that past investments will be adequately maintained and operated. Nevertheless, the rehabilitation of assets which is still needed, and which is being mainly financed by donors, should send a signal to both the Tanzanian Government and to the foreign donors against continued over extension of the Development Budget, in the sense that Tanzania cannot afford the current expenditure '0 See Government Finance Statistics, IMF, 1995. Chapter 2 15 needed to realize the intended return on the investments. These considerations argue for a massive downscaling of the Development Budget to ensure improved efficiency of priority investment expenditures. While detailed project profiles are not available on the approximately 1400 development projects (and in particular, the Regional projects), based on the fact that sufficient funding has not been available in recent years even for the approximately 125 super core projects, argues for a rationalization of the development budget limited to projects that reflect Government's priorities and for which local funding is adequate. As these projects are phased out, capital expenditures in Tanzania can be brought in line with the resources available to utilize them effectively. Breakdown of Recurrent Expenditures Debt Service 2.24 Total debt service payments have risen steadily in the last three years as a proportion of Government's total recurrent expenditure, from about 29 percent in FY86- 92 to 34 percent in FY96 and an estimated 36 percent in the first half of FY97, mainly as result of growth in domestic debt. The ESAF program for FY97 indicates a fall in this proportion to 21 percent, both because of the Paris Club and lower interest rates on domestic debt service. So far in FY97, crowding out from debt service has continued in spite of a decline in domestic interest rates. This has resulted from the reduction in aggregate expenditure, which has been constrained below program levels both as a result of adjustment to lower levels of donor financing and the cautious implementation of the cash budget system, as discussed above. 2.25 Continued refinancing of domestic debt at lower rates as the impact of lower inflation is felt will in due course relieve some of the pressure on non debt expenditure, both recurrent and development. At the beginning of FY97, Tanzania's external debt stock amounted to US$7.4bn, of which about 45 percent is owed to multilateral creditors. Interest on external debt will reach a peak of US$170m in FY98, thereafter falling slightly for the next two years The recent review of Tanzania's debt sustainability" shows that while debt distress continues to plague both the fiscal and external positions, debt indicators will improve steadily through the medium term, assuming strict adherence to the ESAF and Paris Club agreements, and with future borrowing only on highly concessional terms. Ministerial Supply Expenditures 2.26 These include four economic categories: Subventions and Grants. These payments are of two kinds: internal and external grants. The internal grants consist mostly of transfers to non-commercial parastatals and includes a substantial component for their personal emoluments; " See the "Tanzania Country Assistance Strategy", May 1997, World Bank. 16 Chapter 2 the external payments consist chiefly of Tanzanian subscriptions to various international organizations. Contractual and Contingent Liabilities. These are claims on the Government arising from loan guarantees, contingencies and other special payments. Personal Emoluments. This category consists of the Governments wage bill. Other Charges. This is the "residual" left to fund the general operating costs of departments, and includes components such as traveling, office supplies, maintenance and running expenses. 2.27 Table 2.4 provides data on the breakdown of ministerial supply (MS) expenditures into these four categories. It shows that since FY92, the share of subventions and grants has halved from 16 percent to 8 percent while personal emoluments have more than doubled, crowding out other charges. Each of these four categories is discussed briefly below. Table 2.4: Economic Composition of Ministerial Supply Expenditures, 1992-97 (percentage shares) Actual Budget Budget Expenditure Estimates Estimates FY92 FY96 FY97 Subventions and Grants 15.7 8.1 8.2 Contractual and Contingent Liabilities 6.7 6.4 17.7 Net Supply 77.6 85.4 74.1 Personal Emoluments 20.6 37.9 47.2 Other Charges 57.0 47.5 26.9 Total Ministerial Supply 100.0 100.0 100.0 Source: Budget estimate books Subventions and Grants 2.28 Table 2.5 reveals the major recipients of these subventions and grants. In recent years, external grants have accounted consistently for about 5 percent of total grants. Table 2.5: Major Subventions and Grants, 1992-97 (T Sh billion) Actual Approved Budgeted Expenditure Estimates Estimates FY92 FY96 FY97 Total External Grants 1.1 1.8 1.8 Total Internal Grants 23.0 23.8 30.3 University of Dar 3.3 5.0 6.4 University of Agriculture (Sokoine) 1.0 1.7 2.3 Muhimbili Medical Center 2.8 3.6 6.0 Other Domestic 15.9 13.5 15.6 Total Grants 24.1 25.6 32.1 Source: Budget estimate Books Chapter 2 17 These represent payments for the membership of some 150 international organizations. As suggested in the 1993 Cost Cutting Commission's Report, these expenditures need to be reviewed and rationalized to the extent possible. 2.29 Domestic Grants. Of the internal grants, about 50 percent of the funds continue to go to three institutions: the University of Dar-es-Salaam, the Sokoine University of Agriculture and the Muhimbili Hospital. The balance of domestic grants is spread across about 85 non-commercial parastatals, mostly educational and training institutes of various sorts. Employment in these institutions is estimated at approximately 45,000 and their wage bill is an addition to the civil service wage ceiling. It is considered a priority for Government to undertake in FY98 a review of the non-commercial parastatals that receive grants from the budget, with a view to consolidating expenditures and funding fully those institutions that need to remain in the Government sector and provide a high return. Such a review is expected to have a clear payoff. Several of these institutions, such as those for management and accounting, could, in principle, be self-financing. Others, such as the various research institutes that provide core Government services, would continue to need Government funding. The Executive Agencies Project, supported by UK-ODA, has identified 37 agencies that, in principle, would be eligible for privatization (see Annex Table 24 for a list of the proposed agencies). Contractual and Contingent Liabilities 2.30 During the early years of this decade, contractual and contingent liabilities (C&CL) grew rapidly. This reflected the inclusion in previous years of an item for salary adjustments that was distributed amongst the various ministries during the course of the fiscal year. This is no longer, however, included in C&CL. For FY97 a significant increase in this component has been budgeted - from 6 percent of MS to 18 percent. However, this reflects simply the incorporation of the TRA budget (T Sh 15b) and the Road Fund (T Sh 25b) in this category for FY97. The balance within C&CL are currently: (i) the costs of running the Parastatal Sector Reform Commission (PSRC) and the Loans and Advances Realization Trust (LART), the agencies that were set up to oversee the privatization and liquidation of the parastatal sector, respectively; (ii) the cost of servicing the Government bonds that were used to clean up the bad debts of the banking systeml ; and, (iii) genuine contingencies. 12 Beginning in FY93, as part of its financial sector reform program, the Government took over the bad debts of the publicly-owned commercial banks (valued at T Sh 24.1 billion); provisioned for their doubtful debts (valued at T Sh 57.2 billion); and undertook to partially recapitalize the banks (at a cost of T Sh 8.1 billion). Thus, during FY93, the Government incurred a total cost of T Sh 89.4 billion for financial sector reforms, which it met by the issuance of Government bonds. These bonds have a maturity of 20 years and carry an interest rate of 11 percent for bad debts (i.e., for non-performing assets that were transferred to LART and 5 percent for doubtful debts (i.e. for provisioning of those debts that remained with the NBC). The total annual cost of servicing the bonds issued was estimated to be T Sh 6.4 billion. With the issuance of additional bonds since then, the budgetary cost of servicing the bonds is estimated at T Sh 9.2 billion per annum. 18 Chapter 2 2.31 Parastatals: The costs (direct and indirect) of the parastatal sector remain high, and parastatal reform, therefore, is key to revitalizing the Tanzanian economy. Nonetheless, from the budget standpoint any decision to take on additional liabilities from a parastatal needs to be evaluated carefully. In 1992 the Government of Tanzania indicated that in parallel with a policy of privatization of public enterprises, the remaining enterprises would be subject to a hard budget constraint, eliminating previous subsidies to the sector. As part of the present review, the implementation of the hard budget constraint was examined for a sample of 22 of the largest enterprises, including the enterprises which have previously absorbed the greatest volume of public resources. 2.32 The survey has been of value firstly in identifying the major channels through which the hard budget constraint has in practice been evaded since 1993, both within and without Government. The survey revealed a significant volume of direct transfers to parastatals, amounting to T Sh 27 billion during the period 1993-96. The major recipients of such transfers were Southern Paper Mills (T Sh 6.8bn), NUWA (T Sh 12.2bn)and Air Tanzania Corporation(T Sh 7.Obn). The remaining transfers are indirect and are shown in Table 2.6 below: Table 2.6: Analysis of Indirect Transfers (July 96 T Sh), 1993-96 Tsh million % Overdrafts from state banks 21,406 23 Utility arrears 17,404 19 Tax arrears 15,928 17 Customs duty arrears 3,833 4 Loans from state banks 2,944 3 Wage arrears 1,773 2 Dividend arrears from TPDC 15,000 16.5 Other 14,387 15.5 Total 92,676 2.33 There is some encouragement to be drawn from the pattern of subsidies over time. The volume of total subsidies has shown a significant decline from about Tshs 32bn in calendar 1993 to approximately Tshs 1 Obn for the same sample in 1996, although the data are not sufficiently robust to be confident that this trend is not in part an illusion arising from lags in data collection. 2.34 While the investigation of the sample of enterprises has revealed useful information about the type of financial support still provided to the parastatal sector, there is still considerable uncertainty about the total magnitude. Uncertainty also stems from 13 "A study of the Financial Relations between the Government of Tanzania and the Parastatal Sector, 1993-96" December 1996; study carried out for the PER by Coopers & Lybrand with LART, and funded by UK-ODA. Chapter 2 19 the likelihood of substantial additional liabilities that have been incurred on behalf of other enterprises not investigated. This is particularly evident from exploration of the contractual and contingent liabilities for the enterprises incurred before 1993, the stock of which is estimated to exceed T Sh 550bn, of which 75 per cent is related to external debt guaranteed by the Treasury. 2.35 It is clear from the results of investigating the sample of enterprises that further information is urgently needed regarding the total parastatal sector, firstly, to clarify the position on the stock of contingent liabilities (cross-checking the stock with external debt data in order to assess the extent of liabilities additional to debt already recognized and serviced by central Government); secondly, to reach agreement on the liabilities/assets set-off between the parastatal and Government; thirdly, to assess the total flows through the channels identified in the sample. 2 36 While the progress in reducing the financial drain of the parastatals since 1993 is recognized, it is recommended that the Government take immediate actions to limit further costs to the budget. These should include: * tighter discipline, scrutiny and transparency in the approval of loans, overdraft extension and other special dispensations to parastatal enterprises by NBC; * suspension of utilities services to parastatals on non-payment of bills, in accordance with normal commercial practice with regard to other debtors; * application by TRA of full collection powers for taxes due from parastatals, including normal proceedings applied for private sector tax defaulters; * More generally, greater transparency is needed in regard to the liabilities assumed by Government to date and of the inter-parastatal liabilities. Hopefully, the data validation exercise on C&CL on-going in the Accountant General's office for the Presidential Debt Task Force will provide the basis for Government to formulate and adopt a policy stance in this regard for implementation beginning FY98. 2.37 Contingencies. In the past there were two components to this expenditure: about half has been for salary adjustments to be made during the fiscal year, and the other half was for genuine contingencies such as famines, epidemics, etc. The practice of budgeting for large sums of expenditures for contingencies under the Ministry of Finance's vote, however created serious problems of lack of transparency. An additional problem with this budgeting practice was that it gave ministries an impression that there is a large pot of unallocated expenditures that they can draw upon during the course of the year, and reduced their incentive to maintain fiscal discipline. 2.38 There is also a wider issue of the soundness of holding back sums for general contingency purposes. While there is a legitimate need for having contingency funds 20 Chapter 2 since certain expenses cannot be predicted and have to be met immediately, there is also a need to limit the discretionary spending by the Executive branch of the Government. The practice of budgeting large sums for contingency detracts from transparency and from the role of the legislature. If the emergency requires expenditures greatly in excess of the contingency amount in the Budget, presumably the Parliament could be summoned to grant additional revenue and spending authority. Currently there is no framework within which contingency provisions are set. What is required is the adoption of some framework that clearly sets out the rules for how the contingency money should be allocated and which sets a ceiling on the proportion of the total budget that can be allocated to contingencies. The framework should also be clear to the legislature and accepted by it. Personal Emoluments and Other Charges 2.39 The problem of balance between personal emoluments and other charges is recognized by Government as one of the most acute in the budget. It is also a problem which has been stressed and discussed in both the 1989 and 1994 PERs. Table 2.3 above compares the share of Personal Emoluments (PE) and Other Charges of (OC) over the period FY86-FY97. It reveals that PE share increased from around 17 percent of total recurrent expenditures over the period FY86-FY92 to 32 percent in FY96. The increased share, in part, reflects the significant real increases in the minimum civil service wage in recent years, in an effort to reverse the decline in real wages that had occurred (see Chapter 3). 2.40 Simultaneously, the share of Other Charges has fallen from approximately 37 percent in FY92 to some 16 percent in FY96. The strains on OCs have been particularly severe in FY97 as a result of the mismatch between the pace of adjustment of aggregate expenditure and the delays in civil service retrenchment. The 1994 PER concluded that the degree of underfunding was even then acute for critical inputs such as textbooks and medicines, and that it would likely to lead to a further sharp deterioration in the quality of services provided by the Government. This has by now been demonstrated beyond doubt. 2.41 To reverse the real decline in OCs beginning in FY98, it is therefore critical that Government take steps to control future increases in the wage bill and to reduce its volatility-in addition to efforts to increase the efficiency and effectiveness of its expenditures. The implications of controlling the wage bill for civil service reform, both for the reduction in total employment and for pay reform, are discussed below in Chapter 3. The problem of the share of the wage bill, however, goes beyond the size of government to the underlying structure of the budget system and its implementation and this report explores some of the factors that are obstructing resolution of the problem. These issues are dealt with respectively in Chapter 5 (Strengthening Public Expenditure Management) and Chapter 6 (Budget Implementation Issues). Chapter 2 21 2. 42 Since the problem of overcommitment on the wage bill has been recognized for some years, it is helpful to explore some of the factors which obstruct avenues for correcting the problem. Firstly, the problem is widely perceived as a deficiency of allocations for other charges, to which the only proposed solution is higher allocations in future. In other words, there is insufficiently widespread understanding that the present 'cash budget' reflects the real availability of resources and is unlikely to improve greatly in the foreseeable future. Partly, this may reflect the lack of a hard budget constraint in the past and the familiarity with a situation in which budget resources appear to be scarcer than turns out to be the case, because of the avenues previously offered by supplementary expenditure and the lack of control on commitments. But it also appears to reflect an inadequate understanding of Tanzania's real fiscal circumstances, arguing the need for a more collaborative approach to budget making. 2.43 Secondly, even where budget managers recognize that aggregate resources available to their sector are unlikely to be increased, there has been no incentive to propose restructuring of the budget with a better balance between PEs and other charges since there was no mechanism by which savings offered on PEs by an agency would be available for reallocation to that agency's Other Charges. The FY98 Budget Guidelines provide for the first time such an incentive. Its impact, however, remains to be seen. 2.44 The program of O&E reviews carried out in line ministries should in principle address the issue of staffing in the context of program rationalization. However, at present the O&E reviews are proceeding without a clear budget framework within which to match programs, staffing and available resources. It is important that the next stage of the O&E review process is fully coordinated with the medium-term expenditure framework of the RPFB. This should entail establishing broad sector allocations for total expenditure (wage and non-wage) and encouraging the O&E review teams to work with sector policy staff in line ministries to establish appropriate staffing and wage bill allocations as a byproduct of determining priorities for sector policies and programs which can be fully funded within the expenditure ceilings, rather than a process of centrally driven staff reductions 2.45 In addition to controlling the growth in the wage bill to free up resources for essential components of Other Charges, there also appears to remain scope for improving the allocation of expenditures within the Other Charges category by reallocating from non-essential to essential inputs. In the education sector for example, as Chapter 4 indicates, a large proportion of the sector's budget continues to be spent on food and housing for the students and overseas studies, rather than on textbooks and school maintenance. Similar allocative issues are evident in the other sectors. The Government attempted to address allocative issues systematically when it set up the Presidential "Cost-Cutting Commission" in mid-1993 and charged it with the responsibility of identifying areas for savings across a wide range of Government activities. The Report of 22 Chapter 2 the Commission provided a wide range of proposals for cost cutting, including the reduction in the number of Ministries from 23 to 15. Reconsideration of the Commission's recommendations should be undertaken. The importance of such a review is accentuated by the changing role of Government in the economy, which would indicate prima facie at least, the need for a smaller administration. C. CONCLUSIONS 2.46 In view of the need to maintain fiscal discipline and simultaneously improve service delivery, and given that no increase in the Government sector can be supported through the medium-term, it is concluded that: * the cash budget system has played a major part in achieving aggregate expenditure control since early 1996 and, that until the fiscal balance is more robust, some form of a cash budget system should remain. Chapter 6 below suggests means of improving the current system; * notwithstanding the recent improvement in expenditure control, Tanzania's budget remains fragile. Specifically, debt service and wages account currently for over two-thirds of total expenditure, crowding out Other Charges and Development Expenditures. Control of the share of wages and debt service will be central to attempts to strengthen and restructure the budget; * the debt service burden will fall over time with declining inflation and debt relief. In the short-term, however, there is also considerable uncertainty in estimating debt service, which contributes to volatility in the residual expenditures. Efforts to strengthen debt service management systems, including increasing the transparency of parastatal liabilities in the budget framework, are therefore important; * the wage bill share has increased significantly in recent years, notwithstanding a significant retrenchment program. Measures both to control the wage bill and to reduce uncertainties in its estimation, so as to free up resources for other charges and development expenditures are central to the Government's realization of fiscal stability. Proposals are considered in detail in the following chapters; * strengthening of the budget system requires full incorporation of donor funds and an improved capability to forecast aid flows. Chapters 5 and 6 address these issues; * although other charges have born the brunt of fiscal adjustment, potential remains to reduce non-critical expenditures in all sectors and to increase the efficiency of outlays. Chapter 4 focuses on this issue. It is also recommended that expenditures on external grants be reduced and that grants to the 85 non- commercial parastatals be rationalized and consolidated. 3. THE SIZE AND AFFORDABILITY OF GOVERNMENT A. INTRODUCTION 3.1 Over the past decade there have been numerous commissions, reviews and studies of Government performance in Tanzania14 . These investigations were designed to address weaknesses in Government policy making and execution that have been manifested in inappropriate economic policies, failing social services, deteriorating infrastructure, poor civil service performance and increasing public sector corruption. The conclusion from the studies are consistent, namely: * the present scope of Government activities and the range of Government services are no longer affordable also, that they are increasingly inappropriate given the changing role of Government in the economy; * the role and functions of Government need to be redefined within Tanzania's resource envelope; * core programs are seriously under resourced in terms of non-wage operating costs; * the wage bill share is too high; * civil service wage levels are low. 3.2 It is also now increasingly accepted within Government that Tanzania's performance is constrained by weak public sector institutional capacity, including the lack of capacity to plan and manage key reforms, and to establish a professional Government administration focused on core responsibilities. Also, internationally, there is now clear evidence on the significant and positive correlations between efficiency of public institutions and economic growth." 3.3 Low pay and the long-term budgetary constraints have long resulted in declining standards of service delivery and increasing public dissatisfaction with Government services and staff behavior. The public perception of Government services is that they are increasingly slow, bureaucratic, unresponsive and unaccountable. Petty corruption is also common. For example, a recent survey found that approximately one third of people using police, courts, revenue or lands services paid 'extra' to staff' In consequence, people are less reliant on Government services, and increasingly turn to non-Government sources for services. 3.4 Government is addressing these problems through the civil service reform program initiated in 1992, with donor support. While there has been significant progress in reducing the size of the civil service, the measures taken to date have not yet resulted in a lower wage bill, payment of a minimum living wage, increased resources for non-wage costs, or improved services. In fact, despite the retrenchment of about 25 percent of the Government workforce 14 See for example, the Report on Cost Cutting Commission on the Efficiency and Productivity of Government Operations, 1994. 1 See the 1997 World Development Report, World Bank. 1 Service Delivery Survey, Corruption in the Police, Judiciary, Revenue and Lands Services, EDI 1996. 24 Chapter 3 since FY92, the wage bill share of total expenditure during this period increased from 17 percent to 32 percent in FY96. 3.5 A Government which is over-extended and underpaid is not efficient. The frontiers of Tanzania's public sector needs to be redefined so that Government focuses resources on core functions which only Government can provide, and on programs such as social services and poverty alleviation which are not being adequately provided by others. Redefining the role of Government involves both rightsizing the public service and enlarging the responsibilities of local Governments, NGO's, and the private sector 3.6 As the Government workforce gets smaller, the scope for across the board reductions in staff is also reduced. At the same time, Tanzania's current budget frame allows no scope for real pay increases, unless the size of the workforce is further reduced. There is however considerable scope for efficiency improvements within the Government administration. Resources need to be targeted more specifically than in the past in order to improve service delivery in key areas. The O&E component of the civil service reform program has made a useful start and this work should be developed and extended to other parts of the general Government sector. At the same time, a more rigorous approach to the identification and abolition of non-priority activities needs to be adopted and implemented if civil service reform is to realize its objectives. 3.7 Increases in pay levels are necessary to improve civil service productivity and performance (particularly in professional and technical areas), and to address the resulting problems of low motivation, attracting skilled staff, absenteeism, petty corruption and an aging workforce. Because of the limited resources available, pay increases need to be targeted if they are to have a greater effect on capacity building. Improving Government performance will also require a substantial increase in the skill level of the civil service. 3.8 Action also needs to be taken to increase the incentives for ministries to use resources more efficiently, as well as to be more accountable for their use of resources. Over time this will involve greater delegation of decision making to ministries and departments and also measures to ensure greater accountability for their performance. 3.9 The broad issues of civil service reform are beyond the mandate of the PER. Pay reform, however, is both critical to the success of civil service reform and given, the growth in the wage share of the budget frame and consequent crowding out of other charges and the Development Budget, central to the PER. * Specifically, the priority issue to be addressed by Government is to determine what programs, and what level of civil service and extent of pay enhancement, can be afforded within the budget framework. B. INTERNATIONAL COMPARISONS 3.10 International comparisons of the relative size and remuneration of the civil service in Tanzania are shown in Table 3.1. As with all international comparisons of civil service data, the data should be interpreted cautiously. The table indicates that the share of Government Chapter 3 25 expenditure in GDP in Tanzania (19 percent) is below the average for low income countries (28 percent)," as is the size of the Government workforce in relation to population. The share of Government expenditure to GDP and the relative size of the Government workforce in Tanzania is close to that of Uganda, which also underwent a significant downsizing program between 1989 and 1995. Significantly, however, both the ratio of the Government wage bill to GDP and to general Government expenditure is significantly higher in Tanzania than Uganda. In contrast, the size of the Government workforce and the share of the wage bill in GDP in Kenya are higher than in Tanzania, reflecting Kenya's higher levels of social service coverage and slower pace of civil service reform. Table 3.1: Relative Size and Remuneration of the Government Workforce. International Comparisons, (MRE) Population General General General General Wage Bill / (million) Govt* Government Govt. Govt. General Govt. Workforcel workforce to Exp./GDP Wage Bill/ Expenditure ('000) Population (%) GDP (%) (%) Tanzania 29 286 1.0 19.3 4.4 22.3 Uganda 19 170 0.9 20.3 2.7 13.4 Kenya 26 532 2.0 25.8 10.3 32.1 Low income2 1.8 27.8 5.6 20.3 General Government includes the civilian workforce in central and local Government and public sector education and health. It excludes employees of public institutions (e.g. Muhimbili). Low income countries: Sources: General Government Expenditure, two year averages 1983-90, in A Pragmatic Approach to Policy Analysis, 1995, IMF; Pamphlet Series No. 48. Workforce to Population calculated from a sample of low income African and Asian countries, source: EMTPM, IBRD. Sources: CSD, Government of Tanzania; Republic of Uganda, 1995/6 expenditure estimates; Kenya (1993), 1995 Government Finance Statistics, IMF 1995 World Tables, World Bank. C. CIVIL SERVICE PAY 3.11 Approximately 94 percent of the general Government workforce currently earns" less than T Sh 65,000/ month ($105), the estimated minimum living wage in Dar es Salaam in 1996." Almost one quarter of the workforce currently earns less than $60/month. 3.12 Government pay levels in real terms fell by over 50 percent between 1980 and 1985.20 Since then, real pay levels have recovered gradually. However, real wages and salaries only exceeded their 1980 levels in 1995/6", following two years of significant real increase in the Government minimum wage. 3.13 Since the 1980s, and despite real wage increases, Government pay levels have also lost ground to both the parastatal and the private sectors, and are now seriously uncompetitive. In 1996, average civil service earnings over a range of benchmark jobs were 34 percent of average 17 Chapter 2 above and 1995 Government Finance Statistics, IMF. 18 Civil service compensation, including salaries and allowances. 19 Towards an Objective minimum base pay for the Government Service, 1996, Pay Reform Unit CSD. 20 Africa Development Economic Indicators, World Bank 1996. This decline is in addition to that of the 1970s. 21 Africa Development Indicators, and Pay Reform Unit, CSD. The real wage index (1980 = 100) was 87 in 1994, and 103 in 1995. 26 Chapter 3 private sector earnings and 45 percent of parastatal earnings' Civil service earnings were 39 percent of private sector earnings at senior management level, one third of private sector earnings for middle level positions and 58 percent of private sector earnings at the lowest (unskilled) level. However pay levels of professional staff in Government service were only between 10 percent and 30 percent of equivalent private sector positions. 3.14 The relatively favorable public/private sector differentials for low paid workers reflects the pattern of recent Government pay awards in Tanzania. Government's recent pay reform objectives have been two-fold: to move towards the payment of a minimum living wage with a focus on improving the pay of the lowest paid workers; and decompression of the compensation structure, with a focus on increasing salary and grade differentials and increasing the wage bill share for groups in the upper end of the salary scale. This is the pattern followed for example in 1996.' However, in order to comply with the agreed FY97 wage bill ceiling, middle level civil servants and professional and technical staff lost wage bill share, including some staff who experienced absolute wage reductions. The effect of Government pay policy is shown in Figure 3.1. 3.15 The impact of pay policy in Tanzania to date has been a decline in the professional and technical capacity in the civil service. Many middle level staff now lack the necessary skills. At current pay levels and differentials between public and private sector pay, there is little prospect of being able to increase professional and technical employment in the Government service. The civil service has not only lost professional and technical staff but has been unable to recruit successors, resulting in an aging workforce and increasing reliance on contract-based employment (often donor funded) to carry out key management responsibilities. The increasing loss of civil service capacity and the failure to develop future managers is not sustainable and has important implications for the quality of civil service performance. 3.16 The low level of Government remuneration has also led to the widespread incidence of secondary occupations at all grades, and including amongst professional staff in the health and education sectors. Low pay also influences productivity indirectly, for example in the increasing use of travel, training and workshops as means of increasing earnings, since it takes key staff, away from their duties. 3.17 In these circumstances, improving policy making and public sector management is extremely difficult. Without targeted increases for professional and technical staff, the professionalization of the civil service will continue to be held back. Without substantive pay reform toward the minimum living wage, the morale and motivation of the civil service can be expected to decline further; absenteeism and petty corruption to increase; and the productivity increases required to support improvements in service delivery will be difficult to achieve. Thus, a medium-term pay reform strategy is a pre-requisite for improving civil service performance. 22 Civil Service Pay Comparability Survey, Coopers and Lybrand, Dar es Salaam 1996. It needs to be noted that benchmarks apply only to jobs for which equivalents can be found in the private sector; for instance, there are no comparators available for rural primary teachers. 2 The Distribution of Nominal Wage Bill Gains, FY95/6, Pay Reform Unit, CSD. Figure 3.1: Sectoral Comparisons of Pay Levels in Tanzania 800000 700000 Average of /oo00000 Private Sector, Parastatals & / 400000 300000 200000 !Civil Service 200 100000 Gross , 0 Driver Nurse Doctor Watchman Secretary Teacher Computer Analyst Economist Engineer Technician officer Accountant Legal Officer Research Personnel Internal Auditor Chief Executive Registry Assistant Personnel Officer Assistant Engineer PS to Chief Executive Purchasing/supplies TeHead of Department Source: Pay Reform Unit, CSD 28 Chapter 3 D. COMPOSITION OF THE WORKFORCE 3.18. As of end 1996, there were approximately 285,622 civilian Government employees. In FY92, the total civilian workforce was approximately 355,000. Thus by end 1996, some 70,000 civil servants had been retrenched, including an estimated 20,000 ghost workers removed from the payroll." Simultaneously, the wage bill share of total expenditures increased from 17 percent to 32 percent at end FY96. Given available data, it is not possible to track changes in the composition of the Government workforce over time. However, since the retrenchment program has mainly affected general support staff in central, regional and local Government, its impact so far has been to reduce the size of the traditional civil service in Tanzania by 40 percent. The focus of the O&E program has also been on the central Government ministries, which account now for only 17 percent of the total workforce. There is a need to extend the review of staffing levels to other parts of the Government workforce, and in particular, to teachers and health workers in local government. Table 3.2 summarizes the structure of the general Government workforce as of end 1996. Table 3.2: Tanzania: General Government Workforce, December 1996 Category No. of % Staff Workforce in Central Government Ministries (Traditional Civil Service) 49,079 17.2 Judiciary 882 0.3 Police 27,161 9.5 Prisons 10,845 3.8 Ministry of Health - Health workers 3,367 1.2 Ministry of Education -Secondary and Tertiary teachers 9,334 3.2 Total Central Government Employment 100,668 35.2 Regions - Health workers 12,200 4.3 - Others 10,311 3.6 Total Regional Government Employment 22,541 7.9 Urban and District Councils Primary teachers 106,071 37.2 Health Workers 8,809 3.1 Other local Government staff 47,533 16.6 Total Local Government Employment 162,413 56.9 Total General Civilian Government Employment' 285,622 100.00 This excludes military and national service employment and public institutions employment Source- CSD 3.19 Based on present Government policy regarding the implementation of O&E reviews, rightsizing based on efficiency criteria, military modernization and on the rationalization of regional administration, the scope for potential reductions in the FY98 workforce is shown in Box 3.1. 24By end FY97, the Government workforce is expected to fall to 270,000. Chapter 3 29 Box 3.1: Potential Reduction in FY98 Workforce 1. Defense and national service 15,000 2. Regional and District Hospitals 2,000 3. O&E Reviews CentWd Government 6,000 4. O&E reviews: loWd Government 5,000 5. Natural Attrition 3,000 TOTAL 31,000 1. Defense and national service. As part of its military modernization program, Government policy plans to reduce its military strength by 15,000 (a net reduction of 10,000). These reductions include National Service staff. Financing for this program has still to be identified. The mobilization of donor support for military demobilization will depend on agreement to a compensation package. Since military pay scales are higher than civilian pay scales, compensation costs for army personnel will be higher than those in the civil service. 2. Regional administration. The rationalization of regional administration has identified approximately 6000 administrative staff for retrenchment who will be retrenched in FY97. In addition, there are 12,200 hospital staff in regional and district hospitals. This workforce has not been subject to review, but there is believed to be significant overstaffing amongst support staff. It is assumed here that 2000 staff could be retrenched. 3. O&E Phase II reviews are expected to identify a net 6000 workers in low priority areas available for redeployment. Phase II reviews have been undertaken by Technical Task Groups appointed from the ministries concerned. However no budget or staffing ceiling targets were set for these reviews, and it is not clear what criteria have been adopted to identify core and non-core activities. A review of the findings of Phase II Reviews should be carried out by CSD to validate the manning levels and to determine the further scope for rationalization and staffing reductions. An 'alternative options' review of proposals for the 37 executive agencies is to be carried out to determine their scope for privatization and streamlining. 4. O&E reviews have not yet been carried out for the local Government workforce. Personnel controls and management systems in local Government are known to be weak compared to central Government ministries. It is assumed here that efficiency reviews at local Government level will identify the same percentage of the workforce for redeployment as in central Government (10 percent). 5. Natural attrition from the civilian Government workforce in 1996/97 is estimated at 7000 as a result of the total recruitment freeze imposed in early 1996. However, Government policy is to restore some limited and selective recruitment to the civil service in FY98. On this basis, net natural attrition has been assumed to decline to 3000 in FY98. 30 Chapter 3 3.20 There is scope for some further wage bill savings in FY98 and beyond, but it is not possible to quantify them. These include the following: the impact of decentralization on the central Government workforce. Phase II O&E reviews have not taken into account the possible impact of decentralization on the workforce in Ministries. Regional experience suggests that decentralization can lead to significant workforce reductions at the central Government level and in overall levels of employment. The implication of decentralization should be explored through a further round of O&E efficiency reviews which should be undertaken in FY98; * ministerial rationalization. Recommendations for ministerial rationalization are included in the Mramba Commission report, which recommended that the number of Government ministries be reduced from 23 at present to 15. Rationalization will lead to efficiency wage bill savings in senior administration and support staff, and in overhead costs; * Executive Agencies. Phase I and II O&E reviews have recommended 60 organizations, with a total workforce of 12,200 for executive agency status. As part of the Executive Agency program, an alternative options review will be carried out to consider whether all of the above organizations should remain within the public sector. The preparation of organizations for agency status may also, as in the case of the CMS and the TRA, identify staff savings; * transfers to public institutions. At present, approximately 85 public institutions, with an employment of around 45,000 are funded through transfers from central Government. The annual budget support provided to wages and salaries alone in public institutions is T Sh 22 billion in FY97 and T Sh 30 billion in FY98. This amount is in addition to the civilian and military wage bill and adds over 10 percent to the total. This program should be reviewed in FY98 with a view to commercializing, privatizing or closing some of the public institutions; * primary education. Primary school student: teacher ratios (36:1) are relatively low compared to other SSA countries. Gross enrollment ratios at primary level have been falling and staffing costs increasingly dominate the recurrent budget. (Personnel emoluments currently account for 91 percent of the recurrent budget for primary education). Teacher requirements are currently based on the number of classes in schools, rather than enrollments, with most urban schools overstaffed in relation to norms. Many rural schools are understaffed in relation to norms, but also have relatively small enrollments. The efficiency of resource use at primary level could be improved by increasing student: staff ratios and teacher class contact hours. This will require greater use of multi-grade teaching and multi-skilled teachers in rural areas and the introduction of a shift system for overcrowded urban schools. Also, untrained primary teachers should be replaced by trained staff. The increase in the average student: staff ratio would however result in a reduction in the size of the teaching force at any given level of student enrollment. With teachers accounting for over 37 percent of the total general Government work force, a review of teacher requirements and affordability should be undertaken in FY98. These issues are currently being pursued through the Primary Education Master Plan and are discussed further in Chapter 4; Chapter 3 31 * measures to improve the integrity of the payroll and to strengthen personnel information systems. Payroll information system and software in Tanzania is outdated and the Government Computer Center payroll reports are insufficiently detailed. There are also serious concerns as to the integrity of data and the usefulness of the system. Two payroll audits have been carried out in the last 12 months: a physical audit of the workforce by CSD in July 1996, and a financial audit by the Office of the Controller and Auditor General released March 1997. Both audits revealed a high incidence of overpayments. The CSD audit revealed 6000 ghost workers, and the OCAG audit, which was limited to a sample of 4 out of 23 ministries, and 6 RDDs identified over T Sh 314 million in irregular payments for March 1996". The solution to these problems is to strengthen personnel information and payroll systems and replace the existing central payroll. These issues are being addressed through the on-going UKODA PCIS project. However, new payroll software is unlikely to be operational before June 1998. Efforts to accelerate this project timetable should be supported; * Appointments and promotions: Finally, additional wage bill savings, and improvements in personnel management systems can be realized through changes to the present policy on civil service appointments and promotions (see Chapter 6) below. E. AFFORDABILITY 3.21 In Tanzania, Government expenditure on wages and salaries as a percentage of total expenditures increased from 17 percent FY86-89 to 32 percent in FY96, notwithstanding an almost 25 percent reduction in payroll numbers." Simultaneously, as a result of the crowding out by wages, expenditures on other charges fell from 31 percent of total expenditure to 16 percent. 3.22 Measures taken to date to contain the wage bill share, besides civil service retrenchment are: * restrictions on recruitment; * reduction in the number of allowances and benefits designed to strengthen expenditure control over payroll costs; * improved controls and audits to strengthen Government payroll integrity including the abolition of voucher payrolls and the removal of ghost workers. 3.23 Despite these measures, wage bill costs have continued to increase over time, both absolutely and as a percentage of recurrent budget. Most recently, between July and December 1996, the wage bill was inelastic to reductions in the size of the workforce. In July 1996, following the consolidation of the Government payrolls and the introduction of the new pay structure, the non-military wage bill was T Sh 13.3 billion. By December 1996, this had increased to T Sh 13.5 billion (Source CSRS ) despite the reduction in the workforce of 3,700 2s This significantly underestimates the incidence of irregular payments, because of the relative under-representation in the sample of local Government, where personnel control systems are known to be weak. 26 Mostly unskilled and therefore low paid workers. 32 Chapter 3 staff. The main reason for the increase in payroll costs was grade drift (arising from promotions) and unanticipated allowance payments. In the face of this rigidity, further retrenchments of 13,000 by the end of February were required to meet the FY97 wage bill ceiling of T Sh 199.5 billion, agreed as part of the IMF supported ESAF program. 3.24 While there is no one correct size for the civil service, affordability is key and some nominal concept of right size is useful. One such target is the civil service as a percentage of the population: Tanzania's civil service over the past three decades grew at 4.5 percent per annum while population grew at 3 percent per annum. If the civil service had expanded at the same rate as population, then its size today would have been 225,000. An alternative is, in a further round of O&E reviews with budgets covering the Government workforce, to propose a workforce that can deliver required programs and which incorporates a medium-term pay reform strategy. F. PROSPECTS FOR 1997/98 AND BEYOND 3.25 The impact of reductions in the civil service workforce during FY98 on the wage bill have been estimated using the wage model developed by the Pay Reform Unit in CSRS. For purposes of illustration, the impact of reductions in the civilian and military workforce have been developed and these are shown in Table 3.3. Table 3.3: Tanzania - Pay and Employment Reform FY98 FY99 (A) (B (C) (D) (E) Assumptions No further Reduction: Reduction: Reduction: Reduction: 4,000 reduction 15,000 civilian 15,000 civilian 15,000 civilian civilian workforce in workforce start workforce start workforce start start of year workforce of year & no of year & of year & 15,000 change in 15,000 military military military Average Civilian Workforce (1) 270,528 255,528 255,528 255,528 250,058 Pay Structure (2) unchanged unchanged unchanged fully reformed reformed over 5 years (Year 2) Inflation factor, % (3) 15 15 15 15 7.5 Wage bill (billions T Sh) (5) 226.0 226.0 226.0 310.0 248.0 Outcomes Civilians average monthly 53,258 56,369 63,781 91,213 71,741 compensation (T Sh) Real increase in civilians average -3% 3.7% 19.3% 77% 28.6% monthly compensation (4) Minimum Compensation (6) 35,770 37,895 42,959 58,652 47,948 Compensation - economist (7) 97,920 103,146 115,595 170,443 130,322 (1) Attrition in civilian workforce estimated at 2,550 employees. Assumes workers are retrenched at the start of year. (2) Current pay structure reflects situation at December 1996. Reformed pay structure assumes civilian employees are paid 75 percent of the salary of an equivalent position in the private sector. (3) Inflation in the previous financial year. (4) Real wage increase = increase in average compensation - inflation factor. Average monthly compensation-December 1996 47,498 T Sh (5) Assumes military pay is increased at the rate of inflation. (6) Compensation in grade TGOSI, Watchman. (7) Compensation in grade TGS9. Chapter 3 33 3.26 The wage model has also been used to simulate the effects of increasing the competitiveness of the Government pay structure. And, in particular, of reducing the differentials between public and private sector pay rates for professional and technical staff. For this purpose, the model assumes all Government workers are paid according to a pay structure which provides for 75 percent of current private sector pay levels for equivalent grades. The main assumptions underlying the model include a 15 percent rate of inflation in FY97 and a wage bill ceiling of T Sh 226 billion in FY98 and T Sh 248 billion in FY99. 3.27 The principal conclusions of this simulation exercise are: * Scenario A: without further reductions in the size of the Government workforce (civilian and military) during FY98, average real wage levels in the civil service will decline by 3 percent; * Scenario B: assuming a reduction of 15,000 in the size of the civilian Government workforce, but no reduction in the size of the military, average Government real wage levels would increase by 3.7 percent; * Scenario C: reducing both the civilian and the military workforces by 15,000 would enable an average 19 percent increase in Government real wage levels over FY97, allowing significant scope for pay reform; * Scenario D: introducing a competitive pay structure in FY98 is not possible within the FY98 wage bill ceiling, even with the assumed reductions in the civilian and military workforce. Within the present budget framework the new pay structure could only be introduced in stages. 3.28 Assuming a modest further reduction in the size of the workforce beyond the levels reached in FY98, an average real wage increase of at least 10 percent in FY99 would be possible (see Scenario E). This would also allow further reforms in the pay structure. The impact on competitiveness however would depend on movements in private sector remuneration levels over the same period. 3.29 The model confirms that it is not possible to reform the pay structure to increase the competitiveness of the Government pay structure within the present budget framework. Targeted improvements in pay competitiveness are however possible, assuming continuing reductions in the size of the workforce, based on efficiency criteriaP. G. CONCLUSIONS AND RECOMMENDATIONS 3.30 Improvements in the efficiency of Government administration in Tanzania depend on further changes to the structure of Government pay and employment. Further reductions in the size of the civilian and military Government workforce based on efficiency criteria, continued restrictions on recruitment, and improved payroll integrity, are required to release resources for 27For instance, to correct for their wage losses in FY97, a further simulation was prepared assuming that salaries of technical and professional grades increase to 75 per cent of the private sector level. With the reduction in civilian and military workforce, this is achievable within the wage ceiling. 34 Chapter 3 pay reform. A pay reform strategy should be adopted as part of the civil service reform program. The strategy should take into account the competitiveness of the Government pay structure and capacity building requirements in the form of targeted pay increases, especially for professional and technical groups. Annual pay reform targets should be set. Increases in personnel costs through grade drift should be reduced by changing the policy and practice on civil service promotions. 3.31 Wage bill ceilings for FY99 and beyond should be based on the RPFB and budget framework. The implications of wage bill ceilings and pay reform targets for the overall size of the Government workforce should be derived. These should be reconciled with the workforce implications from the O&E program for central ministries and local Government, and efficiency reviews of the workforce in health, education, and law and order sectors. In this way affordability criteria should determine the Government's core program and standards of service delivery. 3.32 The scope for internal mobility within the civil service workforce, and growth incentives, have declined over the last few years, reducing morale and motivation. It is important that the scope for internal mobility and career incentives within the civil service should be increased through the introduction of a promotions policy based on open recruitment, measures to strengthen the civil service commission, the introduction of an executive development program for new entrants and the introduction of an award program for civil servants demonstrating outstanding performance. 3.33 In summary, the gross over-manning of the Government workforce is past but scope for further downsizing remains without hurting service delivery. The wage bill share of Government expenditures remains high and problems remain with the wage levels. These issues present tough decisions for the Government. It is recommended that in FY98: * a medium-term pay reform policy be adopted; * O&E findings should be implemented. As proposed in the FY98 Budget Guidelines, agencies should utilize the flexibility to vire between personnel emoluments and other charges; * O&E Phase II reviews be reassessed by a strengthened efficiency review team in CSRS/CSD to ensure they have taken a sufficiently rigorous approach to elimination of non-core responsibilities; efficiency-based retrenchments should be implemented on an on-going basis; * arrangements to conduct O&E reviews of district and urban councils (including teachers and health workers) be put in place and efficiency-based retrenchments implemented; * efforts to reduce the size of the military and national service, as part of the military modernization program should be introduced as soon as possible. The case for a voluntary program should be considered; * steps be taken to continue the retrenchment of surplus regional administration staff and the transfer of staff to local authorities; Chapter 3 35 * a retrenchment fund should be established to meet the costs of retrenching surplus workers identified through the O&E review program; * a review of the 85 public institutions should be undertaken with the objective of commercializing, privatizing or closing some of them; * a review of privatization and commercialization options for organizations proposed for agency status should be considered by the Executive Agency program at an early stage; * improved payroll controls and audit trails should be introduced along with a new central Government payroll as soon as possible. In the meantime, the existing payroll should be validated through an audit program; * promotions policy in the civil service should be changed as soon as possible. Promotions should no longer be automatic but should depend on vacancies which should be advertised. Promotions should be awarded on the basis of open competition administered through the Civil Service Commission; * restrictions on civil service recruitment should be maintained. 3.34 Finally, the implementation of the above measures largely depends on the capacity within the Civil Service Reform Program Secretariat and the Civil Service Department. Despite the major accomplishments of the CSR program to date, there is evidence that the program is under- resourced. The management demands of the current program have constrained the capacity for future planning. There is a need to review future resource requirements, including staffing, in the Secretariat and Department and to establish new CSRP targets.  4. ISSUES IN STRATEGIC ALLOCATION 4.1 The recent commitment to macroeconomic stability offers a new opportunity to address issues of strategic allocation within public expenditure. The 1994 PER noted the steps which had been recently taken to establish a framework for medium-term expenditure planning in the Rolling Plan and Forward Budget (RPFB). At its inception, this instrument was intended to meet most of the principle objectives of a medium-term expenditure framework with sector expenditure programs, embracing both capital and recurrent expenditure and linking expenditure plans to sector policy options and objectives. 4.2 It is widely recognized that in practice this initiative has not produced the benefits expected from it in terms of extending the horizon of expenditure planning and addressing strategic issues of prioritization in public expenditure. A sufficient explanation for the state of the RPFB as a framework for strategic expenditure planning lies in the lack of macroeconomic discipline for most of the period since its inception. In fact, actual fiscal performance has departed from plan in the early months of the first year of the plan, if not before. 4.3 Assuming that the recent commitment to aggregate expenditure control is sustained and that the principal factor undermining the RPFB will therefore be removed, it is an appropriate time to take stock of what other weaknesses in the existing system will need to be addressed. The institutional issues in medium-term expenditure planning are discussed further in Chapter 5, Strengthening Public Expenditure Management. At this stage it will be helpful to examine some preconditions for addressing strategic allocation issues-a base in a realistic medium-term macroeconomic framework and demonstrable capacity to impose a hard budget constraint at sector level. This chapter then continues with a review of cross sectoral strategic issues which need to be addressed in medium-term planning, with particular attention to education and health. A. SECTOR ALLOCATION ISSUES 4.4 Before turning to specific sectors within public expenditure, it is helpful to examine some basic issues in budget implementation. In general, how effective is the budget process as an instrument of resource allocation, especially at the strategic level of votes/ministries/ sectors and programs? A preliminary answer to this question can be found by examining the comparison between outturn and budget at the broadest level of aggregation within the budget, looking at expenditure by vote/ministry. 4.5 Table 4.1 shows some results on this measure for the recurrent budget during the last two years, using provisional outturn data for FY96. From the perspective of assessing the effectiveness of the budget as an instrument of allocation, overexpenditure and underexpenditure are similarly important. For each vote therefore, expenditure outturn is compared with budget and the absolute deviations are summed across votes to give the total deviation reported in the third row. This figure represents the sum of the 38 Chapter 4 total of deviations in underspending Table 4.1 Budget Variation by Vote ministries (row 1) and the total excess (Percentage of total recurrent budget) in overspending ministries (row 2). FY95 FY96 4.6 The measurement of deviation Budget shortfalls 22.50 27.22 is complicated by the contingency Budget overspending 35.86 10.79 provision within the budget. Total deviation 58.36 38.01 Excluding such contingency Aggregate overspending 13.36 -16.43 provisions, equivalent figures for total deviations by vote would be 50 percent and 22 percent for FY95 and FY96 data respectively. However, as a measure of the effectiveness of the budget as an instrument of allocation, the higher figures are more appropriate, since the large contingency provision is by definition a non-allocation of resources. 4.7 While crude, the sum of absolute budget deviations as a percentage of the total budget provides a useful summary statistic on the credibility of the budget as an instrument of allocation. Firstly, it is intuitively obvious that a substantial improvement in budget discipline and control at sector level is a precondition for addressing any issues of strategic allocation. Medium-term expenditure planning should further aim to develop strategy at sub-sector or program level. The degree of deviation from budget at this lower level of aggregation has not been measured, but is necessarily at least as great as at vote level and is likely to be much higher, an expectation confirmed in investigation of a few sample ministries. 4.8 The levels of deviation evident in Tanzania are high by regional standards. A similar measure for Uganda stood at around 40 percent in the late 1980s, falling to a low of about 16 percent and averaging around 20 percent in the last three years.28 Table 4.2 illustrates some of the gainers and losers in the process of budget rearrangement during implementation, showing in rank order the average deviation from budget for selected votes over FY95 and FY96. 4.9 Any effective progress in strategic allocation at the broadest level requires vigorous action to enforce a hard budget constraint at sector and program levels of the budget. One promising development is the intention to introduce comprehensive monitoring at the commitment level. A major source of reallocation (often over successive years rather than within year) arises from line ministries incurring commitments outside budget limits, which are carried forward as unpaid bills, often to be addressed as domestic arrears in a subsequent year. At the beginning of FY97, these were estimated at T Sh 50 billion. Enforcement of spending limits at the commitment level should assist in eliminating this source of budget deviation. Secondly, greater effort will be needed to control the total volume of supplementary expenditure and to restrict reallocations at vote and program levels. 2 Similar data for neighboring countries are not readily obtainable and raise several issues of comparability where they do exist. Chapter 4 39 Table 4.2: Deviation from Budget for Selected Votes, FY95-96 (percentage) Science Tech and Higher Education 60.0 Defense (inc National Service) 56.8 Presidents Office 46.5 Education and Culture 33.4 State House 26.7 Water 26.7 Office of the Prime Minister 24.7 Tourism and Natural Resources 24.2 Industries and Trade 22.0 Vice President's Office 19.6 Health 17.5 President's Office -Civil Service 13.9 Lands Housing and Urban Development 10.4 Lands Housing and Urban Development 10.4 Foreign Affairs 10.0 Agriculture & Coops 9.7 Police 8.8 Prison service 6.8 Judiciary 5.9 Ministry of Justice -21.6 Home Affairs -21.7 Works -67.7 4.10 The effect of the high level of budget deviation on predictability of resource flows needs to be more widely recognized, especially its effect in undermining attempts to plan public expenditure and hold budget managers accountable for the sector outcomes generated from budget resources. Recognition of hard budget constraints at the sector level, in the view of the PER mission, is the key issue in strategic allocation. Government has recognized the connection between laxity in aggregate expenditure control and undesirable outcomes for price and exchange rate stability and hence private sector confidence and growth. Similar recognition of the costs of laxity at the intermediate level of budget implementation is also needed. 4.11 The sections below examine the key expenditure issues within each of the major sectors. The sectors are grouped into three broad categories: the social sectors consisting of education, health, water and other social sectors; the economic sectors consisting of agriculture, industry, and infrastructure; and public administration consisting of general administration, defense, and law and order. Reflecting Government's priorities, the focus of discussion is on the social sectors. 40 Chcper4 Figure 4.1: Composition of Government Budget by Ministry, 1996/97 (percent shares) RcarentBudget DvelopmentBudget PMO(3.2%) De!NatSer(1.9%)& Home Aff(0.6%) Econ.& Inf&&. (18.40/)ow Cond atedFunds finusty/rde05%1Water(2.5%\ Services (CFS) Works(7.6),Eneg& Mnerals(44% (35.7%) TranSpot& Conmnaicin(33%) Agricullune(2-7%) Eduadon- includes Healh(33%) PMO(CLSY% Reg%o (7.7%) ow Health (0/ol Edue. (0.0 1%) Deferns&Nat.Serv(9.6%) Water(1.6%)Works(0.12%) and Lcal Govt (1.9%) Home Affairs (0.7%1) Econ. ad nfiastr. (2.0%) Agiculture(23%) Educatio - icdes HigherEd(5.1%) Healh (2.8%) Regiore (2.9%) Finance (442%) IalGovt.(13.0%)o/w. (Incxlescrss-seorldonrfimds Education(92%1Heath chm ledtb ughtheTreasry) (2.9%/. Wa ter (03%/1 Works (02/o Admin. (04%) Finance (122/) 10 Oibe (lo2%) Othes(10.8%) 0 Includes Works (0.6%), Water (0.4%), Energy/Minerals (0 2), Transport/Comm. (0.7%) and Industry/Trade (0.3%) 2 Recurrent budget for Regions includes Health (1.4%), Education (0.2%), Water (0 3%) and Works (0 03%) Includes Police, Prisons, President's Office, Vice President's Office, Civil Service Dept., Foreign Affairs, Natural Resources and Community Development Source: 1996/97 Government Budget Books Volumes 11, III and IV Chapter 4 41 4.12 In the discussion below, a distinction is made between recurrent and development budget expenditures. Over 80 percent of the development budget is funded by donors. Given the large number of donors in Tanzania and the large aid program funded by them, and given the Government's limited capacity to coordinate aid, it would be fair to say that the aid program in Tanzania is donor-driven and to a large extent it is the recurrent budget that reflects the Government's priorities more than the development budget. Figure 4.1 below depicts the sectoral distribution of both the recurrent and development budgets for the current fiscal year. It should be recalled, however, that in recent years, the development expenditures have accounted for approximately 10 percent of total expenditures'. B. BUDGETARY PRIORITIES IN THE SOCIAL SECTORS 4.13 Reflecting the significance of the social sectors to Government's overall poverty alleviation strategy, the sectoral analysis of the PER mission focused on expenditures in health, education and, to a lesser extent, water. Throughout, the consistency of budgetary outcomes with the stated intentions of policy was examined, drawing on the overall poverty alleviation strategy and specific sector polices as outlined in the Education and Training Policy (1995), the Health Sector Reform Plan of Action (1996), the broader Social Sector Strategy (1995) and the first draft of the Basic Education Sector Master Plan (1997). This chapter begins from an overview of the trends in sector shares, in order to assess the extent to which it has been possible to implement those recommendations. Trends in social sector expenditures and their implications 4.14 Table 4.3 shows the aggregate pattern of government budgetary allocations to the Social Sectors, defined to include all ministerial, regional and district allocations for health, education and water, as well as other social sector expenditures through the Ministry of Labor, the Ministry of Community Development, Women & Children and the Regional supply vote for Community Developmentf. The aggregate level of budgetary allocations to the social sectors peaked in FY95 and has declined since as a result of the tightening of fiscal management Total allocations fell, in real terms, by 26 percent in FY96 and, although budgetary allocations are somewhat higher in FY97, the real value of allocations remains considerably lower than FY94 and FY95. These trends and their underlying causes are examined below in relation to education and health. Education and Health : Recent Trends in Sector Shares. 4.15 Table 4.4 shows the real and nominal levels of public recurrent expenditure in health and education over FY90 to FY96 and budgeted allocations for FY97. The level of recurrent expenditures is compared with overall GDP, with total recurrent expenditure and with discretionary recurrent expenditure, defined as total recurrent expenditure net of 2 As noted above, estimates of donor disbursements are unreliable. 30 A detailed assessment of expenditures on the social sectors has been undertaken by the World Bank in 1996. The data from this exercise underpins the analysis in this section. 42 Chapter 4 Table 4.3: Real Government Budgetary Allocations to the Social Sectors 1993/94 - 1996/97: Recurrent and Development Budgets (T Sh million, 1995 Prices) Real Government Allocations" 1993/94 1994/95 1995/96 1996/97 (1995 Prices) Actual Actual Rev. Est. Budget Total Social Sector Allocations 163,634 174,582 129,905 134,262 Real Annual Growth: Total Allocations 7% -26% 3% Recurrent 21% -39% 14% Development -38% 52% -22% Social Sector Recurrent 123,254 149,603 91,891 104,495 Education Share 51% 55% 62% 65% Health Share 33% 38% 32% 29% Water Share 12% 4% 3% 4% Other Share 4% 3% 2% 2% Social Sector Development 40,380 24,979 38,014 29,767 Education Share 28% 25% 37% 33% Health Share 48% 40% 37% 39% Water Share 9% 31% 21% 24% Other Share 16% 4% 5% 5% Allocations by Source of Finance: Recurrent 75% 86% 71% 78% Development 25% 14% 29% 22% 'This Table presents a mixture of actual expenditures and budgeted appropriations servicing and other statutory payments.31 It is the share of discretionary expenditure which perhaps constitutes the best indicator of the priority allocated to health and education. (Figure 4.2). 4.16 A number of observations can be made on these recurrent expenditure trends: * over the six year period from FY90 to FY96, the Government was successful in achieving a modest real rise in public expenditures on health and education; * on the basis of extrapolations from the 1988 census, it appears that this increase was sufficient to just preserve real per capita expenditures for education but not for health. However, real per student expenditures on education have declined; * apart from FY95, the social sector shares of the discretionary budget have remained reasonably constant at 11 percent for health and 21 percent for education; * although the Government responded to the recommendations of the 1994 PER with a sharp increase in the share of social sector spending in FY95, the increase was not sustainable. FY96 witnessed a striking fall in both absolute 31 In the Tanzanian budgetary system, these payments are made through the Consolidated Fund Services account, which in addition to debt servicing also covers the operating requirements of the State House. Once approved by Parliament, these allocations are protected so that no virements or reallocations can be made without additional legislation, thus there is no discretion over these items. Table 4.4: Public Recurrent Expenditures in the Education and Health Sector, 1989/90 1989/90 1990/91 1991/92 1992/93 1993/94 1994/95 FY'90-'95 1995/96 1995/97 (Actual) (Actual) (Actual) (Actual) (Actual) (Actual) Average (Rev. Est.) (Budget) Z EDUCATION Normal Recurrent Expenditure 16,576 21,880 27,587 33,055 46,782 82,058 37,990 72,024 97,902 Real Recurrent Expenditure 53,471 57,579 58,696 56,991 63,219 82,058 62,002 57,619 68,232 (Base = 1994/95) Real per capita exp. (T Sh). 2,122 2,223 2,207 2,088 2,250 2,839 2,288 1,940 2,237 % of Total Recurrent Exp. 14.5% 13.6% 13.8% 11.9% 14.4% 21.2% 15.6% 14.7% 15.5% % of Discretionary Recurrent Exp. 21.2% 19.6% 19.6% 16.7% 19.8% 29.6% 21.9% 22.7% 24.1% "%of GDP 2.2% 2.3% 2.4% 2.6% 2.9% 3.6% 2.8% 2.4% 2.7% HEALTH Nominal Recurrent Expenditure 9,009 13,154 16,410 20,381 30,663 56,968 24,431 37,255 43,784 Real Recurrent Expenditure 29,061 34,616 34,915 35,140 41,436 56,968 38,689 29,804 30,515 (Base = 1994/95) Real per capita exp. (T Sh) 1,153 1,1337 1,313 1,287 1,475 1,971 1,423 1,004 1,000 % of Total Recurrent Exp. 7.9% 8.2% 8.2% 7.3% 9.5% 14.7% 10.0% 7.6% 6.9% % of discretionary Recurrent Exp. 11.5% 11.8% 11.7% 10.3% 13.0% 20.5% 14.1% 11.8% 10.8% %ofGDP 1.2% 1.4% 1.5% 1.6% 1.9% 2.5% 1.8% 1.2% 1.2% Basic Data GDP (at market prices) 758,050 935,074 1,130,596 1,267,432 1,635,470 2,284,300 1,335,154 2,998,272 3,620,414 Total Recurrent Expenditure 113,964 161,224 199,670 278,899 323,977 386,573 244,051 488,830 631,906 Discretionary Recurrent 78,2777 111.407 140,505 197,571 236,765 277,197 173,620 316,749 406,467 Expenditure NCPI deflator 0.31 0.38 0.47 0.58 0.74 1.0 0.58 1.25 1.43 Population (Millions) 25.20 25.9 26.6 27.3 28.1 28.9 27.0 29.7 0.5 Notes and Sources Actual expenditures front the Appropriation Accounts for FY90-95 1995/96 data are approved estimates and 1996/97 budgeted appropriations from the latest Budget Books. Data include Ministerial, Regional and District Budgets, with District expenditures based upon release of flunds rather than formal expenditure returns Populations data estimated from 1988 Census with CAGR of 2 8 percent GDP1 and dellator from WB/IMF/BOT estimates. 44 Chapter 4 Figure 4.2: Health and Education Sector Shares (% of GDP) 20% 90o0 18% 16% Ote Dsrtoa 14%r & 10%OteDiceinr Expenditure 8% 6% 4%ReurnEdcto 2% Expenditure 0% levels of expenditures and sector shares, which has only slightly been reversed in budget estimates for FY97. 4.17 In terms of the likely future resource scenario, it is highly improbable that in the short to medium-term, Government will be able to do more than preserve the real per capita level of expenditure. The PER therefore emphasizes the importance for the social sectors of efficiency improvements and intrasectoral reallocations towards primary education and basic health. International Comparisons of Social Sector Outcomes and Expenditures 4.18 Comparing average recurrent expenditures on health and education over FY90-95 as a proportion of GDP with the performance of neighboring countries, Tanzania does not rank well. From Table 4.5, it can be seen that, Tanzania's relative level of expenditure on education is the lowest, although on health it is significantly higher. However, the differences in health and education outcomes are more startling: both primary and Chapter 4 45 Table 4.5: Public Expenditures on Health & Education compared with outcomes in Tanzania and Neighboring Countries (1993) Country Primary Secondary Public Infant Public School Gross School Gross Expenditure Mortality Expenditure Enrollment Enrollment on Education Rate on Health Rate Rate (% of GDP) (deaths per (% of GDP) (%) (%) 1,000 live births) Tanzania 71 6 2.8 84 1.8 Kenya 92 28 6.8 59 1.3 Uganda 99 17 2.9 122 0.9 Ethiopia 27 12 4.9 120 1.9 Source: NB. Tanzania expenditure data represent the average expenditure over FY90 - FY95 as a percentage of GDP. For other countries, the 1993 data are presented. Data on enrollments & IMRs from 1996 World Development Report. secondary enrollments are considerably higher in both Uganda and Kenya and Infant Mortality Rates, lower in Kenya. 4.19 Looking at health and education outcomes in historical perspective, the data for Tanzania are worrying. Having reached a peak Gross Enrollment Rate (GER) into primary education of 98 percent in 1980, rates declined sharply over the 1980s and are only now showing a modest rise.32 Despite the fast expansion of both public and, especially, private secondary schools since 1986, secondary enrollment rates have moved only slightly faster than population growth and the present rate of 5 percent is one of the lowest in Sub Saharan Africa. With regard to health indicators, despite the very significant gains made since Independence, Infant Mortality Rates and Life Expectancy have remained largely unchanged since 1985. Public Expenditure Issues In The Education Sector 4.20 The 1994 PER undertook a thorough analysis of education sector expenditures; this has been deepened through the Social Sector Review which benefited from the new information on household expenditures provided in the 1993/94 Human Resource Development Survey (HRDS), both of which were supported by the World Bank. More recently, further analysis has been undertaken by the Local Government Support Unit of the Prime Minister's Office on the question of teacher utilization and by UDSM and MOSTHE on unit costs in higher education. 4.21 Government's priorities in the education sector reflect significant rethinking of the role of Government in the sector. It is in primary and lower secondary education that there appear to be the greatest externalities. These emerge from the spill-over effects of 32 The MoEC record a GER of 77.6 percent in 1995 as compared with 74.4 percent in 1991. United Republic of Tanzania, Basic Statistics in Education 1991 - 1995, Ministry of Education & Culture, Dar es Salaam, June 1996. As will be noted by comparison with the World Development Report data shown in Table 4.5, this publication tends to give relatively high estimates of enrollments, probably because they are based on enrollments recorded at the beginning of the academic year rather than the end. 46 Chapter 4 literacy and numeracy, from improvements in health due to understanding of hygiene and sanitation issues and from the declining levels of fertility associated with higher levels of girls' education. By contrast, the returns from university level training, although also substantial, are almost fully captured within the higher incomes of graduates. Recent estimates for Tanzania based on the 1993 Labor Force Survey reaffirm these findings (THET, 1997, Mason & Khandler 1996). The Government of Tanzania rightly places these as the priority areas for public investment. Overview of Public Expenditures 4.22 In FY96, the Government of Tanzania spent an estimated T Sh 72 billion, equivalent to T Sh 2,425 per capita, on teachers, school materials and other recurrent expenses. In addition, some T Sh 12 billion (T Sh 409 per capita) were budgeted to be spent through the Development Budget, the bulk of it financed through aid grants and concessional loans. This sub-section examines the trends in education spending and the allocation of expenditures by education program and by input. 4.23 The coverage includes all expenditures undertaken by the Ministries of Education & Culture (MoEC) and Science, Technology & Higher Education (MSTHE) as well as education expenditures at the district and regional levels. It does not include all of the training activities carried out within Government, which are spread across some 18 ministries. In omitting these "other" training expenditures from the analysis, there is therefore an understatement of the total amount spent within the sector. 4.24 From FY90 to FY95, relatively steady real growth in the aggregate recurrent expenditure within the sector was achieved, permitting modest growth in real per capita Table 4.6: Recurrent and Development Budget Allocations by Education Program, 1995/96 and 1996/97 (T Sh Millions) EDUCATION 1995/96 1996/97 Recurrent Devpt. Devpt. Total Recurrent Devpt. Devpt. Total (Rev. Est.) (GOT'/ Donor Allocation (Budget) (Donor) (Donor)" Allocation Total (Nominal) 72,024 3,094 9,056 84,174 97,902 3,478 11,467 112,847 Total (1995 prices) 57,619 2,475 7,245 67,339 68,463 2,432 8,019 78,914 % of Total allocation 86% 4% 11% 100% 87% 3% 10% 100% Primary Education 59% 46% 46% 57% 57% 41% 41% 55% Total Secondary 9%% 4% 4% 8% 8% 2% 2% 7% of which Secondary Technical 2% 4% 4% 2% 2% 2% 2% 2% Secondary 3% 0% 0% 3% 3% 0% 0% 3% Commercial Secondary Ag & HEc 4% 0% 0% 3% 3% 0% 0% 3% Teacher Training 2% 9% 9% 2% 2% 6% 6% 3% Higher & Technical 21% 21% 21% 21% 25% 45% 45% 28% Other 2 9% 19% 19% 10% 8% 5% 5% 8% Notes and Sources 1" The breakdown of the Devpt. Budget was not available and so was assumed to be the same as for Donor funds given that a large part of he Devpt Budget comprises counterpart funds for Donor-financed projects. V'Includes Adult Education, Ministry & Regional Administration and other Education, Science and Culture. Recurrent Revised estimates and budgeted appropriations from Ministerial and Regional Supply Votes, MoF. Development Budget data from Planning Commission; Donor data provided directly by Donor agencies. Chapter 4 47 expenditures over the period. Table 4.6 presents the aggregate picture of allocations to the sector covering both the Recurrent and Development Budgets. However, data from FY96 suggests that it has not been possible to sustain this trend. Budgeted allocations in FY97 represent an increase in real terms over the approved estimates of the previous year but are still below the real average achieved over the five year period to FY95. Moreover, the implementation of the cash budget system has meant that actual releases of funds are below budget levels in the present financial year. 4.25 Expenditure by Program. Table 4.7 shows percentage allocations by program. Some 57 percent of recurrent expenditures are allocated to primary education, disbursed almost exclusively through the local government structure. The data indicate that the relative share of primary education has increased since the early 1990s . The final out- turn of expenditure is likely to differ somewhat from this data and, in the past, it has been secondary school and teacher education expenditure that has benefited from reallocations rather than primary education, due to the failure of the secondary institutions to meet the tight targets for cost sharing and efficiency gains in catering and boarding expenditures set by the Treasury. Nevertheless, a relative reallocation towards primary education has occurred, which is in keeping with the defined priorities and the new role of government. 4.26 The expansion of the primary education share appears to have occurred at the expense of the secondary school allocations which in FY97 will comprise only 8 percent Table 4.7: Education Recurrent Expenditures and Budgeted Allocations by Sub-Sector, 1992/93-1996/97 (T Sh millions) Educations 1992/93 1993/94 1994/95 1995/96 1996/97 Actual Actual Rev. Est. Estimates. Budget Nominal Recurrent Expenditure 33,055 46,782 82,058 72,024 97,902 Real Recurrent (1995 prices) 56,991 63,219 82,058 57,619 68,463 %of Discretionary Recurrent 16.7% 19.8% 29.6% 22.7% 24.1% % of GDP 2.6% 2.9% 3.6% 2.4% 2.7% Primary Education 51% 51% 58% 59% 57% Total Secondary 15% 14% 10% 9% 8% Of which Secondary Technical 3% 3% 3% 2% 2% Secondary Commercial 6% 5% 3% 3% 3% Secondary Ag & H.Ec 6% 6% 4% 4% 3% Teacher Training 5% 4% 3% 2% 2% Higher & Technical 20% 21% 23% 21% 25% Other 9% 9% 7% 9% 8% Notes and Sources: Other includes Inspection, Adult Education, Ministerial & Regional administration. Actual expenditures, approved estimates and budgeted appropriations from Budget books (Ministerial and Regional Supply), Ministry of Finance. 48 Chapter 4 Table 4.8 Average Public Recurrent Expenditure Per Student by Educational Level, 1992/93-1996/97 Education Level 1991/93 1993/94 1994/95 1995/96 1996/97 (Actual (Actual) (Actual) Estimates Budget Primary Ed. (Nominal) 4,806 6,391 12,547 10,973 13,450 Primary Ed. ('95 Prices) 8,287 8,637 12,547 8,779 9,406 Secondary (Nominal) 63,034 78,944 98,343 70,408 80,188 Secondary ('95 Prices) 108,676 106,681 98,343 56,343 56,075 Ration to Primary 13 12 8 6 6 Teacher Training (Nominal) 117,625 118256 151,473 87,898 112,622 Teacher Training ('95 Prices) 202,802 159,805 151,473 70,319 78,756 Ratio to Primary 24 19 12 8 8 UDSM (Nominal) 1,089,476 1,043,888 967,211 1,176,104 UDSM ('95 prices) 1,878,406 1,410,660 967,211 940,883 Ratio to Primary 227 163 77 107 Enrollment: (Projections) Primary Schools 3,507,384 3,732943 3,793,201 3,872,473 4,148,899 Secondary Schools (Govt.) 78,660 82,964 83,441 92,066 97,673 Secondary Schools (Private) 97,116 97,935 102,805 104,309 na Teacher Training Colleges 14,051 15,824 16,252 16,388 17,386 UDSM (FTEs) 3,025 2,937 3,112 3,877 na Notes and Sources Actual expenditures, revised estimates and budgeted appropriations from budget books Ministrial and Regional Supply; Enrolment data 1993-1995 from Basic Stastics in Education, MoEC, June 1996; projections for 1996 are PER mission estimates. University data from UDSM, Facts and figures on Institutional Transformation Program, October 1996. Figure 4.3: Average Public Expenditure per Pupil/Student Education Level,1992193 - 1995/96 1,200,000 1,000,000 800,000 1 600,000 400,000 200,000 I - r] 1992193 Actual S.*1993/94 Actual . E3m 1994/95 Actual u Do M g 1995/96 Estimates Chapter 4 49 of the recurrent education budget. Simultaneously relative allocations for Higher & Technical have risen significantly and given the needs of primary education, raises concern" Central and regional administrative expenditures (included within "Other") have remained constant. 4.27 Per student expenditures differ widely by level of education as Table 4.8 and figure 4.3 illustrates. Per student expenditures in teacher training and in university are very high. These reflect inefficiencies and offer potential for cost-savings. However, their ratio to primary is in part distorted by the unusually low level of expenditures per primary pupil. This is influenced by the high level of parental contributions to primary schooling, where 40 percent or more of the unit costs are borne directly by parents. Notwithstanding, the counter-balancing influence of these contributions, overall unit expenditures on primary education are low and translate into shortages in text books and school materials-most especially for the poorer families who are unable to buy materials. 4.28 Input mix. The nominal expansion of primary education budgets, however, has not arrested the decline in the relative share of non-personnel expenditures (Table 4.9). In particular, the FY97 budgeted allocation for school materials represents a real decline of nearly 20 percent (T Sh 351 per pupil to T Sh 282 in 1995 prices) from FY96 and an even more substantial decline from FY95. Table 4.9: Shares of Recurrent Allocations by Input for Primary, Secondary and Teachers Education, 1994/95-1996/97 (percent) Input type PUBLIC PRIMARY PUBLIC SECONDARY TEACHER EDUCATION 1994/95 1995/96 1996/97 1994/95 1995/96 1996/97 1994/95 1995/96 1996/97 PERSONNEL 71 86 91 37 52 59 41 53 52 OTHER CHARGES 28 15 8 63 48 42 59 47 48 Of which: Travel & Visits 0 0 0 2 1 1 3 2 2 Operations/Maintenance 0 0 0 4 2 5 4 2 5 Pupil Transport 0 0 0 0 0 0 1 0 0 Student Board/Welfare I 1 1 48 33 26 40 33 35 School Materials 7 4 3 6 8 7 2 2 I Exam Expenses 19 2 3 2 3 2 2 4 2 Other I 8 1 1 1 1 1 0 1 Notes and Sources: Table presents a mix of budgetary and expenditure data and should be interpreted with caution. All public primary data are derived from budgetary allocations in the Local Government Budget Books. For Secondary TTCs, 1994/95 data are revised estimates, 1995/96 and 1996197 budgeted allocations. Proposals for Budgetary Restructuring for Education and Health 4.29 It is proposed that budget restructuring proposals for education and health be based on the following underlying principles: 3 For instance, in FY97 Tshs 4.6 billion is allocated to study abroad; this is equivalent to 8 percent of the primary education allocation. 50 Chapter 4 * Budget resources should be made available through a medium-term expenditure program (i.e. the RPFB) involving specified minimum budget allocations and corresponding performance targets. In the case of education, these targets should relate to public enrollments, average per pupil expenditures, and increased pupil: teacher ratios, broken down by broad program areas-primary, secondary, teacher training, technical and higher education; * The focus of management attention at the sector ministry level should be on achieving efficiency improvements within the given medium-term expenditure ceiling; * Donor resources should be made available in a manner consistent with the defined medium-term targets for budget allocations through formal sector development programs. Recommended Budgetary Framework 4.30 Specifically it is proposed for education that: * As a minimum expenditure target, the real per capita expenditures for the education sector be maintained at the FY97 level. Any additionality of funds to the budget should be targeted first to raising per student primary allocations; * To ensure that public enrollments continue to rise in primary, secondary and tertiary education, pupil: teacher ratios will need to be targeted to increase at every level of the educational pyramid, but most especially in secondary and tertiary education; * Increased efficiency in the use of teaching staff and physical infrastructure, combined with further cost recovery on catering and boarding costs, should drive down average per pupil expenditures at the secondary and tertiary levels, thus permitting enrollments to expand within the limits of existing facilities and enabling a resource shift in favor of primary education where quality improvements require increases in average per pupil expenditures. Public Expenditure Issues In the Health Sector 4.31 Over the 1970s and 1980s, Tanzania was consistently able to achieve health indicators that were above the average for its per capita income level.34 Indeed, it continues to have infant mortality rates which are lower than the Sub Saharan average. (See Table 4.10) Yet over the past ten years, the rate of improvement in health indicators has declined: infant mortality rates fell from 250 per 1,000 live births at Independence to 137 in 1975, slowing to 115 by 1985 and now standing at 84 per 1,000 live births. 3 This is well reported in the successive issues of UNDP's Human Development Reports, for example. Chapter 4 51 Similarly average life expectancy rose from 33 years in 1961 to 44 in 1978 and 50 in 1988 but in 1992 remained at 51." 4.32 The level of recurrent health funding has Table 4.10 Infant Mortality remained broadly constant in recent years, per 1,000 Live Births 1992 recording a modest aggregate increase in real Country IMR terms over 1989/90 to 1995/96 and steady per Ethiopia 120 capita increases up to 1994/95. Given these Kenya 59 financing trends, it needs to be asked ask why, Tanaia 84 over the past ten years there has been stagnation Uganda 122 in health indicators during a period when the Zambia 108 pattern of disease has not changed radically and a reduction in poverty is generally acknowledged to Source World Development Report, 1996. have been achieved. 4.33 The 1994 PER concluded that over one third of health expenditures at the time could be considered non-essential. Since than, there has been little adjustment in this underlying situation, largely due to institutional and administrative constraints with the result that the expenditure structure continues to be at variance with stated government objectives. To address these concerns, the Ministry of Health launched the Health Sector Reform Program in 1993 that has lead to the introduction in 1996 of the Health Sector Reform Plan of Action. The discussion in this section provides an overview of the reform program and an assessment of its usefulness in addressing the fundamental difficulties of restructuring health expenditures. Essentially, there are three prongs to the Government's strategy to reform public health services so as to make them more effective, efficient and sustainable: * Measures to raise resources available to the Health Sector - increase Government funding to health to 14 percent of the (discretionary) budget 16 - procure additional resources through cost-sharing and application of user charges backed up by insurance mechanisms. c Measures to reduce Unit Costs of service provision (without prejudice to quality or coverage) - streamline the organizational and management structures for delivery of primary services through the establishment of District Heath Management Boards; - integrate health service delivery systems at all levels, adopting a multi- sectoral approach in the implementation of community-based health care - rationalize health personnel through a thorough function analysis and staff redeployment process; G Proposals for Health Sector reform. MoH, December 1994 and Health statistics Abstract 1996, MoH. 36 The Plan ofAction does not in fact specify whether the 14 percent target refers to the total budget or the discretionary recurrent budget. The latter is the only plausible target however. 52 Chapter 4 - rationalize health training through the elimination of certain cadres of training and through the closure or merger of institutions where possible; - restrict the practice of referring patients for overseas treatment by improving capacity utilization in the existing national referral hospitals and by tight control of government subsidies for overseas treatment. * Measures to reallocate resources more effectively: - redefine the role of Government to focus on policy formulation and regulation; public health interventions, including immunization, hygiene and sanitation, health education, MCH services, treatment of common ailments in primary health care facilities; training of key cadres of health workers not trained by other agencies; clinical service provision or back- up to areas not adequately covered, with emphasis on the vulnerable, disadvantaged and the poor. - reallocate resources to reflect the revised role of Government, in particular to reduce direct expenditure on hospital care. Overview of Public Expenditures in Health 4.34 In FY96, the Government of Tanzania spent an estimated T Sh 37.3 billion, or approximately T Sh 1,255 per capita, on personnel, drugs and the other recurrent costs of the public health care system. A further T Sh 435 per capita was budgeted to be spent through the Development Budget, largely financed by donors. This sub-section examines trends in spending on health care and the allocation of expenditures by facility level and by input. 4.35 Real per capita recurrent expenditures on health over FY90 to FY95 improved. However, in FY96 and FY97 per capita health allocations have plummeted and given the medium-term budget outlook, maintaining the FY97 level is perhaps the best that can be expected. 4.36 Expenditure by program. Budgetary allocations to the Health sector are channeled primarily through three sources: the Ministry of Health, whose budget includes all public health training and the six national referral hospitals (4 general and 2 specialist) in addition to central administration and policy; the 20 Regional Administrations who hold the budgets for the 17 regional referral hospitals and the 70 government district hospitals; and the District Councils who manage primary level facilities-the 302 health centers and 3,500 rural dispensaries which provide services at village level.37 4.37 Despite the objectives of the Health Sector Reform Program, health expenditures continue to be dominated by curative services and in particular by hospital expenditures at the district, regional and national levels. Over the past five years, these have accounted for 80 percent of recurrent expenditures. Fundamentally, the Tanzanian health care system serves primarily to treat rather than to prevent disease. However, the relative significance of curative services has declined over FY93 - FY96 and there has been a " Health Statistics Abstract 1996, MoH, DSM. Chapter 4 53 Table 4.11 Health Recurrent Expenditure and Budgeted Allocations by Sub-Sector, 1992/93-1996/97 (T Sh million) Health 1992/93 1993194 1994/95 1995/96 1996/97 (Actual) (Actual) (Actual) Estimate Budget s Nominal Recurrent Expenditure 20,361 30,663 56,968 37,255 43,784 Real Recurrent (1995 prices) 36,140 41,436 56,968 29,804 30,618 % Discretionary Recurrent Expend. 10.3% 13.0% 20,.6% 11.8% 10.8% % of GDP) 1.6% 1.9% 2.5% 1.2% 1.2% National Referral Hospitals 33% 21% 27% 17% 15% Regional & District Hospitals 34% 33% 32% 34% 34% Health Centers & Dispensaries 23% 35% 31% 33% 34% Sub-Total Curative ('95 Prices) 90% 89% 90% 84% 83% Preventive Services 4% 4% 7% 10% 13% Training & Administration 7% 7% 4% 6% 4% Notes and Sources *Includes Health Training, Ministerial Administration and Other Actual expenditures, revised estimates and budgeted appropriations from budget books Ministerial and Regional Supply MoF. Analysis includes health expenditures under Ministry of Labor and Ministry of Community Development as well as MoH. significant reduction in the share received by the national referral hospitals, to the benefit of the health centers and dispensaries. Over the same period, the share allocated to preventive services more than doubled, albeit from a low base. (Table 4.11) It should also be noted that several of the services provided at the dispensary and health center level (MCH patrols, monitoring of immunization, etc), are in fact preventive rather than curative. Table 4.12: Shares of Recurrent Allocations by Input and Facility Level for Curative Services 1994/95-1996/97 Health Curative National Referral Regional & District Health Centers and Services Hospitals Hospitals Dispensaries 1994/95 1995/96 1996/97 1994/95 1995/96 1996/97 1994/95 1995/96 1996/97 Nominal Total (Tsh. MM) 15,381 6,333 6,568 18,230 12,667 14,887 17,660 12,294 14,887 Total (1995 Prices) 15,381 5,067 4,593 18,230 10,133 10,410 17,660 9,835 10,410 PERSONNEL 36% 0% 60% 25% 41% 53% 67% 43% 60% OTHER CHARGES 64% 100% 42% 75% 59% 47% 33% 57% 40% of which Travel & visits 0% 0% 0% 1% 1% 1% 0% 0% 0% Operations/Maintenance 10% 0% 4% 2% 3% 4% 1% 14% 2% Training 0% 0% 0% 0% 0% 0% 0% 0% 0% Overseas treatment 17% 11% 9% 0% 0% 0% 0% 0% 0% Food 4% 0% 2% 3% 2% 2% 1% 1% 2% Materials & Supplies 1% 0% 1% 4% 2% 2% 2% 2% 2% Internal Subventions 14% 47% 11% 42% 19% 22% 0% 0% 0% Other 6% 42% 1% 9% 22% 3% 3% 20% 3% Notes and Sources: Table presents a mix of budgetary and expenditure data and should be interpreted with caution. Data on HCs/Dispensaries show budgetary allocations from the Local Government Budget Books. For hospitals, 1994/95 data are revised estimates, 1995/96 and 1997/98 budgeted allocations Internal Subventions include some allocations for Personal Emoluments which could not be fully separated. 1996/97 breakdown for the referral hospitals was derived from the hospitals' original budget submissions. 54 Chapter 4 Table 4.13 Recurrent and Development Budget Allocations by Health Sector Program 1995/96 and 1996/97 (T Sh Millions; Government & Donor Funding) 1995/96 1996/97 Health Recurrent Devpt. Devpt. Total Recurrent Devpt. Devpt. Total (Rev. Est.) (GoT) (Donor) Allocation (Budget) (GoT) (Donor) Allocation TOTAL(Nominal) 37,255 2,234 10,686 50,175 43,784 3,812 12,841 60,437 TOTAL (1995 Prices) 29,804 1,787 8,549 40,140 30,618 2,666 8,980 42,264 % of Total allocation 74.3% 45% 21.3% 100.0% 72.4% 6.3% 21.2% 100.0% Referral Hospitals 17% 33% 0% 14% 15% 3% 0% 11% Reg./District Hospitals 34% 10% 6% 27% 34% 24% 6% 27% Dispensaries/HCs 33% 33% 41% 35% 34% 48% 41% 36% Sub-Total Curative 84% 76% 47% 76% 83% 75% 47% 75% Preventive Services 10% 10% 53% 19% 13% 17% 53% 22% Training & administration* 6% 14% 0% 5% 4% 8% 0% 3% 100% 100% 100% 100% 100% 100% 100% 100% Notes and Sources: * Includes Health Training, Ministerial Administration and Other Recurrent Revised estimates and budgeted appropriations from Ministerial and Regional Supply Votes, Development Budget data from Planning Commission; Donor data provided directly by Donor agencies. Analysis includes health expenditures under MoL and MCDWC. 4.38 Input mix. Table 4.12 shows budgeted allocations by input and facility level for 38 curative services . The use of internal subventions for the financing of the main national referral hospitals and several other institutions complicates the analysis of the input mix. Although the institutions are required to submit a fully detailed set of estimates, this is rarely financed in full and the item by item changes which are later made to conform with lower levels of financing are not always forwarded to the Treasury, whilst formal expenditure returns are consistently submitted late if at all. Nevertheless, the predominance of Personal Emoluments at every level of the curative structure is clear and is almost certainly understated in the data presented. 4.39 Application of Cost Sharing Systems. The policy to introduce formal cost sharing through user fees for services provided by public health facilities was endorsed by the Cabinet in June 1993 and approved by the National Assembly in August of the same year. An evaluation of the implementation of cost sharing was undertaken in early 1996 by the Ministry's cost-sharing implementation team.39 Reported collections grew steadily over FY93 and FY95 but fell short of targets in each section with the exception of the national referral hospitals. According to the monthly flash reports, user fee collections in government hospitals in FY95 amounted to T Sh 537 million, or 70 percent of projected collections. Preliminary data for FY96 indicate a lower collection rate. Projected collections amounted to no more than 2-3 percent of total recurrent budget or 6- 10 percent of non-wage recurrent budget for FY97. However, data collected at the 3s The corresponding breakdown of preventive, training and administration recurrent expenditures by input has not been tabulated because it is of considerably lesser interest given that recurrent allocations in these areas are dwarfed by the Development Budget allocations. 3 MoH, Implementation of Health Services User Fees in Tanzania: An evaluation of progress and potential impact, DSM, August 1996. Chapter 4 55 facility level through the survey suggests a much higher level of collections, where they comprised 34 percent of total non-salary recurrent expenditure and thus a much higher proportion than originally projected. 4.40 Despite the decline in user charges in FY96, there is much that is positive. There is general acceptance and understanding of the charges and they have boosted sector finances - significant achievements, given their radical departure from past practices. However, problems are emerging, largely of a sequencing nature. The agenda for explicit cost recovery as a step towards withdrawing Government subsidies from tertiary services in favor of preventive services as envisaged in the Reform plan is a complex one and the choices are difficult, and only made harsher by the budget environment. It is recommended that: * a standard reporting system for user charges be introduced effective the beginning of FY98 to ensure that Government effectively monitors funds collected and their subsequent use; * a tracking study of public expenditure in primary education and health be undertaken for recent years in a sample of districts in FY98 with a view to establish an effective tracking system beginning in FY99. Proposals for Budgetary Restructuring 4.41 There have been positive actions to initiate the Health Sector Reform Plan of Action, including Budgetary reallocations in favor of preventive and primary level services; measures to reduce subsidies to overseas treatment; and steps to extend and deepen cost sharing arrangements. However, these budgetary actions have yet to be supported by the more substantive changes in policies and expenditure structures which would permit sustainable improvements in efficiency and effectiveness and a permanent re-allocation of resources consistent with the newly defined role of Government in the health sector. 4.42 The sector remains over-stretched with public resources continuing to be spread over too many activities to the detriment of operational effectiveness. It has proven difficult for the Ministry to address this fundamental problem or to correct the large-scale inefficiencies in utilization of staff and physical capacity which persist beneath the general pre-occupation with "underfunding". There are institutional reasons which mitigate against these types of changes; yet, if the Ministry are to accelerate implementation of the Health Sector Reform Program, which is their clear objective, then more detailed attention must be given to the underlying resource issues and to the institutional and administrative structures which hinder the achievement of efficiency. 4.43 From the budgetary perspective, actions that might be taken to facilitate these changes include measures to provide a more stable medium-term financing framework; to facilitate decentralization of responsibility for district resources; and to promote managerial autonomy within the national referral hospitals. Specifically: 56 Chapter 4 * As a minimum expenditure target, real per capita expenditure on health be maintained at the FY97 level; * The sector strategy should be completed in FY97 and fully integrated with the RPFB. It should include measures to restructure expenditures and raise efficiency, detailing the specific organizational, administrative and legal changes entailed and provide performance targets, including for infant mortality and life expectancy; * Complete the preparation for the establishment of District Health Management Boards in FY99; * Review the existing plan for Human Resource Development with a view to accelerating and deepening the proposed measures for efficiency improvement; * Develop a coherent framework for the reduction in the relative share of Government financing devoted to hospitals, based upon a rationalization of district and regional hospitals, a reform of the managerial structures for all hospitals, beginning with the National referral hospitals and the introduction of a fee-for-service structure, supported by carefully targeted subsidies and insurance mechanisms; * Proceed with the development of the proposed Health Insurance Scheme for civil servants with a view to its introduction from FY99. Other Social Sector Spending 4.44 In addition to spending on education, health and water, the government has allocated T Sh 3 billion (supplemented by an estimated T Sh 7 billion of donor funds)40 for other social sector spending in FY97. These expenditures are channeled through the Ministry of Labor and Youth Development (MOL) and the Ministry of Community Development, Women's affairs and Children (MOCD). The two ministries are relatively small and funded largely by donors funds. Funds are to be spent on various training institutions and programs, labor issues, youth recreation and development, "folk development colleges" and social welfare programs. Approximately two-thirds of the FY97 budgets of the two ministries is devoted to training institutions and programs. The programs tend to be small with relatively high unit costs and cannot be considered core activities. 4.45 While the MOL and MOCD operate some social welfare and community development programs that might serve an important need in society and are not being replicated elsewhere, they also operate education and health programs that could readily be consolidated with those run by the Ministries of Education and Health, respectively: * Consideration should be given to operating these programs on a fee for service basis and the public resources allocated to them reallocated to priority programs. Simultaneously, the labor, youth and welfare programs could 4 This compares to approximately Tshs 10 billion for health and education in FY97. Chapter 4 57 usefully be evaluated to see if their rates of return are as high as those for basic health and education; * As recommended elsewhere, consideration should be given to merging the MOL and MOCD into one Ministry of Social Welfare (to reduce overhead costs) and the expenditures incurred by these two Ministries on Education and Health be rationalized and consolidated with basic expenditures in these two sectors. The resources saved by this merger could then be reallocated to basic education and health; * Institutional closures and rationalizations aimed at raising physical capacity utilization and increasing average trainer: trainee ratios towards more efficient levels; Water Sector" 4.46 Water sector allocations are made at the Ministerial, Regional, District levels, and until recently through the National Urban Water Authority (NUWA). Total budgetary allocations for FY97 amounted to T Sh 16.8 billion, some T Sh 385 per head of the population, a 40 percent real decline from the FY94 level (Table 4.14). FY95 saw a sharp drop in allocations as Government recurrent funding was withdrawn from the National Urban Water Authority, on the understanding that it should be able to cover its operating costs from the collection of water rates. 4.47 In practice, NUWA failed to reach a satisfactory operating position, in part because controls remained over the level of water rates and, in part because it was extremely lax in collections. In the event, these constraints forced NUWA to restrict normal maintenance functions with a consequent increase in the rate of leakages within Dar es Salaam-a problem which has exacerbated the 1997 water crisis. The newly established Ministry of Water has taken steps to establish a clearer operating framework for the urban water authorities, including drafting of the legislative amendments. Key features of the revised operating framework are as follows: * the establishment of individual water boards for each major city and town; * the replacement of NUWA with DAWASA, with responsibility only for Dar es Salaam; * the development of performance contracts for the newly established water authorities based upon target reductions in accounts receivable, in the number of leakages, and in the response time to reported leakages alongside improvements in the quality of water; * a management framework based upon accountability to the beneficiaries (through a locally-nominated Board and through a defined complaints procedure), financially self-sustaining operations and managerial autonomy with regard to procurement, hiring and firing and tariff structures. 4 The information presented in this section was obtained in a series of interviews with the newly established Ministry of Water and from documentation made available on the new draft water sector policy. Table 4.14: Water Sector: Recurrent and Development Budget Allocations by Adminisrative Level 1993/94 -1996/97 (T Sh. Millions; Government & Donor Funding) 1993/94 1994/95 1995/96 1996/97 Water Sector Rec. Devpt. Devpt. Total Re. Devpt. Devpt. Total Rec. Devpt. Devpt. Total Rec. Dcvpt. Dcvpt. Total (Actual) (GOT) (Donor) Budget (Actual) (GOT) (Donor) Budget (Rev. Est.) (GoT) (Donor Budget (Budget) (GoT) (Donor) Budget Total (Nominal) 10,735 1,337 1,303 13,375 5,355 1,432 6,308 13,095 4,016 1,706 8,277 13,999 6,374 2,280 8,153 16,807 Total(1995prices) 14,507 1,807 1,761 18,074 5,355 1,432 6,308 13,095 3,213 1,365% 6,622 11,199 4,457 1,594 5,701 11,753 % of Total allocation 80% 10.0% 9.7% 1000% 40.9% 10.9% 482% 1000% 28.7% 122% 59.1% 100.0% 37.9% 13.6% 48.5% 100% of wich. Ministry of Water" 1,474 786 128 2,388 1,619 1,007 5,193 7,819 1,245 1,093 5,405 7,743 2,548 1,547 6,611 10,706 Regional Authorities 2,631 460 1,175 4,266 2,524 275 275 3,074 1,570 386 2,872 4,828 2,023 478 1,542 4,043 District Councils 1,407 - - 1,407 1,213 42 - 1,255 1,200 227 - 1,427 1,803 255 - 2,058 NUWA ' 5,223 91 - 5,314 - 108 - t08 - - - - Real Capita Exp. 516 64 63 643 185 50 218 453 108 46 223 377 146 52 187 385 % of Total Recurrent 3.3% 04% 04% 41% 1.4% 04% 1.6% 34% 08% 0.3% 1.7% 29% 1.0% 04% 1.3% 2.7% % Discretionary Rec. 4 5% 0.6% 06% 56% 1.9/ 0.5% 23% 4.7% 08% 0.3% 1 7% 2.9% 1 6% 0.6% 20% 4.1% %ofGDP 07% 0.1% 01% 0 8% 0.2% 0.1% 0.3% 06% 0.1% 0.1% 0.3% 0.5% 0.2% 0.1% 0.2% 0.5% Noies and Sources 1/ The Ministry of Water was established only in 1996/97 and had previously formed part of the Ministry of Water, Minerals & Energy. Relevant divisional expenditures are shown for the earlier years 2/ The National Urban Water Authority became a self-financing parastatal in 1994/95 and in 1996/97 was again restructured to focus on Dar es Salaam. Data from various Government Budget Books, MoF and Planning Commission; PER Review of Subsidies to Public Enterprises 1993-96, Vol 3, 1996 Coopers & Lybrand. Chapter 4 59 4.48 The proposed framework is sound and, notwithstanding the problems associated with the recent drought, there is no reason to depart from these proposals nor to revert to a structure reliant upon Government funding and more traditional parastatal structures. However, the priority in the short term will be to find appropriate ways to assist the newly established water authorities through the early development phase. In many cases, substantial capital works are required to upgrade water installations and introduce metering systems; management training and support is also needed. In both these areas, there would seem to be a useful role for donor support, so long as it is managed and guided by the Ministry of Water in line with the newly established principles of autonomy and self-financing operational structures. C. BUDGETARY PRIORITIES IN THE ECONOMIC SECTORS 4.49 The Tanzanian government is trying to disengage from the commercial activities that it has been pervasively involved with for the last two decades. It has recognized that production and distribution of most goods and services is best left to the private sector, and that it should focus its attention on providing the few core government services that will improve the operations and equity of results of private market activities. These include infrastructure and agricultural research and extension. Recent allocations are summarized in Table 4.15. Table 4.15: Share of Ministerial Expenditures in the Economic Sectors, FY95-97 (percent) Sector FY95 (Actual) FY96 (Approved) FY97 (Budget) Recurrent Dev.1 Recurrent Dev.1 Recurrent Dev. Agriculture 1.7 15.1 2.4 4.9 2.3 2.7 Tourism/Natural Resources 0.3 2.6 0.3 1.6 0.3 2.2 Industry and Trade 0.3 1.4 0.4 0.8 0.3 0.5 Infrastructure2 1.8 16.8 0.6 9.2 1.4 12.0 All Economic Sectors 4.2 35.9 7.7 16.4 4.3 17.4 1 Total Development Expenditure 2 Includes Works, Communication and Transport, Energy and Minerals and Lands Source: Table 13, Statistical Appendix 4.50 Infrastructure: As in the development of human capital, the Government has an important role to play in financing infrastructure. During the 1970s and 1980s, the transport system had deteriorated to the point that it was severely hampering Tanzania's economic recovery. This was true of all sub-sectors: roads, railways, ports, and civil aviation. The two main reasons for the deterioration were an inadequate allocation of resources to the sector and weak institutional structures. Since then the Government, together with the donor community, has made a concerted effort to improve the quantity and quality of infrastructure in Tanzania. Since 1990, there have been numerous large joint-donor investment projects in the infrastructure sector, totaling several billion dollars in investment and, reflecting Government's priority to infrastructure, new investments are being planned in the roads and energy sectors over the next few years. However, 60 Chapter 4 continuing emphasis on strengthening of institutional capacity and pricing and management issues are key to the successful implementation of these projects. The major public finance issue in the infrastructure sector is the recurrent cost implications of this growing portfolio of investments. Much of the current investment in rehabilitation is necessitated by inadequate attention to the maintenance of assets in the past. To avoid repeating the costly mistake of neglecting maintenance, in the roads sector, for example, in 1991 the Government established a roads fund to fund the maintenance and repair of roads. The charges for this fund consist of a tax on fuel (T Sh 60 per liter), and other charges, primarily vehicle licensing fees. By FY96, the roads fund was supposed to fully fund the total road maintenance costs. Prudent fiscal policies need to be adopted to ensure the proper maintenance of other public assets as well. 4.51 Agriculture: Agriculture is central to the Tanzanian economy, accounting for almost 60 percent of GDP and over 80 percent of its employment. A prime role of government in agriculture is the finance of public goods and services in support of agricultural development by the private sector. Accordingly, the key activities for Government remain support for research and extension. In addition, there are important effects on agricultural development from providing basic physical and social infrastructure, particularly rural access and feeder roads, and basic health and primary education. Support for research and extension is essential if ongoing efforts to increase agricultural growth are to be successful. The PER therefore supports the priority accorded by Government to research and extension. 4.52 Savings from the agriculture could be obtained from two sources: reduction in allocations and elimination of non-core activities. A review of the MOA budget reveals two items which appear to be arbitrarily large, and incidental to the operations of the Ministry as a whole: (i) T Sh 3.5 billion for an Agricultural Inputs Trust Fund; and (ii) T Sh 0.75 billion for the operation of the Strategic Grain Reserve. The inputs fund has little justification. It seems to represent an effort to guarantee inventory finance for importers and traders of agricultural inputs. While the trade may indeed face financial difficulties, it is not appropriate that the Government step in and cover these needs from the budget, with funds equivalent to approximately 15 percent of annual fertilizer consumption requirements. It is therefore recommended to eliminate this fund. 4.53 Some allocation to the Strategic Grain Reserve is required to cover the costs of delivering food to needy populations in times of crisis. However, the allocation seems unnecessarily large, representing over 25,000 tons of cereals, a quarter of the total stocks to be held by the Strategic Grain Reserve. Given the financial constraints facing Government and the understanding that stocks are reasonably high, an allocation of T Sh 500 million should suffice to meet emergency needs and to facilitate stock management. Reducing the budget by the value of the inputs fund, and 35 percent of the allocation to the Strategic Grain Reserve would free up some T Sh 3.75 billion. 4.54 A second approach to freeing up funds in the budget is to assess those activities, currently carried out by the MOA, which are inappropriate in a market-based agricultural Chapter 4 61 economy. It is recommended to eliminate funds in the recurrent budget for staff and operation of commercial farms breeding livestock or seed and the provision of mechanization and other services which private enterprises can provide as (or more) effectively. These operations could be sold to the private sector, and prior to sale subject to the same hard budget constraint as in other commercial parastatals. Recommendations for reform in agriculture: * Eliminate the Agricultural Inputs Trust Fund; * Improve the management of the strategic grain reserve and reduce its operating costs to T Sh 500 million in FY98; * continue the priority attached to agricultural research and extension through the medium-term; 4.55 Industry and Trade: Given Tanzania's budgetary outlook and the on-going efforts to redefine Government's role, consideration should be given to whether there is a role for a Ministry of Industry and Trade. Regardless, the Ministry would benefit from a reorganization to bring its structure more in line with its role and functions, similar to that in the Ministry of Agriculture. Its current organizational structure results in a thin scattering of its limited resources over a variety of programs that differ widely in priority. D. BUDGETARY PRIORITIES IN PUBLIC ADMINISTRATION 4.56 This section includes a brief discussion of expenditure issues in the public administration sector, which includes general administration, defense, and law and order. While intrasectoral issues were not examined in these sectors, the appropriateness of the size of Government spending in these sectors relative to others was examined, and is discussed below. 4.57 Table 4.16 provides a breakdown of central ministerial expenditures (i.e., excluding expenditures of regional and local governments) on public administration. It shows that in FY97 almost 40 percent of the recurrent budget, and 32 percent of the development budget (local funds), is devoted to public administration. It should be noted that these expenditures exclude public debt service and grants to local governments.42 Within the share of public administration in recent years, general administration has increased. largely at the expense of defense. 4.58 Figure 4.4 compares the intersectoral allocation of ministerial expenditures in Tanzania to the average of these expenditures in three Sub-Saharan African (SSA) countries: Ghana, Kenya, and Mauritius43. All sectors have been classified into four 42 Prior to FY92, the local Government grants were channeled through the Prime Minister's Office (PMO) and would have been included under the general administration category above. However, in FY92, they were moved out of the PMO's office and are now channeled through the regional governments. It would be useful to include local government expenditures in the analysis, but such data are not easily available for the comparator countries. 62 Chapter 4 Table 4.16: Share of Ministerial Expenditures on Public Administration, FY95-FY97 FY95 FY96 FY97 Recurrent Dev.' Recurrent Dev.' Recurrent Dev.1 Defense 14.7 7.4 10.2 10.2 9.6 7.7 Law & Order 7.0 18.6 7.2 8.9 8.1 8.4 General Administration 40.1 6.2 49.7 9.7 53.5 16.31 (Excl. debt Service) 17.81 - 21.8 - 21.5 - Total Administration 61.8 32.2 67.1 28.8 71.2 32.4 (Excl. debt service) 39.51 - 39.2 - 39.2 - I Local funds only Source: Table 14, Statistical Appendix Figure 4.4: Intsmational Copaison of Misierial Expenctm as %Sham of Total GoverMWrt Bpendie 45. 0%/ 40.0%D/ 35.0/6-- gTan.aia in FY97 30.0% 25.0% 20.0%/o.- 3Qury Averagein 15.0% "j"1993; (Knya Gian 10.0%ad hitius) 5.0%) 0.0/ 0L/) I ) 0((D groupings: general administration (including law and order); defense; social services; and economic services. The figure reveals that in comparison to these countries, there is scope for Tanzania to reallocate resources away from general administration and defense towards the social sectors. A comparison with Ghana reveals that Ghana spends two- third as much as Tanzania on public administration: Ghana's total spending on all administration (including defense) is 3 percent of GDP as compared with Tanzania's 5 percent. Part of these savings come from Ghana's considerably lower defense expenditures: only 0.4 percent of GDP as compared with Tanzania's 1.7 percent. Chapter 4 63 * In order to free up additional resources for the social sectors, the PER recommends that expenditures on public administration be reduced from their current level of 5 percent of GDP over the next 3 years. Recommendations for Reform in Administration 4.59 The savings sought from expenditures on administration and defense could be achieved by implementing additional recommendations of the Cost Cutting Commission's report. The report had made several recommendations for reducing administrative costs, including the following: * National Service. The Commission reviewed the benefits from the National Service (NS) program, and in light of "our inability even to provide adequately for the regular army and our universities and secondary schools" recommended that the NS be abolished, and the savings be used to strengthen basic education. In FY97, the abolition of the NS would have freed up T Sh 5.5 billion for basic education (10 percent of Government primary school funding); * Size of the Army. The Commission recommended that the reduction in the size of the army, which was initiated in 1987, be accelerated. Specifically, it is recommended that priority be given to downsizing the military and National Service by 15,000 in FY98; * Moving the Capital to Dodoma. The Commission noted that the economic and financial situation of Tanzania in 1973, when the decision was made to move the capital to Dodoma from Dar es Salaam, was completely different from the situation today. In light of this, the Commission recommended that the idea of making Dodoma the legislative capital and keeping Dar es Salaam as the administrative capital "should be considered seriously". According to the Commission, this decision could be reconsidered when the country's financial situation had improved; * Prison Management. The Commission saw scope for "significant savings" in other services like the prisons, where it recommended measures to reduce costs, including a review of the current remand arrangements and regulations to reduce the remand prisoner population; * Costs of Foreign Operations. The Commission recommended that foreign missions and embassies be restructured to enhance performance and reduce costs. To date, 4 of the 12 missions proposed for closure in the June 1993 Budget Speech have been closed. It also recommended that the costs of membership of international organizations be reduced. 4.60 While some of these recommendations have been accepted in principle by the Government, most have yet to be implemented three years after the Commission submitted its report to the Government. The adoption of these measures would free resources for improvement in the delivery of social services. 64 Chapter 4 Conclusions 4.61 Overall, the strategic allocation of funds across sectors is largely consistent with Government's priorities and, in recent years, there has been progress in increasing allocations to high priority activities. The key issue in strategic allocation is the development of realizable, medium-term sector development programs, consistent with the macro objectives, and the enforcement of these hard budget constraints at the sector and program level. 4.62 Until this fiscal year, deviations from the sector allocations was a serious concern. At the beginning of FY97, the overhang of domestic arrears amounted to an estimated T Sh 50 billion. The placement of accounting officers within agencies and the on-going implementation of the Central Payments Systems is expected to significantly reduce this source of budgetary deviation. 4.63 Given Tanzania's constrained budget outlook over the next several years, there is little prospect for increasing real allocations to the priority sectors or programs. Thus measures to improve the efficiency of expenditures are vital - especially in the social sectors, where Government is facing impossibly harsh decisions. Nonetheless, this chapter has identified activities that can be cut immediately or deferred. Combined with the potential for increased cost sharing, these amount to approximately T Sh 25 billion - or 5 percent of recurrent expenditures (excluding debt service). 4.64 The poverty focus of expenditure reallocation. The priority allocation of public resources to basic health, primary education and economic infrastructure should in itself lead to an improvement in the living standards of the Tanzanian population in general. However, it is possible and desirable to direct additional expenditures towards the poorer segments of the population. Since the bulk of local government expenditures are on basic health and basic education, one way of improving the delivery of social services to the poor is via the efforts to strengthen the local governments and to channel at least some of the additional resources through them to the poor. 5. STRENGTHENING PUBLIC EXPENDITURE MANAGEMENT 5.1 Earlier sections of this PER report have reviewed the present level and allocation of public expenditures in Tanzania and made recommendations on their future level and composition. Earlier chapters have made reference to the importance of the institutional framework in determining the pattern of public expenditures. The present section brings together the PER's analysis and recommendations concerning the institutional and organizational aspects of public expenditure management. It outlines the system and standards of public expenditure management Tanzania should aim for, reviews the progress there has been in strengthening public expenditure management and identifies practical short- and medium-term steps towards further improvement. 5.2 This part of the report is arranged as follows: * Section A provides the analytical framework and outlines the essential features of a well-functioning public expenditure management system; * Section B reviews the present institutional framework in Tanzania against this yardstick, assesses progress that has already been made in strengthening the system and considers what still needs to be done; * Section C considers the implications for organizational strengthening and how aid agencies can support the necessary improvements. 5.3 As is apparent from earlier sections, many of the findings of this PER echo those of PER89 and PER94. There has been significant progress but many of the fundamental issues are essentially the same. This Chapter therefore pays particular attention to the problems of strengthening systems and their implementation, trying to understand and to address the institutional factors that have caused obviously unsatisfactory patterns of public expenditure to persist in spite of efforts at reform and in light of this, makes recommendations for strengthening expenditure management. A. SCOPE OF THE PUBLIC EXPENDITURE MANAGEMENT 5.4 There is now wide agreement within Government and amongst Donors, that it is necessary to take a broad view of public expenditure management-one that is concerned with all stages in the budget cycle, including implementation as well as formulation of the budget. Specifically, that the effectiveness of the system depends as much on budget implementation (expenditure control, accounting, monitoring and audit) as on planning and budget preparation. Thus, there needs to be a comprehensive view of public expenditures - including recurrent and development expenditures, both on-budget and off-budget items (the latter still include a large proportion of aid flows and significant flows to parastatals) and revenues. It is important too, to consider personnel as well as financial resource management. 66 Chapter 5 5.5 Against this background, it is useful to distinguish between three main levels at which the public expenditure management system operates: * macro: the control of budget aggregates in a way that supports overall economic management; * strategic: the allocation of public resources in ways that reflect public policy priorities; * operational: the efficiency and effectiveness of resource use in implementation. 5.6 In reviewing public expenditures, there is a natural tendency to focus on allocative shares (reflecting the competitive process of the annual budget and competition for development funds etc.), but macro and operational dimensions are equally important. Moreover, there are tensions between the different levels, which are very well illustrated by Tanzania's case. Specifically, competition for strategic allocations invariably creates pressure to adopt a budget that is too ambitious to be sustainable in terms of macroeconomic management. This is at the root of the 'underfunding' (overcommitment) problem that has been so apparent for so long. As earlier sections have noted, the present government's determination to enforce macro fiscal targets has simply made the degree of overcommitment starkly clear: but the problem of adopting an affordable budget is far from simple-as the Budget Commissioner noted at the November 1996 PER workshop, it is much more difficult to agree on something small but affordable than something bigger but unattainable. In turn, the built-in pressure for overcommitment of Government resources has direct implications at the operational level: because there are insufficient resources adequately to finance all the activities that Government attempts, implementing agencies find that nominal budgets are less than the full cost of an effective service, and, often, that funds released are less than those budgeted. This undermines the efficiency and effectiveness of the services that are provided, not least because, in the short term, meeting staff costs takes priority over complementary inputs and leads to the familiar syndrome of poorly paid personnel operating from inadequately maintained premises and without the materials they need to do the job. 5.7 'The Government' is not a single, rational, entity making central decisions about the budget and its implementation. Rather, the configuration of planned and actual expenditures is the outcome of interactions between many participants whose behavior can be understood only by considering the incentives and signals to which they respond. Thus, for example, budget guidelines to be followed, or if they do not expect actual releases of funds to correspond to the amounts budgeted. They are likely to seek funds separately from different sources (government and donors, recurrent and development budgets) if they see this as an effective strategy for maximizing the resources available to them. They will be reluctant to share information that might work to their disadvantage, and they are unlikely to reduce the scope of their activities or offer savings in expenditure if they anticipate that doing so will simply reduce their share of total resources relative to other agencies. Chapter 5 67 5.8 The incentives to which players actually respond may be different from those embodied in the formal rules of the game. An important part of what the present Government is doing is to try to restore a culture of compliance after years in which breaking of formal rules (e.g. by incurring expenditures in excess of budgeted allocations) has gone unsanctioned. Important factors in ensuring compliance with established rules are transparency (are the rules clear and understood not only by those to whom they apply but also by other participants) and accountability - do players have to report and justify their actions? Mechanisms for accountability have to be reasonably certain and timely if they are to be effective, so that, for example, an audit process is unlikely to be effective if it only takes place long after the event and is not expected to lead to serious disciplinary action. 5.9 Traditional accountability mechanisms for public expenditures have focused primarily on propriety and control. The standard audit process has been concerned mainly to ensure that expenditures have been properly authorized and that the stipulated formal procedures have been followed. The focus is thus on accounting for inputs. There is now an international trend towards focusing more on the outputs of public expenditure (e.g. what health care services were delivered by the resources allocated to the ministry of health?) and on the ultimate outcomes (what differences in people's health status did these services achieve?). This implies holding managers accountable for outputs while giving them more discretion over how inputs are deployed. Such a results-oriented approach has the potential to achieve greater efficiency and effectiveness in service delivery, but is inconsistent with a tightly-specified line-item budget handed down from the center. More decentralized decision structures are needed: it is not practical or efficient to collect all information at the center and make detailed decisions there; not only is a centralized approach unwieldy in terms of the information flows required, it fails to provide an incentive structure that is conducive to prioritization or efficiency. 5.10 Different timescales are relevant for the different levels of expenditure management distinguished in para 5.5. For example, macro control of fiscal aggregates in the short term is often difficult to reconcile with medium term strategic reallocation of resources. The dilemmas of the cash management system illustrate this point: it is entirely understandable that in the short term wages are treated as a priority for the allocation of available cash; but, in order to achieve a budget that is sustainable in the longer term, there needs to be a strategic reallocation that is based on a different prioritization between sectors and functions. Reorganizing public services to use resources more efficiently requires managers to take at least a medium-term view, which is difficult if they cannot confidently predict their actual budgets even in the very short- term. 5.11 There are potential conflicts between different reform objectives for-example, the need to re-establish a culture of compliance may be at odds with the desire for more decentralized management, where managers are given more discretion over how they use resources. Personnel management is a case in point in Tanzania: it is desirable to give 68 Chapter 5 line agencies more discretion to achieve an efficient balance between personnel and other inputs, but important not to lose overall control of employment in the public service. 5.12 In summary, it is not enough for a report such as this one to specify a level and pattern of public expenditures that, for Tanzania's present economic circumstances, reconciles resource availability with mutually consistent policy priorities. There is a deeper need to develop a public expenditure management system that can continually reconcile macro, strategic, and operational objectives, adjusting to changing circumstances from one budget cycle to the next (and also during budget implementation within each fiscal year) and which avoids getting stuck in suboptimal (because highly inefficient) patterns of resource allocation and use. 5.13 This implies: * A need to build a national consensus about objectives, resource levels, and the strategic priorities for government so as to limit the extent to which, at technical level, bids lead to overcommitment of resources, and to counter the repeated experience where political decisions disrupt expenditure plans early in the budget year; * A need for discipline in implementation-this is both about control at macro level so as to serve the objectives of fiscal stabilization, and about predictability of resources at implementing agency level, so that the limited resources that are available can be allocated to priority services and used efficiently; * A need for a comprehensive approach to public expenditure management (encompassing aid, development, recurrent, implicit flows, deployment of personnel, etc.); * A need to link short and medium term perspectives, so that short-term imperatives do not prevent the strategic realignment of public expenditures that is needed. B. PUBLIC EXPENDITURE MANAGEMENT IN TANZANIA 5.14 The first Tanzania PER (PER89), thoroughly documented the mismatch between the role Government was attempting and the resources available to it. Among other points, it highlighted the need to redefine the role of Government in the economy; the need to avoid focusing primarily on the development budget and capital projects; to make the efficiency and effectiveness of recurrent expenditures a central concern - the link between public expenditure rationalization and civil service reform; the need to revamp the heavily distorted tax structure; and the need to strengthen aid management as part of a comprehensive approach to the management of public expenditures. It identified institutional changes in planning and budgeting systems that would facilitate rationalization and better management of expenditures. A key recommendation was to strengthen the link between planning and budgeting by moving towards a system of Chapter 5 69 rolling three-year plans addressing the entire budget (as opposed to the traditional 5-year fixed, capital-focused plan). The 1994 PER endorsed the same approach and focused on the practicalities of carrying forward the implementation of the recommendations, especially in relation to rationalization of the development budget, accounting for aid and linking budgetary reform to a systematic overhaul of the civil service, and also focused on the social sector as budget priorities. 5.15 The present PER recognizes that the measures implemented so far have not yet achieved the transition to a sustainable pattern of public expenditures focused on the effective and efficient delivery of core public services. This does not mean that past prescriptions were inappropriate, although the difficulties of implementation may have been underestimated. Given the interconnections between different elements of the public expenditure management system, necessary reforms must complement each other. Many of the necessary pieces are in place, and the requirement is not to abandon past efforts but to adjust and build on them. The paragraphs which follow give the PER team's assessment of four main elements of the reforms and identify the ways in which they need to be reinforced. These elements are: - civil service reform; - the RPFB; - specific issues relating to the development budget and the integration of recurrent and development budgets; and - issues concerned with program costing, planning and management at the sector level. Civil Service Reform 5.16 The PER team recognizes that the CSR Program (CSRP) has made impressive progress on a number of fronts: it has succeeded in developing a climate for change across central government and in getting donors to focus assistance on a coherent reform program. It is important that the exercise is not perceived simply as a retrenchment program: retrenchment is a means towards achieving an affordable, adequately remunerated public service that can help deliver core public services more effectively, and the long term institutional goals of restoring morale and professionalism are vital. Recognizing the progress that has been made, the following paragraphs highlight a number of areas in which strengthening and consolidation of the CSRP are especially important. 5.17 First, the links (in both directions) between civil service reform and the budget need to be strengthened. Already there is increasing co-operation between the Civil Service Department and the Ministry of Finance on the implementation of the Phase II O&E reviews and also in the day to day management and control of the payroll. At a higher level, the CSRP, and the ministerial O&E reviews in particular, need to be informed by a more definite resource framework. The ongoing civil service affordability 70 Chapter 5 study" will help to establish realistic targets for the future size and cost of the public service. Within the RPFB process (see below) Ministries/agencies need to be given more definite guidelines on future funding that reflect national decisions on priorities between different government activities and services. Otherwise, however conscientious the sector-level reviews are, their aggregate resource demands are certain to exceed what Tanzania can sustain, and will perpetuate the operational inefficiencies that result from overcommitment. Both CSRP and RPFB require a thorough co-ordination of the resources of Government, and should be co-ordinated and timetabled as a combined exercise. 5.18 Secondly, while it is extremely important that those affected be closely involved in the process, too much is being expected from the largely internal, bottom-up, O&E reviews within each ministry. The absence of a clear resource framework for these reviews has been noted. In addition, it is unlikely that a predominantly internal exercise will be radical enough in redefining and cutting back ministry functions. In the light of the affordability study, there will need to be a strong top-down element of decisions about which functions are to be sharply cut back or eliminated altogether. In the past, Governments have proved reluctant to implement recommendations on these lines, such as those proposed by the Presidential Cost Cutting Commission in 1994. 5.19 Third, there is the question of the overall ministerial structure of Government. It is encouraging that the Government is now acting to prune the Regional tier of administration. There is still a case for a significant reduction in the number of central government ministries (the Cost Cutting Commission argued that 12-15 would suffice). There would be direct savings by reducing the extra administrative costs associated with separate ministries, but consolidation could also help to reduce the fragmentation of some sectors between different, competing ministries; it would thus facilitate the sector-wide approach to expenditure planning that is advocated below. 5.20 Fourth, the CSRP has a valuable role in helping to develop innovative management and budgetary structures that are more conducive to efficiency and effectiveness. The pattern that is being developed for Executive Agencies is likely to be relevant for a wide range of public service activities where privatization is not contemplated. The Rolling Plan and Forward Budget 5.21 The RPFB represented a substantial break with previous traditions, but, in the fourth year of the RPFB system, there is a widespread feeling that the RPFB has not delivered what was expected of it45. The present status of the RPFB is unclear. The main volume of the 1996/97 -1998/99 RPFB was prepared, but has not been reproduced and 4Undertaken by the ESRF and funded by the IDA-supported Public and Parastatal Reform Project. 45Project Completion Review of the Rolling Plan and Forward Budget, Government of Tanzania, ODA and Mokoro, August 1995. Chapter 5 71 circulated. In the context of proposed changes to the role and structure of the Planning Commission, it is not clear whether the production of the RPFB is intended to continue. The O&E proposals for the Planning Commission makes reference to a "three year Strategic Public Investment Program" and a "National Investment Plan" without making clear how these relate to, or differ from, the RPFB. Beginning with the FY98 budget, Government, in line with the Cost Cutting Commission's recommendation, has announced that the Ministry of Finance, instead of the Planning Commission, will be responsible for the preparation and administration of the development budget. 5.22 Meanwhile, budget administration has been dominated by the Government's emphasis on expenditure control through the cash management system. This has reinforced skepticism about the feasibility of realistic medium term expenditure planning, given the difficulties experienced in predicting funds available in the short term. On the other hand, constraining expenditures to match current revenues has clearly demonstrated that the budget originally approved for the year was both unrealistic and unimplementable, and hence, the need for non-marginal restructuring of expenditures. 5.23 In several respects, the RPFB has represented a distinct improvement on previous approaches to planning and budgeting in Tanzania. The Ministry of Finance, the Planning Commission and the Bank of Tanzania, have cooperated in using a simple macroeconomic model to prepare consistent economic projections. The annual policy statements have reflected the changing policy stance of the Government, and have ostensibly linked policies to expenditure priorities. The link between recurrent and development budgets has, at least in principle, been acknowledged. And some progress was made in rationalizing and prioritizing the development budget. However, there have also been serious shortcomings. Expenditures under the RPFB have always departed from the projected path and projections for the forward years have not been taken seriously. Although line ministries have prepared and presented combined submissions for their recurrent and development estimates, recurrent and development submissions have in practice been reviewed separately by Finance and Planning. A large proportion of aid flows still fail to be recorded in the Government budget and accounts and until FY98, there has been only a limited reduction in the number of projects (see below). 5.24 There are several reasons why the RPFB has so far failed to make the concept of a medium-term expenditure framework properly operational. Budget projections were overoptimistic, thus evading the problem of attaining a realistic initial budget. The Governments political commitment to the budget figures it announced was weak: thus there were frequently large reallocations early in the budget year. General expenditure discipline was poor, and budgets were also thrown off by expenditure arrears carried forward. And weaknesses in estimating, monitoring and accounting were another source of uncertainties. Nonetheless, the principles and objectives of the RPFB approach remain appropriate. Given the present Government's seriousness of purpose, the RPFB, if suitably modified, can still be an effective instrument. The principal requirements are: 72 Chapter 5 * Firm political support for the RPFB; * Changes to the timetable and sequence of budgetary discussions; so as to advance the review of the medium-term program ahead of the budget; * Strengthening of the technical foundations of the RPFB; * Greater involvement, and a more responsible role for, sectoral ministries; * Better integration of recurrent and development budgeting (including aid funds). 5.25 Political support. Effective political support cannot be based on post hoc endorsement by politicians of a document prepared by technicians and submitted for political approval only as a final step. Political commitment to the RPFB implies both that the document is treated as an authoritative statement of Government policy and priorities, and that the process of RPFB preparation and revision is treated as the appropriate forum for continuing to resolve issues of strategy and expenditure prioritization. Hence, there would be discipline in adhering to the resource allocations embodied in the RPFB (there would inevitably need to be adjustments, but these should be based on transparent criteria - a modification, not an abandonment of priorities). Indeed, the credibility of the RPFB depends on transparency and on a public demonstration of Government's political commitment to it. Such commitment would be revealed in a number of ways: the document would be published and widely circulated; it would take precedence over the annual budget (and the Economic Review) in presentation to Parliament - national debate over the overall resource framework and resource priorities should precede detailed consideration of consequential annual budget appropriations; it would be consistent with the projections and undertakings agreed between Government and the Bretton Woods institutions in the PFP, and would thus simultaneously serve domestic and international audiences. Its credibility and practical value as a guide to Government policies could be increased by including practical, monitorable statements of the main actions that will be undertaken by each ministry during the year, with each roll-over reporting back on implementation. The Government's political commitment to the RPFB must be manifested to donors too, by making it clear that the RPFB provides the basic framework for aid relationships as well as guiding the allocation of domestic resources. 5.26 Timetable. To achieve credible commitment to the RPFB, there needs to be an appropriate sequence of preparation, in which broad issues and fiscal macro targets are first reviewed and agreed at the political level, as the basis for subsequent elaboration in the annual budget and the detailed RPFB. Aggregate macro targets and their breakdown onto sectoral expenditure frameworks would thus be agreed before line ministries completed the detailed preparation of their budget submissions (thus emphasizing the responsibility of the line agencies for optimizing expenditure patterns within known constraints, rather than submitting optimistic bids that perpetuate the inefficiencies of underfunded activities). For as long as the Government's economic strategy depends on approval and financing from the Bretton Woods agencies, it is important to synchronize discussions with them and with the domestic budget/RPFB timetable. The main Chapter 5 73 substantive discussions with the IMF should take place at the stage when the budget frame for the coming year is being agreed, and before detailed budget preparation takes place. Belated review and adjustment to meet IMF criteria (which at times has continued after the budget year has commenced) undermines the credibility not only of the budget/RPFB preparation exercise but also of the Government itself. However, even with RPFB preparation following a more appropriate sequence and timing, it would still have to be accepted that RPFB is not an immutable blueprint, and that its annual rollover provides scope for change and revision. 5.27 Technical underpinning. While political commitment to the RPFB is critical, it needs to be reinforced by strengthening of the technical foundations on which both preparation and implementation of the RPFB are based. Key areas (for most of which relevant initiatives are already taking place) include: * Strengthening of macroeconomic analysis and modeling. The macroeconomic analytical capacity of the Ministry of Finance is dangerously thin (especially when contrasted with the number of staff available to the Bank of Tanzania); the restructuring of the Ministry should give due weight to the establishment of a well-staffed macro analysis unit; * Continued strengthening of budget discipline, supported by the strengthening of accounting and monitoring functions; * Reducing the likelihood and scale of shocks to expenditure management through better cash management techniques and procedures improved debt management, strengthened personnel management and better forecasting and control of the wage bill and better accounting for aid resources; * Improving the basis for strategic expenditure planning by better costing of options. At sector level there is a need for better analysis of the costs of properly funding efficient services. At aggregate level, the civil service affordability study will be an important input; * General strengthening of the capacity of the Ministry of Finance (current initiatives are discussed in Section C below); * The general raising of standards and productivity across the public service, as supported by the CSRP. 5.28 While the recent demonstration of commitment to aggregate expenditure control is promising for improved strategic allocation, it is by no means sufficient for providing a satisfactory macroeconomic basis for medium-term expenditure planning. Recent improvements in fiscal discipline have focused on immediate, within year controls, rather than medium-term strategy. As discussed above, improved control has not developed from improvements in capacity in macroeconomic monitoring and management. For the medium-term as for short-term macroeconomic management, there is an urgent need to strengthen capacity in this area. 74 Chapter 5 5.29 While the RPFB exercises have been diluted by rapid departure from the main macroeconomic assumptions in the plan, it is also not evident that the proposed macroeconomic framework was in fact consistent with maintenance of macroeconomic stability. The budget component of the macroeconomic model used to prepare the RPFB is not fully integrated with the rest of the model, which gives rise to problems of internal consistency. It should be recognized that continued preparation of the macroeconomic component of the RPFB represents a significant achievement in view of the few staff conversant with the framework and a background of persistent non-implementation. However, in the context of renewed seriousness in macroeconomic management, it would be appropriate to overhaul the existing framework and the institutional arrangements for its production, with the following objectives: * full integration of the projections for central government operations with the real sector, monetary and external sector components. This might entail rebuilding a core framework with the central government operations determined jointly with the other sectors on the basis of an explicit set of common assumptions on key factors. For ease of handling, this might initially require some simplification of framework components such as the detailed multi-sector treatment of the real economy; * review of data requirements and quality and timeliness of existing data inputs, with a view to strengthening key statistical inputs and assisting timely production of RPFB, earlier in the budget cycle. A first step would be to define key stages in the budget cycle at which the framework should be fully updated, with documentation of the required timetable for data collection and input to the framework; * closer coordination between fiscal authorities, the Planning Commission and the research department of the Bank of Tanzania in preparation of the framework; * a higher profile for key macroeconomic issues in the RPFB process, such as submission of a mid-year Cabinet paper, highlighting proposed medium-term strategies on domestic revenue, debt, aggregate expenditure, etc. 5.30 Sector ministry roles. By providing sector expenditure guidelines within an overall medium term resource framework, it is intended to give line agencies both the responsibility and the discretion to identify priorities and to allocate and use resources more effectively within the sector. In future, line ministries need to take more responsibility and ownership for the sectoral inputs to the RPFB. Integration of Development and Recurrent Budgets 5.31 The Ministry of Finance in early 1997 assumed the Planning Commission's responsibility for preparation and administration of the development budget. This provides a critical opportunity not only to secure better integration in the planning of recurrent and development expenditures but also to achieve a rapid and radical Chapter 5 75 rationalization of the portfolio of development projects. In the longer term, the Government should reconsider whether maintaining a separate development budget continues to be appropriate. In the short term, the priority must be to ensure that, both at sector level as well as in the center, expenditure programming takes place on the basis of a comprehensive view of all resources applied to the sector. This objective can be served by: - organizing the staff of the Budget Division in sector teams, rather than maintaining separate sections for recurrent and development budgets; - ensuring that submission, review and meetings on estimates all deal jointly with recurrent and development proposals; - issuing budget guidelines that focus on the total ministry allocation of domestic resources, rather than beginning with an arbitrary division of available funds into recurrent and development; - applying a consistent chart of accounts to both budgets (so as to facilitate sector-wise analysis across the budgets and simultaneously to allow expenditures in both budgets to be identified as recurrent or capital in character); - undertaking an immediate and radical rationalization of projects (see below): - close liaison with donors, both to ensure that government's information about donor projects is complete and consistent, and to ensure that donors observe government priorities in project funding (including withdrawing and abstaining from projects government designates as non-priority). 5.32 Integrated management of recurrent and development budgets will place new demands on the Budget Division. Hitherto it has been concerned primarily with administration of the annual budget. Its new role requires a much greater emphasis on analysis and forward programming of both recurrent and development expenditures. Increasing the capacity of the Division thus requires the rapid introduction of new skills and procedures as well as additional personnel. The Development Budget 5.33 The Development Budget accounts for only a small fraction (10 percent) of the recorded budget (and for an even smaller proportion of actual disbursements), but this yardstick understates its importance because of its influence on a much higher volume of aid flows. Historically the Development Budget has been central to the Planning Commission's role, representing the resources over which it had direct control, and reflecting its aspiration to guide both Government and the economy by directing capital investment. 5.34 The existence of separate recurrent and development budgets is widespread, but its appropriateness is increasingly questioned. Traditionally the development budget was conceived as an investment budget focusing on capital projects. In principle there are some advantages in this. Capital projects are typically one-off activities with a multi-year 76 Chapter 5 time scale, and in practice much of the development budget was financed by donors, and most donor funds were directed towards such projects, so there were certain advantages in having a separate administrative channel for them. But the existence of a separate development budget can have serious disadvantages too. It tends to build in an upward pressure on government expenditure commitments: this happens because of the natural tendency of the responsible agency to try to expand the resources it controls, exacerbated by the fact that it is not directly responsible for the eventual recurrent costs of the projects it sponsors. As public expenditures come more under pressure, there is a tendency, as in the recurrent budget, for resources to be spread too thinly across too many projects. At the same time, the development budget comes to be seen as an alternative source for essentially recurrent funding. The separation of responsibilities for recurrent and development budgets makes coherent sector planning more difficult. As projects proliferate there is a danger that the planning agency becomes preoccupied with administrative detail, losing sight of strategic issues. 5.35 All these weaknesses have been very apparent in Tanzania. In addition, the deterioration of standards of administration and financial management led to an increasing proportion of aid funds bypassing the formal budget. Consistent themes of recommendations since PER89 have been: to reincorporate aid flows in the budget, to seek more consistency between recurrent and development budgets, and to prioritize development projects while sharply reducing their numbers, so as to achieve a more compact, but fully funded, set of projects. An attempt to pursue rationalization through a computerized database of all projects (PROMIS) was ultimately fruitless, but there have also been efforts at rationalization within the context of the RPFB. There has been a reduction in the number of separate projects (partly through consolidation rather than completion or termination of projects). As recommended in PER94, projects were classified in three levels of priority (supercore, core and noncore). A second volume of the RPFB was prepared, showing this classification and grouping projects in a smaller number of programs (but no Volume 2 of the RPFB has appeared for 2 years). Box 5.1 gives an overview of the development budget for 1996/97. 5.36 As Box 5.1 illustrates, efforts to rationalize the development budget have had only limited effect. The number of separate projects is unmanageably large, projects are severely underfunded, and the supercore/core classification has not protected the flow of funds to designated projects. The development budget still fails to capture a large proportion of donor funding. The potential of the program classification incorporated in RPFB Volume 2 has not been realized: the classification has been retrospective and has not actually been used as a framework for forward programming. Nor has Volume 2 been used as a basis for checking with donors the completeness and consistency of development budget information about donor projects. Moreover, as already noted, the intention to coordinate the review of development and recurrent budget submissions during RPFB preparation has never been realized in practice. Chapter 5 77 Box 5.1: Summary Data on the 1996/97 Development Budget At present, Government allocations to the Development Budget are minimal and analysis of the Development Budget is therefore mute. Only a fraction of the budgeted 1995/96 GOT funding for the development budget was released, and in the current FY there were no releases of GOT funding before November 1996. This emphasizes the need for a radical rationalization of the development budget to bring it into line with a realistic assessment of GOT's financing capacity. The following table highlights key data. 1996/97 Development Budget All Projects Number T Sh'000 Supercore 125 4,173,958 Core 1,197 21,671,564 Noncore 20 1,583,428 Total 1,342* 27,428,950 1995/96 actual GOT releases 4,017,140 Projects with donor funding in 1995/96 250 Underfunding: 1996/97 submissions were 263 percent of ceiling Regional and Local Government Projects Total Regional and LG projects of which 850 5,233,329 donor funded in 1996/97 50 * Includes only projects with budgeted funds (local and/or foreign) for 1996/97. There are approximately 60 additional projects in the Rolling Plan for 1996/97 - 1998/99. The total number of projects included has declined from a peak of about 2,000 at the time of PER94. The GOT funds budgeted for 'supercore' projects this year exceed the total amount released for all projects in 1995/96. Only a tiny proportion of projects have been designated noncore. The extent to which budget submissions exceeded the ceiling, and eventual allocations, indicates that the projects included have not been fully funded. Thus it appears that all the fundamental weaknesses of the development budget persist, and have been made more obvious by the Government's determination to control its aggregate expenditure more effectively. Numerically, regional projects predominate, but, because they are typically small, they account for less than 20 percent of budgeted GOT funds and fewer than 6 percent of these projects are shown as having any donor funding. The Government is undertaking a radical reduction in the regional tier of administration, which would make a thorough review and rationalization of these projects very timely. 78 Chapter 5 5.37 The present situation of minimal GOT funding of the development budget highlights the need for a rapid and thorough rationalization of projects. Criteria for inclusion of projects in the budget are, inter alia: * the project is in the Government's priority sectors of education, health, water, economic infrastructure and environment; * the project has adequate documentation, justifying its financial and economic viability; * has not been under implementation for more than, say 10 years; * is not recurrent in nature; * could not be undertaken in the private sector; * projects with donor funding will only be included in the budget if there is a firm commitment for disbursements prior to completion of the budget documents; * projects that do not report on donor funds will be excluded from the Government budget. Based on these criteria, it should be possible to reduce the number of projects from over 1,400 to a manageable 300 or so over the next two years. Undertaking such a rationalization, however, will require strong political support, as part of a broader Cabinet commitment to pursuing the RPFB and CSRP in a coordinated and carefully sequenced fashion. Close coordination with the donors in the closure or divestiture of existing projects will be needed. Public Expenditure Management at Sector Level 5.38 At sector level the requirements for better use of public expenditures are two-fold: first, there is a need to carry through a restructuring of public expenditure in the sector so as to match government activities to the level of resources that is available on a sustainable basis; second, there need to be improvements in the efficiency with which available resources are used. Earlier sections of this PER have reviewed social sector expenditures in particular, but the same principles of sector-level expenditure reform are broadly applicable to all sectors. While there have been significant changes in Government policies concerning its sector activities, these have not, in most cases, been adequately reflected in budgets and expenditure programming. 5.39 Part of the problem is that implementing agencies have continued to live in hope of an increase in real funding levels that would allow existing activities to be fully funded, and have persisted with short-term tactics of spreading available resources across all activities, rather than make a radical adjustment that would involve curtailing some activities in order to fund other properly. But even where the need for a restructuring of activities and expenditures is recognized, the existing framework of public expenditure management inhibits reform. Sector ministries have not been given a clear, or politically Chapter 5 79 credible, indication of the resource framework on which their medium term expenditure planning should be based. Although there have been exhortations to make savings and efficiency improvements, the incentive to do so is lacking when line agencies' perception is that any savings they make are likely to be appropriated by the center and reduce the sector's future resources even further. The present institutional framework inhibits fundamental expenditure reform in other ways too: a comprehensive sector-wide view of requirements is inhibited by the fragmentation of expenditure planning between recurrent and development budgets, and the further complication that aid funding is often determined on a piecemeal, project by project basis. Reviews of government's role in each sector (notably the CSRP-coordinated organization and efficiency reviews) have not been given clear resource parameters to start from. The established line-item incremental approach to budgeting does not support the non-marginal adjustments that are needed; it focuses on the costs of inputs and does not facilitate analysis of the costs associated with different outputs. It is difficult to discover the current costs of different programs, let alone what their costs would be if adequately funded. Thus it is difficult to present policy makers with clear expenditure options and trade-offs. 5.40 Efforts to restructure and improve the efficiency of public expenditures at sector level need to proceed along the following lines: * The ongoing O&E exercises need to be reinforced by a closer link between RPFB and CSRP processes. Line ministries need to base their reviews on explicit medium-term budget projections that are politically endorsed as reflecting Government's sector priorities; * These sector resource guidelines would give sector ministries scope to reallocate expenditures within the sector without fear of being penalized for making savings; * There should be priority for efforts to analyze existing expenditures according to program objectives. Emphasis should be on rapid, pragmatic review and analysis of program costs: this does not require a formal shift to an integrated system of program budgeting, and the revised chart of accounts that is being introduced will facilitate more program-oriented analysis and monitoring; * It is essential that options in each sector are reviewed on a sector-wide basis. Line ministries have already made progress in linking recurrent and development budget preparation, in as much as unified budget committees have been established. There also needs to be sector-level coordination with donors to ensure that project interventions are consistent with a coherent and sustainable sector expenditure strategy; * Sector ministries should be given more responsibility for preparing their sections of the RPFB, which should include explicit links between policies and sector expenditure proposals, together with time-bound sector action programs that would provide a framework for monitoring implementation; * As described in Chapter 4, there should be continuing moves to increase the efficiency of resource use by giving more autonomy to implementing 80 Chapter 5 agencies, and increasing the scope for competition between public and private providers; * This PER has not focused on the reform and strengthening of local government; however, the balance of responsibilities as between control and local government bodies will be a critical issue for review in most sectors. C. ORGANIZATIONAL IMPLICATIONS AND DONOR SUPPORT 5.41 This section reviews the organizational implications for public expenditure management of the reform strategy that has been outlined. There are direct implications for administrative reform and capacity building, and the role of donors is relevant both in terms of their direct support to improvements in public expenditure management and in terms of the way aid programs as a whole are managed. Organizational Requirements 5.42 The need to strengthen different aspects of public expenditure management has been recognized for years, and numerous relevant initiatives are under way, supported by a variety of bilateral and multilateral aid agencies. The main initiatives are summarized in Table 5.1 updated from the September 1995 World Bank report on the "Strengthening of Planning, Budgeting and Resource Management". As that report noted, these initiatives have been largely donor-led, and in future should be much more systematically controlled by the Government. 5.43 In the context of the present PER, the following points deserve emphasis: * The importance of the link between political and administrative levels of government cannot be overemphasized. Technical improvements in planning, budgeting and accounting will be effective only if there is political support for expenditure control and rationalization; * Similarly, systematic links need to be developed between the CSRP and the other initiatives summarized in Table 5.1; * The role of the Ministry of Finance in expenditure planning and management is pivotal, and its strengthening must be a priority. It needs to become proactive in coordinating the various technical assistance programs that focus on the expenditure management system; * The most urgent priorities for strengthening within MoF are (a) the Budget Division, whose role needs to be extended from short-term administration of the recurrent budget to encompass medium-term strategy programming of both recurrent and development budgets and (b) the macro analysis function, which is fundamental to the Ministry's capacity to think strategically and to play a full part in macroeconomic management of the economy, but which is grossly understaffed; * There is an urgent need to clarify the respective roles of the Ministry of Finance and the Planning Commission in relation to medium-term policy formulation and expenditure planning. Table 5.1: Main Interventions Related of Planning and Resource Management Intervention/Dates Main Elements and Current Status Government Agencies Donors Involved _______________________Principally Involved Rolling Plan and Forward Introduction of new RPFB system encompassed elTorts to: Intended to be joint ODA: repeat short term Budget (RPFB) -strengthen macroeconomic forecasting through development of a macro model leading role for Planning consultancies and workshops Discussion of present system (MACMOD) and link this to budget targets Commission and Ministry for RPFB preparation, up to began seriously with first PER -provide authoritative statement of macroeconomic and sectoral policies of Finance. 1995. (published 1989). Design -strengthen the link between recurrent and development budgets workshop for RPFI3 held in March -develop three year forward budget to introduce strategic element to (considering In practice RPFB NORAD: repeat short term 1992, and first RPFB covered recurrent and development expenditures jointly) and facilitate restructuring of public Coordinating Committee consultancies plus training to period 1993/94-1995/96, with expenditures (including also CSD and develop and refine MACMOD annual roll-over since -rationalize the existing portfolio of development projects PMO) never functioned and assist in its use. (WB -incorporate aid resources within the planning and budgeting system and most coordination has financed earlier development of been by Planning BUDMOD.) Macro support Ri de1 has supplanted former production of Annual and Five Year Plans. Signiicant Commission. now resumed after hiatus. progress in mhacro modelig (with cooperation etveen Planning and Finance). Forward budgeting element has been ineffective, and intention to consider recirient UNP: assistance in and development budgetsxjoinly never realined R has tended to he rowded out development of PROMIS by urgency of Annual Budget preparation, and even annual expenditure projections projct data base, including one have been quickly de-railed by expenditure indiciphneand hidure to anlicipale and Jong feri TA. Fulure control coimit ents effectively. oevelopment Budget rationaliation disappointing; assistance subsumed tnder progn -oriented Volume 2 ofrRe11 has not been followed through, and AMAI (see below). supercore/core classification failed to lead to snubstarial reduction in projects or protection of funding for priorities 1996/97 RPFwr was not circulated, and Volume 2 not prepared. toconsiderrecun Budget Management Long term effort to improve the quality of budget preparation and management and Located in Budget SIDA. 1-2 long term TA. plus Development Program (BMDP) improve the capacity of the Budget Division and budget staff in implementing Division, Minist of training, consultants, Successive phases since 1986/87. agencies. Aimed to facilitate the formulation of policies and the setting of priorities Finance, equipment Terminated in 1997 by developing adequate systems, procedures and competence of budget managers and technicians. Elements included Steering Committee training fco budget division staff and for budget managers in spending agencies,- includes senior MoF staff -use of computers in preparation of the budget books; and Planning Commission, preparation of a Budget Handbook; but joint donor-government -systems improvement (e.g. work on revision of budget classification, support to meetings regularly note lack RPFB exercise). of sufficient high level involvement and Provided support to RPFB3 exercise. More ambitious effort to introduce more commitment. (But not clear analtical techniques, forward budgeting approach etc. failed to change the pattern that Budget division itself of Budget Division focus on the immediate issues oC preparing and implementing an was strongly committed to under-resources incremental line-item budget. Useful assistance with radical change. tcomputerization of budget preparation process. Co Table 5.1: Main Interventions Related of Planning and Resource Management Intervention/Dates Main Elements and Current Status Government Agencies Donors Involved Principally Involved Health and Education Budget Linked to agreement that EU counterpart funds should he used to ensure adequate Based in Ministry of European Union: two long term Improvement recurrent finance for basic health and education. Aim to improve analytical and Finance, but aiming also to advisers (one was in MoF 1993-1996 budget preparation capacity in concerned line ministries as well as Ministry of assist the Ministry of Budget Division, one in Local Finance. Intended to support RPFB process by deepening analysis of expenditures Health, the Education Government Support Unit, and strengthening mechanisms for the implementation of change. Ministries, and Local Prime Minister's Office), short Also supported establishment of Local Government Support Unit (LGSU) in the Government. (LGSU in term consultancies workshops Prime Minister's Office, to support improved planning and management of health Prime Minister's Office.) and training. and education funding in a number of pilot districts. Linking counterpart funding to budget improvement in this way proved unsatisfactory, since counterpart funds could be interrupted for macroeconomic reasons not directly related to budget management. Civil Service Reform Program Aims to redefine roles and functions of Government, control government Executive Secretary and Secretarial funded by World (CSRP) employment, improve pay and productivity and support decentralization throuigh Secretariat attached to Bank and UNDII. PCIS reform of local government. Six main components are: Civil Service Department. component funded by ODA. Formally launched 1992, drawing -Organizzlion and lfficiency (Q&E1) reviews ofe entral government ministries. Implications for all GOT Several donors contributing to on several precursor projects and -Development of Personnel Controland Ifrmation System (PCIS) agencies; central resource redeployment trust Fund. studies. -1ay reform agencies. CSD, PMO, -Capacity Building MoF, PlanComnm - seen as -Retrenchment & Redeployment prime movers. Steering -Local Government reforms Committee consist of PSs from these agencies plus Chief Secretary. I Ias presided over significant retrenchments (though mostly of lower grades of staff) ad recent restructuring of pay and allowances. Progress of Q&E reviews has been slowv, and most recommendations emerging have been rathier conservative. Lack of clear links to budget restructuring process and RPFB. Government Accounts noight elements identified by preparatory study Focused on Accountant SIA: I long termTA plus Development Project (GADP) -Objective I Capacity Building (including establishment of a Systems General's Department, short term consultancies; Development Unit, Implementation of management development program, revision Ministry of Finance, finance for local development Commenced October 1994, of financial legislation) team, training, equipment following earlier study phase -Objective 2: Systems Development in Accountant General's Department (includes Proposed senior chart of accounts and budget classification, public debt administration and management reference Project seen within a I10-year accounting system; salary system) group with MoF not yet perspective and additional -Objective 3: Establishment of Regional Sub-treasuries operational. donor support is being sought -Objective 4: Establishment of Central Payments and Expenditure Control system for components not funded by -Objective 5: Development of Ministry Accounting systems SIDA. ODA is supporting -Objective 6: Management Development Program subtreasuries and management -Objective 7: Dvlopmeniit of Revenue Accounting systems (superseded by development program. establishment of Tanzania Revenuc Authority) -Objective 8: Internal audit Table 5.1: Main Interventions Related o Planning and Resource Management Intervention/Dates Main Elements and Current Status Government Agencies Donors Involved Principally Involved C Initial phase of the project was intended primarily to support the first two Objectives plus the establishment of regional subtreasuries (Objective 3), but expanded to incorporate the establishment of central payments and expenditure control system. A finther objective: Strengthening of Debt Management has been added. Accountant-General's Dept. will take full responsibility within GOT for all operational aspects of debt administration, accounting and statistics. Auditor General Capacity building project in the Office of the Controller and Auditor General, OCAG ODA assisting in the implementation of the Department's Strategic Plan, prepared in Computers and vehicles Phase It of project commenced January 1993 Main aims are to: provided through World Bank August 1994 (Phase I was 1989- develop systems for auditing development contracts and introduce value for supported Finance and Legal 92) money auditing techniques, Sector Upgrading Project (FILMUP) . address problems in local government audit; . develop and introduce a strategic computerization plan; . Strengthen and develop senior management group and train a successor cadre; . Strengthen OCAG's in-house training capacity; . improve the electiveness of the annual CAG report to Parliament Aid Management and Concerned with all aspects of aid management and accounting. Based in Ministry of UNO1. initially two resident TA Coordination System .AMACS) Participative system reviews and studies in Phase I to lead to specific Finance, External Finance (CTA and inronnation systems recommendations for implementation during Phase II Seven working groups Division led by Assistant specialist); local staff, including (Now re-tilled Aid Management established to consider: commissioner (Aid computer trainer; local and andAlccouintability Project (AM'IAP) -infonnation technology issues related to aid management, Management). Project team overseas short term training -project appraisal, proposal and approval; (working groups) of workshops; equipment. March 1994 -development plan and budget, including government funding; approx. 30 officials from -resource mobilization; Treasury, Planning Under AMAP phase, long term -implementation monitoring and impact evaluation; Commission, BOT, CAG extemal TA phases out in favor -finance and debt management, aid accountability; and PMO. of a local management team (on -technical assistance and NGOs. leave from the public service). . Proposed development of aid Management Inlbrnation System (AMIS) which Steering commite drawn would link to project (PROMIS) and debt (DRMS) databases in Planning from Finance, PlanComm, Bilateral funding has been Commission, MoF and Bank of Tanzania BOT and CSD. secured, at least in principle Under the AMAP agreement, this program is now intended to share an inter-agency proposed AMAP components. steering committee with other resource management programs Revenue Strengthening Tanzania Revenue Authority became operational July 1996. The TRA will incorporate Ministry of Support from World Bank, the existing revenue departments within an autonomous structure outside of the core Finance/Revenue Authority ODA, Danida and USAID. Decisive action on TRA began in civil service. This is intended to allow better resourcing and management of the 1995. Other initiatives have much exeputive functions of revenue collection. World Bank, EU and UNDP longer history. have been supporting VAT,l In addition, a number of revenue strengthening measures which preceded theSTRA Income Tax Computerization decision are ongoing. These include preparations for introducion ofmVAM, and ASYCUDA withTA eoinputcrizalion of the income tax system, and adoption of the ASYCUDA system o v equipment. fmaintaining customs records. Table 5.1: Main Interventions Related o Planning and Resource Management OTHER RELEVANT INTERVENTIONS AND ACTIVITIES oo A number of other activities and programs also have a bearing on resource management reforms, including: * Other Major Elements of the Reform Program, including parastatal sector relorm and the relbrm of financial institutions, have a direct bearing on Government 's overall role, its immediate and long term budgetary obligations, and the appropriate roles and functions of sector ministries. * Bureau of Statistics: SIDA has been principal source of institutional support to BOS for many years, but it is intended to phase this out by 1999. Various other agencies have assisted with specific elements of statistics collection and analysis (e.g WB has funded The National Accounts Strengthening Project under the PPRP Project) * Resource Management in Sector Ministries. Development programs with institutional and resource management components are under way in a number of key sector ministries (e.g. ASMP in Agriculture and the IRP in Transport) * Policy and Sector Studies and Reviews- Government has undertaken a number of important policy and reform studies (including the 1994 Mramba Commission) that provide analytical support for resource management reform. Government and donors have collaborated on a number of sector studies and reviews (e g preparatory work for the social Sector Strategy). * Other Support to Resource Management Agencies: This paper does not attempt a comprehensive review of all donor support to the main resource management agencies Various other capacity-building projects and projects linked to specific aspects of planing/ policy/executive work has been implemented or are ongoing (e g. Commonwealth Secretariat assistance to MOF and BOT in developing Debt Recording and Management System, UNDP's SDA project in the Planning Commission, etc.). The IMF in late 1996 undertook review of information systems requirements in the Ministry of Finance. tA Chapter 5 85 Implication for Aid Management 5.44 Despite Government's objective of reducing dependence on aid, aid continues to underpin public expenditures. The relationship between Tanzania and its aid partners has often been suboptimal on both sides, as is well described in the Report of the Group of Independent Advisers on Development Cooperation Issues between Tanzania and its Donors (the Helleiner report). Aid flows and project management structures, have often bypassed regular government structures in an attempt to avoid the weaknesses of public administration and expenditure management. However, these independent actions have tended to perpetuate those weaknesses and inhibit strategic national and sector planning and expenditure programming. This is a situation that needs to be corrected. 5.45 The approach to public expenditure management set out in this PER requires Government and donors to collaborate in ensuring that: * Government takes the lead in planning, prioritization and expenditure programming leadership of the institutional reform process; * a coordinated CSRP/RPFB exercise provides the framework for articulating national priorities and programs, both domestically and in relations with the donor community; * a sector-wide approach to planning and aid coordination is increasingly adopted, with donors collaborating to support coherent sector programs that address both the recurrent expenditure and the investment requirements of each sector as a whole; * Government is more proactive in liaising with donors both during budget preparation and during budget implementation to ensure that aid flows are anticipated and captured; * aid agencies take seriously their commitment to allow Government to take the lead, and respect the priorities Government adopts; * both Government and aid agencies as far as possible avoid project management arrangements that bypass existing organizational structures; and * a coordinated budget calendar is developed so that necessary Government dialogue with the international financial agencies is integrated with the national expenditure programming cycle, and not superimposed on it.  6. BUDGET IMPLEMENTATION ISSUES 6.1 Previous Chapters have reviewed recent trends in public expenditures and their current configuration. In this Chapter we review budget implementation issues. As indicated by the review of recent public expenditure outturns, through FY96 there has been a large gap between the budgeted allocations of resources at the beginning of the fiscal year and what agencies have actually received by the end of the fiscal year. Also, the sum of discrepancies at agency level has been considerably greater than the discrepancy at aggregate level (Table 4.2). Thus, there is not a simple sequence of deciding allocations during budget preparation and then executing the pre-determined budget: in practice, there are many allocation decisions in-year, and there is considerable uncertainty about how the budget will turn out both for the central agencies, that are trying to control the aggregates, and for implementing agencies, which cannot be sure whether or when they will receive the resources budgeted. These uncertainties themselves create inefficiency in the use of resources: it is difficult for agencies to plan the efficient use of resources when they do not know what resources they will have next month, let alone next year. And, because not all components of expenditure are equally flexible in the short term, adjustments that have to be made suddenly are more damaging than those which can be anticipated. As noted above, short term adjustments have fallen largely on development expenditures and other charges. 6.2 In the subsections which follow, three interrelated aspects of budget implementation are important: the ability to anticipate expenditures (and revenues), the ability to monitor the use of resources as the budget is implemented, and the ability to control both the level and the composition of expenditures. This Chapter contains three sections, reflecting the major concerns with implementation identified in Chapter 2: * cash management: an issue that emerges is the need to improve both the controls and the information on which budget implementation is based, so as to realize macro control in the least disruptive way; * implementation issues related to the wage bill; * other issues of control and information during implementation; A. CASH MANAGEMENT 6.3 Chapter 2 noted the much improved performance in aggregate expenditure control beginning in 1996 and the role of the cash management system in restoring aggregate control. Because of the underlying imbalance between what the budget was attempting to finance and what Government's resources can actually finance, the restoration of control in 1996 has been painful, and there has been a tendency to blame the instrument (cash management), and to hope that it can be a temporary expedient. In this section we: 88 Chapter 6 * describe how cash management has operated and argue that cash management is a necessary part of any Government's responsible fiscal strategy; * suggest ways in which, in the short and medium-term, the cash management system can be improved, and in doing so become less conspicuous and disruptive. Objectives of Cash Management 6.4 Before looking at what improvements might be desirable in the existing system of expenditure control, it is important to see the use of cash management in the context of the wider problems of macroeconomic management. Stabilization policy in Tanzania has relatively few effective instruments available. In the medium term, successful stabilization depends above all else on adhering to a sustainable fiscal stance. But even over the very short term, stabilization policy must also rest very heavily on fiscal adjustment. In view of the lack of financial depth in the economy and the weak condition of the financial sector, only limited contribution can be made by monetary instruments. Development of monetary instruments itself imposes fiscal costs, where for instance Treasury bill sales are expanded in conditions where markets are thin and confidence in price stability beyond the very short term is very limited and only likely to improve very slowly. 6.5 Fiscal adjustment within year need not, in principle, rely exclusively on expenditure reductions. Turning to the revenue side, frequent changes in tax rates impose even more severe penalties in disruption to private sector planning and confidence. However, a single mid-year tax adjustment could be a more efficient and less disruptive instrument than short- term changes in expenditure. This could be feasible with a tax, such as a VAT if successfully implemented, of broad incidence and high enough aggregate yield for a modest rate adjustment to generate significant macroeconomic adjustment. At present, such an instrument is not available and in any case the principal constraint on revenue performance lies primarily with tax administration, improvements in which are inevitably gradual and not easily predicted. In these circumstances, Government has no alternative but to continue to rely on short term expenditure adjustment as the principal instrument of stabilization within year, in spite of the costs that this necessarily imposes on expenditure planning. How then can these costs be minimized? 6.6 In general, the requirements of a cash management system as a tool of economic management are to respond to macroeconomic shocks in a manner which ensures (1) maximum adherence to macroeconomic objectives of price and exchange rate stability, and (2) minimum volatility in release of resources and maximum extension of the horizon of firm available resources for planning by budget managers. Box 6.1 below describes the contribution of cash management to macroeconomic management in Uganda. Chapter 6 89 6.7 In the medium term, the most efficient strategy to deal with shocks is to eliminate them as far as possible through better monitoring, forecasting and planning. A cash management system should not be a substitute for such a strategy. As emphasized throughout this review, high priority needs to be given to improvements in forecasting and planning in Tanzania, both for the macroeconomic framework and expenditure planning and monitoring at sector level. However, better forecasting and planning systems take time to develop, and macroeconomic objectives have to be met as well as possible in the interim. 6.8 It is helpful to distinguish three categories of possible shocks: * revenue shocks (domestic revenue, aid) * expenditure shocks (unforeseen spending requirements) * balance of payments shocks A high level of dependence on aid will entail greater difficulties in forecasting total revenue. Domestic revenue tends to be more difficult to predict where, as in the case of Tanzania, Government is attempting to achieve a rapid increase in the aggregate volume of domestic revenue. Where this rests in part on administrative improvements, the forecasting problems are more acute. External shocks may adversely impact on the budget, not only directly through effects on domestic revenue but, also through unwelcome exchange rate developments which the authorities find themselves unable to correct without recourse to fiscal adjustment. 6.9 In Tanzania's circumstances, it must be recognized that disruption to expenditure planning will be inevitable during the transition to more rigorous expenditure control. Budget managers have operated for a long time without a hard budget constraint in practice and there will need to be a period during which expectations are revised and expenditure commitments are adjusted to the real level of resources available. 6.10 In Tanzania, significant over-expenditure has occurred in the form of commitments unmatched by available cash resources, and resulted in a growing volume of domestic arrears by the end of FY96. The need to settle such arrears within a cash limited budget at the expense of budgeted expenditure plans represents a further distortion of budget allocations and further delays improvement in expenditure planning. Correction of such distortions will depend not only on revising budgets, but also correction of budget managers' expectations that additional resources can always be found by supplementary expenditure or unauthorized commitments. As discussed below, such adjustment can be accelerated by steps towards better monitoring at the commitment stage, tighter controls on supplementary expenditure and disciplinary action on unauthorized spending. 90 Chapter 6 Box 6.1: Uganda's Experience with Cash Management as a Tool of Macroeconomic Management In early 1992, the Government of Uganda responded to inflation in excess of 60 percent with a major initiative to bring public expenditure under control and establish macroeconomic stability. President Museveni gave strong personal support to the initiative making public his determination that Government should live within its means. Senior management in the newly merged Ministry of Finance and Economic Planning reorganized the structure of economic management, mobilizing a strong macroeconomic policy department to spearhead the effort to monitor macroeconomic performance and identify the fiscal adjustments required to achieve and maintain macroeconomic stability on a continuous basis. A key instrument which was developed within weeks of the institutional rearrangements was the Government Cashflow, a twelve month spreadsheet projection of central Government operations, coordinated and produced each month by the Macroeconomic Policy Department (MEPD). A Cashflow Committee was appointed to review macroeconomic outturn and recommend a revised cashflow framework each month, which is submitted to senior management in the Ministry of Finance as the basis for decisions on expenditure releases during the subsequent month. The Cashflow Committee is chaired by MEPD but includes in its membership the Treasury Office of Accounts, the departments of Aid Coordination, Expenditure, Economic Planning and Tax Policy and representatives of Uganda Computer Services. The Ministry of Public Service is represented, covering issues of wage bill implementation and projections. Prominent in the membership are representatives of the Bank of Uganda, especially the Research Department which normally presents a review of the latest outturn on inflation, the exchange rate and monetary conditions as a prelude to discussion of proposed revisions to the cashflow. The Cashflow Committee has emerged as the most regular forum for coordination of monetary and fiscal policy at the technical level, along with the Balance of Payments Group, an MOF/BOU committee responsible for jointly examining balance of payments projections and reviewing forecasts for aid inflows each month prior to the revision of the cashflow projections. Following a rapid reduction in inflation to single digits by late 1992, the cashflow management system has not normally been required to deliver such drastic corrections in expenditure as were implemented in the final quarter of FY92 This has permitted a gradual increase in the predictability of resource flows and a steady extension of the expenditure planning horizon as smaller and less frequent adjustments are required in projected expenditure. Nevertheless, external shocks such as the coffee boom and surges in private capital inflows have required significant within year adjustment of aggregate expenditure targets to maintain price and exchange rate stability during the period, designed by MEPD in consultation with the Cashflow Committee. In each of the three years FY93-FY95, response to within year external shocks has reduced aggregate expenditure below annual program targets, even though domestic revenue performance has performed largely in line with program during the period. Much larger shocks are attributable to within year over-expenditure in some sectors and consequent reallocation. In response to these pressures, the cash management system has recognized certain Strategic Areas of priority expenditure, including primary health, primary education , road maintenance, rural water and some expenditures on police and judiciary. The system is widely recognized in Uganda as central to the achievement and maintenance of price and exchange rate stability during the past four years. Several positive aspects of the system can be identified: * Leadership from a strong macroeconomic policy team ensures that budget implementation is continuously responsive to economic policy objectives rather than uncritical implementation of an annual program; * The wide membership of the Cashflow Committee provides an effective means of coordinating and validating data with key informants and holding relevant members accountable for forecasting or record keeping failures. It also spreads understanding and ownership of the sometimes difficult adjustments required by macroeconomic management and provides an opportunity for the costs of adjustment to be voiced and taken into account. The use of monthly releases in a twelve month framework (itself linked to a three year macroeconomic framework) has provided the means to gradually extend the predictable horizon of expenditure planning for budget managers while retaining the tools for rapid macroeconomic adjustment when required. Chapter 6 91 6.11 There are some lessons to be drawn for the next steps in Tanzania's efforts to regain control of aggregate expenditure: * For the foreseeable future, given the limited availability of other instruments of stabilization in Tanzania at present, a system capable of delivering short- term adjustments in expenditure will be needed to meet the macroeconomic objectives of price and exchange rate stability; * The existing cash budget system has been the main instrument of improved fiscal and macroeconomic performance during 1996. However, it has a number of weaknesses, that Government has been addressing in the course of FY96. These may be summarized as follows: - the system rests on mechanical adherence to annual targets with little regard to macroeconomic performance or the need for adjustment so as to meet macro objectives; - the current system initially generated expenditure patterns different from budgeted levels without an adequately transparent or accountable process for determining revised allocations and protecting priority areas of expenditure. Directions for Improvement of Cash Management 6.12 These conclusions lead to some proposals for strengthening the cash management system in Tanzania: * Extend the requirements planning horizon: Improving agency estimates of their monthly requirements over the budget cycle would facilitate planning of the present system of cash releases. While releases should be maintained on a monthly basis, a year's horizon for requirements would give the option of more gradual adjustment and more warning of impending constraints to budget managers; * Strengthen revenue forecasts: One of the major reasons for the move to a cash management system with a short planning horizon lay in the volatility in revenue receipts. Predictability of revenue flows appears to be improving, though the track record is too short to be very confident. Further improvements will help reduce the recent volatility in expenditure flows; * Extending the scope of cashflow analysis The present system essentially responds only to shocks arising from shortfalls in domestic revenue, by setting an aggregate cash limit based on aggregate domestic revenue. The cash management framework would be strengthened by introducing monthly analyses of the annual projections for central Government operations, capturing monthly flows for domestic and other revenue (including aid), the main components of expenditure and domestic financing and projecting such monthly flows a year ahead. As in Uganda, with appropriate checks and balances, revisions to the framework could then be made as necessary on a 92 Chapter 6 monthly basis, including required revisions to expenditure ceilings. As the major short term instrument of macroeconomic management, setting of cash limits should be sensitive to outturn on stabilization objectives, especially inflation and the exchange rate; * Better forecasting on key expenditure commitments: In FY97, cash management has been affected by problems in reliably estimating the monthly wage bill and debt service. Wage bill estimates do not adequately capture recruitment or attrition ( or at best do so with long lags), nor is there adequate control and budgeting for promotions; * Protected expenditure : social and economic infrastructure, law and order, etc. Implementation of expenditure ceilings as required under cash budgeting should observe priority areas in the budget, which should be protected as far as possible. These priority areas should be identified explicitly by budget item code and should include key components of spending on social and economic infrastructure and law and order. Since the recognized priority areas will in effect represent the actually implemented budget in conditions of expenditure tightening, it is important that the priorities are transparently based on national priorities, preferably subject to Cabinet scrutiny. B. IMPLEMENTATION ISSUES RELATED TO THE WAGE BILL 6.13 The wage bill has been the focus of considerable attention because of the perception of an overstaffed, unaffordable public service as one of the main drivers of excessive public expenditure and hence, the focus of ESAF to limit the total wage bill. Chapter 3 above addressed the longer term issues of control of public service employment and pay - trying to return to a pay structure that is adequate, efficient and transparent. However, in addition to the problem of longer term control of the wage bill, there is also a serious problem with its short-term volatility. With the introduction of the cash management system, the shortcomings in Government's ability to estimate PE obligations has become more apparent. For instance, despite a declining workforce, PE expenditures in the first months of FY97 continued to increase. Analysis of the data shows that the increase stems from arrears, promotions and the underestimation of allowances. In view of these difficulties, the PER took a special look at wage bill issues. The main concerns center on the budget cycle, payroll administration and within year monitoring of the payroll, and the promotions process. The budget cycle 6.14 At the beginning of the annual budget cycle, Ministries agree a Schedule of Emoluments with CSD. This Schedule includes data on staff-at-strength (produced by the Government Computer Services), requests for increases in staff, proposed promotions and the required financial provision. Prior to the freeze on recruitment in FY1992/93, requests for staff increases would have been compared with Establishment lists (i.e. lists Chapter 6 93 detailing the number and deployment of staff for each vote if it were operating at full strength). If the requesting Vote was operating below Establishment level, the practice was to allow increases. Although promotions are supposed to be proposed with the Schedule and their cost incorporated in the budget, in practice the promotions process occurs after the beginning of the budget year, and hence are a major cause of uncertainty in the wage bill. For instance, in FY97, the Teaching Service Commission alone has submitted 10,000 requests for promotions, without any attempt to establish their additional costs to the wage bill. 6.15 There are two major problems with the Schedule: the integrity of the payroll data and the uncertainty introduced by the promotions process. With regard to the payroll data, and despite numerous attempts to clean the Establishment data base, ghost workers remain on the payroll. There is also a considerable body of evidence that payroll- generated data contain material inaccuracies. For example, the March 1997 Report on the Special Payroll Audit prepared by the OCAG for 4 ministries and 6 districts for March 1996 found payroll irregularities of approximately T Sh 314 million plus T Sh 18 million in payments to ex-employees. Similarly, a 1996 audit of five Votes (headcount 11,086) by the Office of the Auditor General (OCAG) supported by the Personnel Management Information Unit (PMIU) revealed 137 duplicates. These were members of staff assigned a single check number but drawing multiple salaries under different sub-Votes. It is believed the scale of the abuse is larger than this because no account was taken of staff fraudulently using multiple check numbers. It is significant that this incident was detected using computer-assisted audit tools designed by PMIU. This means that the Government is therefore in a position to minimise payroll fraud and improve data integrity by institutionalising the payroll audit. There is also evidence of weak links between the providers and users of data, such that it would be useful to allow payroll information users to have an input in the design of the new payroll system and the audit program. 6.16 The PMIU is in the process of compiling an up-to-date, computerized staff data base, known as the Personnel Control Information System. Preparation of this Establishment base however, has been delayed by inadequate resources. There are also plans for a comprehensive payroll audit by the OCAG in collaboration with the PMIU. Both of these initiatives are essential to strengthening the integrity of the wage bill and facilitating the monthly estimation of payroll: * It is recommended that these initiatives by provided priority resource allocation so as to permit acceleration of implementation. Currently, the new payroll software would not be ready for implementation until late 1998. 6.17 The recruitment and promotion of Civil Servants is the responsibility of statutory appointing bodies, namely: the Office of the President in the case of the most senior cadre; the Civil Service Commission (CSC) for middle training civil servants (approximately 150,000 officials); the Teaching Service Commission; the Police & Prisons Commissions; the Judicial Commission; and the Local Government Commission. 94 Chapter 6 In addition, each ministry has a committee (KAMUS) attached to it with responsibility for recruiting support staff. The process of establishing posts, however, is as follows: the creation of posts is the responsibility of CSD, while the funding of posts is the function of the Treasury and recruitment and promotions are carried out by CSC. On the issue of need for posts, CSD is meant to carry out annual reviews of Establishment in order to keep in line with current Government objectives. This practice has lapsed with the freeze on recruitment, but should resume with the completion of new Establishments with Phase II of the O& E reviews. In the past, Treasury has funded all posts created, without reference to budgetary limits. 6.18 The promotion process requires line ministries to compile ahead of the budget submissions a Seniority List (SL), in response to a call circular from the appointing authority. The SL is based on the employees Scheme of Service and officials are eligible for promotion three years after their first posting and every third year after that. Line ministry decisions to promote are meant to be based on need and the existence of a post. The relevant Commissions review the performance appraisals of the candidates prior to approval. The practice of carrying out appraisals has become largely erratic and promotions are virtually automatic. A major weakness of the system is that appointing authorities are not required to verify the availability of funds for promotions but merely assume there are funds to cover all requests for promotions received. Under their empowering statutes the Commissions Civil Service Schedules of Appointment are not required to take account of budgetary provisions. The basis for promotions and their related cost to Government is therefore weak, provides little scope for budget constraints to come into play and does not represent value for money; rather promotions are a means to enhance pay. 6.19 In an effort to contain the FY97 wage bill, Treasury is withholding funds for payment of promotions and it is not known when these arrears will impact on the payroll. Government, however, is aware of the need to develop a long term solution to the volatility generated from the promotions process. It is proposed that in the near term: * discussions be initiated to review the statutory promotions process with the Chief Secretary, Principal Secretary, CSD and the Secretaries of the Commissions. Major issues will be the entitlement of staff to promotion after given periods of service as opposed to merit; the authority of appointing commissions to promote staff without consideration to budgetary constraints; and, the desirability of linking promotions to available funds; * a policy decision be made on whether Government can afford to promote officials every three years or whether after three years an official should merely become eligible for promotion if new vacancies exist at a level senior to his current level and whether the promotion can be borne by the budget; Chapter 6 95 * if the decision is taken to make the possibility of promotion dependent on availability of vacancies, Schedules of Appointment should be amended accordingly; * submissions for promotions should be accompanied by an analysis of the impact of the promotions on the budget and should be reviewed by the PCU , CSD and Budget Division, Treasury prior to final approval; * consideration should be given to introducing a promotions freeze for FY98, except for key positions where added value can be demonstrated. 6.20 Further, with regard to the Civil Service Commission , a recent review of the Commission was carried out under CSRP. The review noted numerous administrative reasons for the poor performance of the Commission and its inability to meet budget deadlines, including its lack of manpower and equipment. In particular, the review noted the lack of capacity to process the promotions data with existing manual systems. The report outlined plans for the computerization and strengthening of the Commission's management and recommended that the CSC be made ultimately responsible for all Civil Servants . * Early consideration of these proposals would assist the Government in regaining control of the wage bill. Payroll Administration 6.21 During the monthly cycle, line ministries update their individual payrolls using data sheets. Data sheets are documents on which staff particulars, including emoluments due are recorded. They are filled in by line ministry administration staff. Updates include recruitments, promotions, increments, changed pay stations and other data. The updated payroll should be submitted to PCU/CSD for verification, however, it appears that not all vote-holders adhere to this new requirement introduced under PCIS. On receipt of the verified payrolls, Treasury makes an Exchequer Issue (Release of Funds) to each Vote. The Votes then allocate the Issue amongst their sub-Votes by means of Warrants of Funds. These Warrants of Funds would be expected to tally with the payrolls of the sub-Votes to whom they are made. 6.22 However, actual PE payments deviate from targets and/or fluctuate for a variety of reasons, including: inconsistencies in staff numbers and outgoings on allowances; the lack of comprehensive manpower budgeting; poor payroll management (including administrative delays, late data entry and coding errors), and insufficient management information which would allow key players to take within-year corrective action and mid- term adjustments. Early implementation of the PCIS and a Government wide payroll audit, will help to reduce the uncertainty in the monthly payroll. Simultaneously, the introduction of cost center or program based budgeting linking staff (through the establishment database) to activities with funding allocated first to priority programs would assist Government in regaining control of the wage bill. These changes would also facilitate within-year performance reviews which could help isolate problems in 96 Chapter 6 implementing the budget eliminating the need for hastily planned within-year cuts and/or retrenchment in order to remain within expenditure limits. C. OTHER BUDGET CONTROL AND IMPLEMENTATION ISSUES 6.23 The main report has addressed in detail the commitment of Government since early 1996 to expenditure control and the difficulties inherent within the current systems to estimate, monitor and account for monthly expenditures on key aggregates and especially, debt and wages. In recent years, this has contributed during budget implementation to a serious underfunding of other charges and the Development Budget. This section addresses in some detail the volatility stemming from both the system used to record and forecast debt, and the expenditure control and information systems. The section assesses the extent to which these systems impact on the Government's ability to control the budget and ensure effective financial management and makes recommendations for strengthening these systems. Debt Management 6.24 The debt monitoring and forecasting system is largely dysfunctional. As a consequence debt allocations, which are provided under the Consolidated Fund Services vote, are mainly derived from demand notices received from lenders, statements received from the Bank of Tanzania's Commonwealth Debt System and the previous year's payments in the case of foreign loans. A major limitation of the system is that data on the current debt management system is incomplete, making it difficult to estimate repayments, and as a consequence undermines the Government's planning resource management functions. The system has also handicapped Government in its debt rescheduling efforts, in particular, with the Paris Club. 6.25 To address this issue, on November 1, 1996 a task force was appointed by the President to take stock of public debt, including both domestic debt (bonds, guarantees to parastatals and treasury bills) and external debt and to recommend a debt strategy. A preliminary report was prepared in December 1996. The approach adopted is to obtain data from the Bank of Tanzania's CS DMRS system and validate the data against the Treasury's records. The stock taking exercise on Government's liability for parastatal obligations (Chapter 2) will also be an important step in increasing transparency and will provide the basis for Government to formulate a policy for redressing these obligations. Once completed, the debt data base will be accessible to all agencies with a need to use debt data. The task force is expected to finalize its report to the President in late-1997. 6.26 Following the establishment of the 1996 Debt Task Force, a program has been developed aimed at strengthening the AGD's ability to manage debt through an institutional strengthening project, supported inter alia by SIDA, the World Bank and MEFMI. The Government's recent actions to strengthen debt management are sound and should lead to a substantive improvement in debt management, budget resource planning, Chapter 6 97 and also support debt strategy formulation. The new system should be fully functional by December 1997. Disbursement Systems 6.27 The Government's efforts to regain control over expenditure since early 1996 have been aided by progress realized in several on-going, donor-supported activities. These efforts are not just to contain the level of expenditures and prevent arrears, but also to tighten controls to ensure that funds are spent as authorized and not informally reallocated. These include the programs to strengthen the overall cash flow through centralizing the management of Paymaster-General accounts at the Treasury, the Sub- Treasuries project, the development of a more detailed budget classification and various other institutional strengthening activities. These activities have already made over the past year (and will continue to make) a significant contribution to the level of accountability that the Government can provide. However, the success of initiatives will depend on (1) the Government's commitment to continue implementation and to enforce the new systems after years when formal rules and regulations have been disregarded with impunity and (2) capacity building in Treasury, especially the Accountant General's Department, in which the Government Accounting Development Project (GADP) has been a vital catalyst. 6.28 Until June 1996, the Government maintained an antiquated Exchequer system utilizing an estimated 200 plus Paymaster-General (PMG) accounts at the Bank of Tanzania. Disbursements through the PMG accounts had resulted in a virtual collapse of the Government's system of budgetary control. As accounting officers failed to monitor commitments and were not held accountable for unauthorized expenditures or late submission of accounts. Central Payments System and Sub-Treasuries 6.29 Thus, the Government in 1994 decided to tighten controls through the development and subsequent introduction of the Central Payments System (CPS) and the Sub-Treasuries, which aim respectively to control spending agencies' accounts centrally at Treasury and in each region at a Sub-Treasury level. These approaches are sound and already impacting positively, but it is not possible to reverse overnight all the weaknesses of the old system. The CPS was introduced in mid-1996 and as of January 1, 1997, 15 ministries' recurrent and development expenditure accounts were held on the Central Payments System. Plans are to increase coverage to all ministries by 1 July 1997 (except Defense), and to eliminate direct disbursements to Paymaster General Accounts controlled by the ministries. The Government has also established 9 Regional Sub- Treasuries which control the accounts of the Regional Administration Secretariats (RASs), Local Authorities and Ministerial Offices based in Regions, District Councils and Commissions. The Government plans to expand the number of Sub-Treasuries to 20, so as to provide coverage nationally. 98 Chapter 6 6.30 Some of the benefits of the new arrangements are as follows: * Government is able to ensure that the implementing agencies do not overdraw May, on accounts maintained by the CPS and Sub-Treasuries; * a new, 7 tier budget classification was introduced in FY97 and this provides a significantly higher level of desegregation on expenditure. Government will be able to report on expenditure by: - object code or vote; - program; - sub-vote; - cost center or category of support; - type of transaction. For example, payment or receipt or asset or liability; - project code; - sector code; * timely reporting: monthly flash reports for the fifteen ministries are now available and, on request, more frequently; * simplification: to date, more than 120 of the 200 PMG accounts in BOT have been closed. 6.31 The CPS system is an interim system, adopted from an in-house package previously running at the Ministry of Education and Culture. Over the next two years, the Government plans to adopt purpose-designed systems that:- * are commitments based rather than cash based. This will enable the Government to capture data on liabilities incurred by spending agencies. This is particularly important, for instance for bills from suppliers of goods and services (who are owed an estimated total amount of T Sh 50 billion) and also for the wage bill (arrears and allowances); * are integrated to other books of account. For example, vote books, payroll and planning models; * account for all Government expenditure accounts. For example, miscellaneous deposit accounts and fund accounts are still controlled by ministries outside of the CPS and may be subject to abuse; * distinguish between payments made in the current financial year and in previous years. This would be in line with best practice which advocates matching of expenditures; * are flexible enough to upgrade. * record revenues; 6.32 Further development of the CPS system is currently constrained by resources and by staff capacity within the AGD. Over and above funding to the GADP (provided by SIDA), it is estimated that total costs of implementing the new system in its entirety would amount to approximately $1.5-2.0 million. Although the AGD invests heavily in Chapter 6 99 the development of finance staff, the Government is unable to retain them once they have completed training as current remuneration is considerably lower than incentives offered in the private sector. As a result, the number of well trained staff within the AGD is low. The CPS project is well-conceived and will considerably strengthen Government's budget monitoring, forecasting and accountability capacity. Efforts at accelerating implementation should be supported. 6.33 It is proposed that by the beginning of FY98 the Government should:- determine its position on the future of Sub-Treasuries, as there is uncertainty over their role if the Government goes ahead with its plans to decentralize financial management to a Local Government level. The Mission view is that, regardless of where the Sub-Treasuries are located, the function is essential and that the Sub-Treasuries are therefore needed and should be expanded nation wide; * review pay structures of accounting cadres under the pay reform initiative of the CSRP with a view to retain accounting staff once they have undergone training; * develop an organization structure for the accounting cadre for when the CPS is fully operational in 1997/98, so that accounting staff are utilized effectively; * the Government together with donors should consider ways to support accelerating the implementation of the new systems. In particular, the implementation of commitments and revenue systems, and cost accounting systems as part of a general ledger module. Accounts outside the CPS 6.34 To further strengthen Government's control over expenditures, a policy decision is required in the short-term on the future treatment of accounts outside the CPS, such as miscellaneous deposits and fund accounts. As it is the intention of the Government to regulate or close such accounts, it is recommended that in the beginning of FY98:- * a Task Force be established to take stock of all such accounts; * a schedule of accounts be prepared including a description of their purpose and amounts held; * the AGD, together with Accounting Officers and representatives from OCAG agree on which accounts to close and ways in which the remainder may be regulated; * the AGD prepare guidelines detailing future arrangements. 100 Chapter 6 Accounting for donor funds 6.35 The strengthening of the Government's disbursement system does not directly address the issue of donor funds. The Government's forecasting process for aid inflows remains weak and amounts recorded for donor funding received are incomplete. For instance, a recent audit by NORAD showed that only 15 percent of their funds disbursed in FY96 were recorded. A number of donors although skeptical about the Government's ability to provide transparency and be accountable for funds, appreciate that a separate disbursement mechanism contributes to undermining Tanzania's ability to plan and manage her own development effectively. 6.36 To consolidate efforts being made in implementing agencies, an Aid Management and Accountability Program (AMAP) has been developing since 1994 an integrated approach to the planning of development expenditure and management of donor funds. This involves players from Treasury, Planning Commission, Bank of Tanzania and CSD. AMAP is expected to run over a further three year period. At the time of the PER mission, the AMAP team had prepared a number of papers, including a program document detailing short and longer term activities together with resource requirements. However, a 3 year time frame for AMAP implementation is too long and it is recommended that the project objectives be refocused and accelerated, ideally with an operational system in place for Treasury by mid-FY98. Accounting for user charges 6.37 Some spending agencies are now allowed to maintain revenues collected for services rendered to fund other charges required to run their agencies. This has the advantage that agencies have an incentive to meet expenditures and hence help to reduce demands on other charges in the budget. However, there is no system to report collections and returns and year end appropriation accounts do not fully reflect amounts collected or spent from fees. To ensure that Government effectively monitors funds collected and their subsequent use, it is recommended that a standardized reporting system be introduced effective the beginning of FY98. Finally, it is recommended that a tracking study of public expenditure in primary education and health be undertaken for recent years in a sample of districts in FY98 with a view to establishing an effective expenditure tracking system for FY99. Monitoring of aggregate revenue and expenditure 6.38 Strengths of the current monitoring processes are that a relatively simple model is used to consolidate the Government's position on a monthly basis of which a consolidated flash expenditure report can be used to highlight over-expenditure by Accounting Officers. Consolidation reports are normally available for circulation to managers in the Treasury by the end of the following month. Limitations are as follows: Chapter 6 101 - supervisory controls at spending agencies are generally weak and as a result monthly reports are likely to be inaccurate or submitted late. Further, no disciplinary action appears to be taken for non-compliance with Financial Orders; - expenditure amounts reported in the flash reports are based on cash book entries which are sometimes not up to date and as a result it is impossible for the Government to detect over expenditure or estimate the effect of unrecorded cheques (check floats) on money supply; - monthly flash expenditure reports do not account for commitments as returns are based on cash outflows. This has resulted in the Government incurring bills from suppliers estimated at T Sh 50 billion and T Sh 10 billion for recurrent and development expenditures respectively; - monthly flash expenditure reports do not provide reasons for any adverse or favorable variances; consolidated flash expenditure reports are not always accurate as Exchequer Issues amounts are incomplete. For example, in the October 1996 flash report total Exchequer Issues are stated as T Sh 171 billion. Mission computations show T Sh 196 billion worth of issues were made during the same period; - revenue receipts at Treasury are subject to considerable delay as revenues are first collected and deposited in TRA, RASs and Sub-Treasury accounts held with the National Bank of Commerce and then subsequently transferred to the PMG accounts and finally to the Exchequer Account. It is therefore necessary for the Government to track flows by reconciling flash reports to amounts deposited. However, this activity is currently not practiced and as a result, the Government is not able to monitor inflows and react to large variances in the revenue float that may arise. For example, the June 1996 flash report indicated revenues collected in 1995/96 as T Sh 383,743 million whereas amounts deposited in the revenue account at Treasury during the same period are shown on bank statements as T Sh 368,062 million, implying a revenue float of T Sh 15,681 million; - revenue arrears are not reported in flash reports; - few returns are submitted by ministries to support non-tax revenues collected and/or retained to fund other charges. For example, at the time of our review, over half the district hospitals had still not submitted year- end returns for the 1995/96 financial year; - the Internal Audit Function is generally perceived to be ineffective and its ability to monitor aggregate revenue and expenditure currently depends on individuals; - Accounting Officers rarely provide the OCAG with explanations for audit queries raised on a timely basis and often responses are only received just before submission of the audit report to the Public Accounts Committee. We understand that the OCAG is currently looking at a mechanism for obtaining responses to audit queries on the spot. Further, there is a two year backlog on Exchequer account reconciliations which means that the 102 Chapter 6 Government's consolidated accounts cannot be finalized and as a consequence, accountability is undermined. Further, there is a backlog on Exchequer account reconciliations which means that the Government's consolidated accounts cannot be finalized and as a consequence, accountability is undermined. Actions being taken 6.39 A special Task force consisting of accounting staff at spending agencies and headed by the Assistant Accountant-General in charge of financial management has been established to prepare the backlog of bank reconciliations on the Exchequer accounts. They have completed the 1994/95 Exchequer account reconciliations and are currently working on the reconciliation of the 1995/96 accounts. This entails a considerable improvement in completion of reconciliation over earlier years. 6.40 In addition, on 1 July 1996 the Government introduced a computerized Exchequer management system which had been developed internally by the Systems Development Unit at Treasury. The aim of the system is to provide a tool for Government to manage its cash flows. The system currently records all Exchequer issues, and some of the revenues collected including treasury bill receipts provided by the Bank of Tanzania. The system is not yet able to provide a net cash position. However, Treasury expects that the system will be fully operational by mid-1997 and will provide net cash flows, trend analysis and cash forecasts. Recommendations for Improved Reporting and Compliance on Aggregate Revenue and Expenditure 6.41 It is proposed that to improve reporting: * Personnel within the Treasury be required beginning immediately, to highlight variances on expenditure. The AGD and Commissioner of Budgets should then request spending agencies to provide explanations for any variances on expenditure each month. This would also be useful information when allocating monthly expenditure. To improve compliance: * The AGD needs to take stronger measures to ensure that monthly returns and any explanations that may be required are received on a timely basis. * Development of an internal audit strengthening program be given immediate consideration; * Government financial regulations (page 126 of the Financial Orders) permits the Public Accounts Committee to penalize officers who fail to comply with Chapter 6 103 their duties. However, these punitive measures are rarely implemented and when Government has done so, court action has ruled in favour of employees. Nevertheless, there is a need to have some precedent set, for example, demotions or deductions from salary. The proposed introduction of a code of ethics and conduct for Civil Servants under CSRP, would provide a framework for ensuring greater compliance with Financial Orders. Its implementation will however require a high level of policing as current schemes of service are inadequate, and therefore Civil Servants are more likely to misuse any powers delegated to them. The Civil Service Department is aware of the financial indiscipline issue, and proposes to tackle it through the introduction of incentives such as improved payscales for the Civil Service. Conclusions 6.42 In conclusion, numerous institutional strengthening measures have been initiated in the past several years. Many of these developments are just becoming operational and a considerable improvement in overall budget control can be anticipated in the course of 1997. This section has also recommended numerous additional actions and the acceleration of several programs. Some of these will entail additional technical assistance but are considered essential to strengthening budget implementation.  30. UGAßD . :..- -, Lake k Vidtoria ukob-a RWANDA KENYAJr- KENYA JÄ TAU 77ý-'~o P. 'AFRICA Arusha ) Ms BURUNDI r"1 -ARC6 hinyanga Lke rLake Eyaisi Manyara oma - STabora EI Sing d ZAIRZE DOD MAr \ Dar es Morogoro Mtera Lake Rejservoir GreatN Tanganyika R..h. R Lake naR Rukwa The boundänes, colors, Mbeya on th.s do Teold TANZANIAup .y ierýwora -MAIN ROADS e)REGION CAPITALS -+ RILOASNATIONAL CAPITAL -oge SRIVERS REGION BOUNDARIES Soge --INTERNATIONAL BOUNDARIES mo 100 200 KILMETER OZAMBIQUE 10 30 36 38