Report No. 47639-GH GHANA 2009 External Review of Public Expenditures and Financial Management (In two Volumes) Volume II: The Medium Term Expenditure Framework May 2009 PREM 4 Africa Region Document of the World Bank TABLE OF CONTENTS 1. INTRODUCTION..............................................................................................................................................1 2. MTEF: DEFINITION OF TERMS AND CONTEXT....................................................................................2 A. DEFINITIONOFMTEF ANDSTRATEGIC BUDGETINGTERMINOLOGY.................................................................2 B. CONTEXT .........................................................................................................................................................4 3. OVERVIEW OF MTEF PROCESS/CURRENT PRACTICE ......................................................................4 A. MTEF/BUDGETPREPARATION PROCESS .........................................................................................................5 Top down process (the setting of inter-sectoral/inter-MDA ceilings).................................................................5 Bottom-up process (the setting of intra-sectoral/intra-MDA budgetary allocations).........................................7 Consultative process and interaction between top down and bottom-up processes ...........................................9 Cabinet approval and Parliamentary scrutiny..................................................................................................10 B. DOCUMENTATION ANDSTAFF TRAINING........................................................................................................12 4. ASSESSMENT OF CURRENT PRACTICE ................................................................................................13 A. MEDIUM-TERM PERSPECTIVE TO BUDGETARYPLANNING...............................................................................13 B. REALISTIC ANDCREDIBLE MACRO-FISCALRESOURCE FRAMEWORK ..............................................................14 Resource projections.........................................................................................................................................14 Ceiling analysis.................................................................................................................................................14 C. COMPREHENSIVENESS -SETTINGOUTTHE OVERALL LIKELY LEVELOF ALLPUBLIC SECTOR RESOURCES OVER THE MEDIUM TERM ..................................................................................................................................................16 Reliability of MTEF Parameters (Revenues and Expenditures) over the medium-term...................................17 D. EXPLICIT LINKAGE OFSECTORALBUDGETARYALLOCATIONS TO GOVERNMENTPOLICIES (ALLOCATIVE EFFICIENCY).............................................................................................................................................................18 Assessment of the process of allocating inter-sectoral resources (top-down process).....................................19 Assessment of the process of allocating intra-sectoral resources (bottom-up process)....................................21 Inter-action between top-down and bottom-up processes.................................................................................22 E. MORE EFFICIENTPLANNINGANDUTILISATIONOFRESOURCES (EFFICIENTSERVICE DELIVERY).....................23 Efficiency of budgeting ­ fragmentation of budgetary resources .....................................................................23 Efficiency of MTEF/budget preparation process..............................................................................................24 Efficiency of budget implementation.................................................................................................................25 F. INSTITUTIONAL/CAPACITY ISSUES..................................................................................................................27 G. SUMMARY ­ IMPACTOFMTEF ON PFM PERFORMANCE ...............................................................................29 Aggregate fiscal discipline................................................................................................................................29 Strategic allocation of resources (Allocative efficiency) ..................................................................................29 Efficient service delivery (Technical efficiency) ...............................................................................................30 5. CONCLUSIONS AND RECOMMENDATIONS.........................................................................................31 A. CONCLUDINGREMARKS.................................................................................................................................31 B. RECOMMENDATIONS......................................................................................................................................32 A more results-orientated focus on budget preparation ...................................................................................43 Implications for GoG........................................................................................................................................43 Tables: Table II.1: Difference between indicative and final MTEF ceilings by spending item (GH¢ million).....15 Table II. 2: Comparison of MDAsIndicative Ceilings and Budget Submissions 2008 Budget (GH¢ ,,000) ....................................................................................................................................................................16 Table II.3: Expenditures included in the MTEF prioritisation process: 2008 Budget (GH¢ Million)........17 Table II.4: Accuracy of Revenue Projections over Time1 (GH¢ Million).................................................18 Table II.5: Budget Deviations by MDA, 2003-20071.................................................................................26 Table II.6: Summary of ERPFM Recommentations on Strengthening Strategic Budgeting/MTEF..........38 ii Diagrams: Diagram II.1: Current Budget Preparation Process.....................................................................................11 Diagram II.2: Proposed budget preparation process...................................................................................37 Boxes: Box II.1: Strategic Budgeting Terminology .................................................................................................3 Box II.2: Evolution of the MTEF, late-1990s to the present ........................................................................5 Box II.3 : Budget preparation timetable......................................................................................................11 Box II.4: Roles of Stakeholders in Budget Preparation..............................................................................12 Box II.5: Summary of comments from recent MTEF training ...................................................................28 Box II.6: Overview of Performance Budgeting..........................................................................................44 Annexes: Annex 1.......................................................................................................................................................45 Attachments: Attachment 1...............................................................................................................................................65 Attachment 2...............................................................................................................................................73 Annex Tables: Table A1.1: Tracking Expenditure Information in Individual MDA MTEF Volumes...............................46 Table A1.2: Difference between indicative and final MTEF ceilings by spending item (bn originally denominated Cedis) ....................................................................................................................................47 Table A1.3: Comparison of MDAsIndicative Ceilings and Budget Submissions ­ 2008 Budget (GH¢ Thousand) ...................................................................................................................................................49 Table A1.4: Accuracy of MTEF Forward Projections (bn originally denominated Cedis)........................50 Table A1.5: Accuracy of Revenue Projections over Time1 (bn originally denominated Cedis) ................52 Table A1.6: Expenditures included in the MTEF prioritisation process: 2008 Budget (GH¢ Million)......54 Table A1.7: Budget Deviations by MDA, 2003-20071...............................................................................59 Table A1.8: Deviations between MDAsBudgeted and Actual Expenditures by Spending Item, 2005- 20071 ..........................................................................................................................................................60 Table A1.9: Difference between MDAsquarterly cash plans and cash ceilings, 2007 (GH¢ ,,000)........61 Table A1.10: Difference between MDAsquarterly cash plans and cash ceilings, 2007 (GH¢ ,,000)......62 Table A1.11: Quarterly distribution of MDAscash plans, cash ceilings and revenue receipts, 2007......63 Table A1.12: Analysis of 2006 Budget Implementation by Spending Item1 (bn originally denominated Cedis)..........................................................................................................................................................63 Table A1.13: Analysis of 2007 Budget Implementation by Spending Item1 (bn originally denominated Cedis)..........................................................................................................................................................64 iii Annex Figures: Figure A1.1: Differences between indicative and final ceilings by MDA (largest differences), 2005......48 Figure A1.2: Differences between indicative and final ceilings by MDA (largest differences), 2006......48 Figure A1.3: Differences between indicative and final ceilings by MDA (largest differences), 2007......48 Figure A1.4: Accuracy of MTEF Revenue Projections..............................................................................52 Figure A1.5: Overview of Revenue Performance, 2004-2008 ...................................................................53 Figure A1.6: Trends in Actual Revenue Performance, 2004-2007.............................................................53 Figure A1.7: Total variance between MDAscash plans and cash ceilings ­ 2007...................................62 Table A1.10: Difference between MDAsquarterly cash plans and cash ceilings, 2007 (GH¢ ,,000)......62 iv CURRENCY EQUIVALENTS (Exchange Rate Effective as of May 7, 2009) Currency Unit= Ghana Cedis GH¢1 = US$0.70 US$1 = GH¢1.42 FISCALYEAR January 1 to December 31 WEIGHTS AND MEASURES Metric System ABBREVIATION AND ACRONYMS AFTFM Africa Financial Management AFTP4 Africa Poverty Reduction and Economic Management 4 AFTPC Africa Procurement AFTPR Africa Public Sector Reform and Capacity ARIC Audit Reports and Implementation Committees BFP Budget Framework Paper BPEMS Budget and Public Expenditure Management System CAGD Controller and Accountant General Department CF Consolidated Fund DACF District Assemblies Common Fund DANIDA Danish International Development Agency DEO District Education Officer DFID UK Department for International Development DHIS District Health Insurance Schemes ECG Electricity Company of Ghana ERPEFM External Review of Public Expenditures and Financial Management ERPFM External Review of Public Financial Management FAA Financial Administration Act GAS Ghana Audit Service GCAA Ghana Civil Aviation Authority GDP Gross Domestic Product GES Ghana Education Services GETF Ghana Education Trust Fund GHAPOHA Ghana Ports and Harbors Authority GNPC Ghana National Petroleum Company GoG Government of Ghana GPRA Ghana Petroleum Regulatory Authority GPRS II Ghana Growth and Poverty Reduction Strategy for 2006-2009 GTZ Gesellshaft für Technische Zusammenarbeit HIPC Heavily Indebted Poor Country IAA Internal Audit Agency IFMIS Integrated Financial Management and Information System IGF Internally Generated Funds IMF International Monetary Fund IPPD-2 Integrated Personnel and Payroll Database IRS Internal Revenue Service v KFW Kreditanstalt für Wiederaufbau LPG Liquefied Petroleum Gas MDAs Ministries, Departments and Agencies MDBS Multi-Donor Budgetary Support MDGs Millennium Development Goals MDRI Multilateral Debt Relief Initiative MMDAs Metropolitan, Municipal, and District Assemblies MoF Ministry of Finance MoFEP Ministry of Finance and Economic Planning MOH Minister of Health MTEF Medium-Term Expenditure Framework MUV Manufactors Unit Value NDPC National Development Planning Committee NEPAD New Partnership for Africas Development NHIC National Health Insurance Council NHIF National Health Insurance Fund NHIL National Health Insurance Levy NPA National Petroleum Authority OECD-DAC Organization for Economic Co-operation and Development ­ Development Assistance Committee OHCS Office of the Head of Civil Service PAC Public Accounts Committee PEFA Public Expenditure and Financial Accountability PETS Public Expenditure Tracking Surveys PFAU Project and Financial Analysis Unit PFM Public Financial Management PIF Permanent Insurance Fund PPME PPAs Public Procurement Model of Excellence tool PREM Poverty Reduction & Economic Management RAGB Revenue Agencies Governing Board RCCS Regional Coordination Councils SOE State Owned Enterprise SSNIT Social Security and National Insurance Trust ST&MT AP Short and Medium-Term Action Plan for PFM UNCITRAL United Nations Committee on International Trade Law VALCO Volta Aluminum Company VAT Value-Added Tax VRA Volta River Authority Vice President: Obiageli K. Ezekwesili Country Director: Ishac Diwan Sector Director: Sudhir Shetty Sector Manager: Antonella Bassani Task Team Leaders: Daniel Boakye Carlos Cavalcanti vi 1. INTRODUCTION 1.1 In the late nineties the Government of Ghana (GoG) adopted the Medium Term Expenditure Framework (MTEF), with the objective of moving away from incremental budgeting towards a more strategic approach to budgeting. After nearly a decade of implementing an MTEF, GoG has found it necessary to review progress made so far and to define the next set of reforms to improve the effectiveness of budget formulation and execution. The annual budget is intended to be formulated within the context of the Governmentsmedium- term national development strategy and the second Growth and Poverty Reduction Strategy (GPRSII), covering 2006-2009. 1.2 Within the context of the annual External Review of Public Financial Management (ERPFM), GoG requested a review of the MTEF process in order to improve the effectiveness of its public spending. The review aimed to identify how GoG can strengthen the link between policy objectives and budgetary outcomes, and thereby improve the use of budgetary resources to achieve better results and better value-for-money. The review was undertaken by a World Bank mission to Accra during July 14-25 2008, with a follow-up mission taking place during November 12-26, 2008. Volume II of this years ERPFM provides the detailed results of the MTEF review, which are summarized in chapter 4 of main report (Volume I). 1.3 As the MTEF should be an integral part of the overall budget preparation, it is not possible to isolate it from the broader budget formulation process. Thus, the discussion below sets the MTEF in this wider context and aims to address the question of how the process of strategic budgeting may be strengthened. 1.4 The rest of this report is structured as follows: Section 2 sets out the policy context for the MTEF in Ghana and provides a description of strategic budgeting concepts used in the paper. Sections 3 includes a description of the MTEF process in Ghana, whilst an assessment of these current practices and procedures is contained in Section 4. A set of short and longer-term policy recommendations, as well as a suggested Action Plan, is contained in Section 5 and Attachment 1, respectively. A typology of MTEF processes is presented in Attachment 2, whilst a series of annexes provides the detailed technical analyses supporting the conclusions in the report as well as more detailed guidance on individual proposals for the MTEF. 2. MTEF: DEFINITION OF TERMS AND CONTEXT A. DEFINITIONOF MTEFANDSTRATEGICBUDGETINGTERMINOLOGY 2.1 Broadly speaking, an MTEF (also referred to as a strategic budgeting process, as indicated above) may be defined as the process of linking public sector resource allocations to Government policies over the medium term. As such, it represents a mechanism to bring together Government policies, plans and budgets and thereby provide the strategic framework underpinning the annual budget. In this way, MTEFs can act as a key tool for operationalizing broader strategic policy initiatives, including PRSPs and MDGs. This is in contrast to the more commonly-found type of budgeting methodology, that of annual, incremental line-item budgeting, whereby central (e.g. Ministry of Finance) and sectoral (line) ministries (MDAs) budget for one year at a time by allocating an additional percentage, usually to cover inflation, to each line ministrys allocation and within this to different items of expenditure. 2.2 In short, strategic budgeting (or, the MTEF, as in this report) may be defined as the process of public expenditure planning which aims to allocate public sector resources (domestic and external) over the medium term to Government policy priorities. Thus, within a strategic budgeting process, annual budgets over time will ensure, respectively: (i) that future budgetary implications of policies are compatible with the likely macroeconomic and fiscal capacity of the government; (ii) that resources are reallocated from lower to higher policy priorities; and (iii) that there is greater certainty over resource envelopes for managers, thereby creating the conditions for better planning and implementation of activities. Strategic budgeting also entails a focus on the objectives, outputs and, where possible, the outcomes of government activities, thus going beyond the exclusive focus on inputs reflected in line-item budgeting. 2.3 Despite the existence of numerous terms related to MTEFs/strategic-based budgeting concepts (see Box 1), they all refer to certain elements of the broader process described above. In practice, MTEF is the generally-accepted (umbrella) term for an initiative to increase the strategic focus (i.e. the policy basis) of the budget. 2.4 In the case of Ghana, there does not appear to be a clear consensus on the meaning of what constitutes the MTEF. The MTEF may refer to the medium term macro-fiscal framework, including MDA budgetary allocations, contained in the Budget Statement (referred to as the MTFF in Box II.1 below), and/or to the detailed MDA volumes setting out a comprehensive annual budgetary allocations by objective and activity. - 2 - Box II1: Strategic Budgeting and Budget Costing Terminology Whilst a Medium-Term Expenditure Framework (MTEF) is a generally-accepted umbrella term for a strategic phase to the budget, other terminology is sometimes used to refer to particular aspects of an MTEF process or to distinguish different levels of sophistication of such an exercise. These include: Medium Term Fiscal Framework (MTFF): Generally refers to the macro-fiscal framework. It usually contains fiscal policy objectives and a set of integrated medium-term macroeconomic and fiscal targets and projections. Medium Term Budgetary Framework (MTBF): brings in broad allocations of aggregated medium-term resources to individual sectors or spending agencies (ceilings), based on overall budget policies and some costing of priorities. The objective of a MTBF is to allocate resources according to the governments strategic priorities and to ensure that these allocations are consistent with overall fiscal objectives. It includes medium term projections for cross-cutting budgetary parameters such as recurrent/investment ratios and level of the public sector wage bill. Medium-Term Expenditure Framework (MTEF): As indicated above, general umbrella term for a strategic phase to the budgeting process. When distinguished (i.e. identified separately) from MTFFs and MTBFs, it can refer to a more advanced form of an MTEF process, one which incorporates sector expenditure strategies (in the form of e.g. programme, activity or output-based budgeting). On the sector side, a variety of terms are used in budget costing and in making intra-sectoral budgeting allocations according to policies, depending on the methodology used. These include: Sector Expenditure Strategies/Sector MTEFs: linking sector policies, budget plans, and intra-sectoral budget allocations over the medium term. Programme-budget/programme-based budget: most frequently, but not necessarily, found in a medium-term budgetary framework, often for individual ministries, setting out objectives, broad or detailed programmes and activities, and linked to annual or multi-annual budgetary allocations that are based on the objectives of the programmes. To the extent that they focus on ministries rather than sectors as a whole, programme-budgets are narrower than sector expenditure strategies/sector MTEFs. Performance budget: strictly defined, a budget that explicitly links each increment in resources to an increment in outputs or other results. Broadly defined, any budget that presents information on what Government organisations have done or expect to do with the money provided to them.1 Output budget: a budget system that links appropriations to specific outputs. Strictly interpreted (also known as "accrual budget"), appropriations are measured on an accrual basis, rather than on a cash basis, and managers are engaged to deliver outputs through "contracts" negotiated with ministers. Types of budgeting techniques used for assessing the costs of activities to be included in the MTEF include: Zero-based budgeting: in theory, this is a budgeting process which consists of evaluating all programmes each year and preparing the budget from scratch, in practice, it has not worked as it creates an administrative burden that exceeds its potential benefits. Activity-based costing: an approach to costing which apportions costs to overheads through aiming to identify the cost driver for a particular category of overhead. It is basically a managerial accounting technique of determining unit costs, rather than a strategic approach. As in the case of zero-based budgeting, it creates an administrative burden that exceeds its potential benefits. Terms for non-strategic budgeting include: Incremental budgeting: a budget process which is based on adding an increment to the previous years budget allocation. Line-item budgeting: a budget process whose basis is the allocation of resources according to the economic (line item) classification. Source:Betley,2009. 1OECD Budget Practices and Procedures Database, Final Glossary, December 2006. - 3 - B. CONTEXT 2.5 GoG indicates the importance it places on strengthening the MTEF process. In its PFM documentation, GoG emphasizes that its budgets are formulated under the MTEF. The Governments Short and Medium-Term Action Plan (STAP) indicates that the priorities for the medium term (2006-2009) are to deepen and strengthen the capacity of staff for the MTEF process, strengthen the capacity to link the budget to the GPRSII process, and to ensure expenditures are in line with priority areas. These priorities are echoed in the latest Annual Progress Report (APR) from the National Development Planning Commission (NDPC). 2.6 At the same time, the Budget Guidelines (e.g. the 2009-2011 Guidelines) indicate that, as part of the deepening of the MTEF process, budget ceilings are being given for the 3-year period, 2009-2011,2 and that MDAs should stay within their indicative ceiling since they "may not deviate from those indicated" (a more detailed discussion on this point is given below). 3. OVERVIEWOF MTEFPROCESS/CURRENT PRACTICE 3.1 This section presents a brief overview of the MTEF process in Ghana, including the timetable and the roles of stakeholder institutions. In line with the discussion above, although the discussion is focused on the MTEF, it cannot be separated from the rest of the (annual) budget process, since the former should be the basis of the latter. 3.2 The process of strategic budgeting was introduced by GoG in the late 1990s, as part of the PUFMARP reforms, and, as such, there appears to be reasonably widespread awareness of the process amongst MDAs and Parliament. To these stakeholders, the MTEF is usually understood to be the 3-year rolling budget framework presented in the detailed MDA volumes (though, as the assessment shows, the MTEF and the contents of the MDA volumes do not necessarily represent the same thing). However, this review considers the MTEF also to include the macro-fiscal framework in the annual Budget Statement and Economic Policy (Appendices 6-8 in the 2008 Budget Statement), which would be considered to be the top-down (aggregate) part of the MTEF, as well as the summary tables at the beginning of each MDA MTEF volume. Box II.2 summarizes the evolution of the MTEF since its introduction. 2Previous ceilings were given in detail (by item) for year 1 and then in summary form (just showing the MDA total). - 4 - Box II.2: Evolution of the MTEF, late-1990s to the present From late 1990s-early 2000s Introduction of new classification system, which simplified the categorisation of economic expenditures Removing dual budgeting 2003-2005 Greater macro-fiscal stability achieved First GPRS, 2003-2005 Greater links between planning process and GPRS APR process (first APR in 2003) Achieving of HIPC completion point, leading to greater availability of resources, to be targeted to investment; included in medium-term macro-fiscal framework 2005-2006 2006 budget was first one passed before beginning of the coming budget year GPRSII 2006-2009 STAP, emphasising the strengthening of MTEF Greater PAC involvement in budget process Incorporation of MDRI resources in medium term macro-fiscal framework 2007-2008 Greater analysis of links between budgets and GPRS Changes in classification; streamlining of MTEF activities First introduction of policy hearings Greater focus on technical hearings: attempts to place greater emphasis on justification of budget submissions for allocation of additional resources Preparation of initial MTEF submissions by pilot MMDAs MTEF training Source: Betley,2009 based on discussions with MoF officials. A. MTEF/BUDGET PREPARATION PROCESS 3.3 The MTEF/budget preparation process essentially involves four separate processes: (i) the top-down process of preparing and finalising the overall macro-fiscal resource framework and the determination of inter-sectoral/inter-MDA ceilings, which primarily involves MoF (and should also involve Cabinet); (ii) the bottom-up process of MDAs preparation of their budget estimates (intra-sectoral/intra-MDA allocations); (iii) the interactive process between MoF and MDAs of finalising the draft budget estimates; and (iv) Cabinet and Parliamentary discussion and approval of the final budget. A description of each of these processes is set out below, followed by a summary of the budget preparation timetable (based on the most recent budget process), and a summary of the roles of each of the stakeholders in the process. 3.4 Prior to the beginning of the budget preparation process, MoF aims to consult with Ghanaians through placing advertisements in the media, requesting suggestions from civil society and the general public for budget priorities. Those individuals or groups who respond with suggestions are listed at the front of the Budget Statement. Top down process (the setting of inter-sectoral/inter-MDA ceilings) 3.5 Macro-fiscal framework. The first step in developing the medium term and annual budget framework is determining the resource envelope, covering both domestic and external resources, which is based on the medium-term macroeconomic framework. At present, GoG uses a basic - 5 - but appropriate macro-economic model,3 based on the financial programming methodology. In terms of making resource projections in preparation for the budget, MoF (Budget and Policy Analysis and Research Divisions) works with the Revenue Agencies and with MoFsTax Policy and Non-Tax Revenue Units to agree on likely domestic tax and non-tax revenues (IGFs) for the coming 3-year period, and with the External Resource Mobilization and Debt Management Divisions on the projections for external resources. 3.6 The setting of MDA ceilings. Based on the macro-fiscal framework and the level of total resources likely to be available, MoF decides on indicative ceilings for the coming MTEF (three- year period). These indicative ceilings are provided by MDA and by source of funds (e.g. GoG discretionary, DP, and IGFs). For GoG discretionary expenditure, separate ceilings are provided for each of the four spending items.4 The setting of indicative ceilings involves an iterative process which is item-specific (see below). Initial ceilings are set on the principle that no MDA should be worse off; there is an adjustment for once-off events and an attempt is made to incorporate the effects of new policies, as well as GPRS policies. However, the indicative ceilings do not include an indication from Cabinet on annual or medium-term priorities. 3.7 By design, indicative ceilings for GoG discretionary expenditure are set such that there is scope for Cabinet to authorize the allocation of additional resources following the technical budget hearings. In the 2007 budget, for example, the ceilings (GoG discretionary) in the Budget Guidelines represented only 77% of the final ceilings in the Budget Statement; in the 2005 and 2006 budgets, the share of additional resources in the final ceilings not covered by indicative ceilings was lower, at around 17% (i.e. the indicative ceilings covered around 83% of the final ceilings). As such, they are intended to be viewed as a minimum, and as a guide for MDAs. The bases for the calculation of initial ceilings are actual expenditures from the previous year and expected (revised) expenditures from the current year.5 Ceilings for payroll. Indicative ceilings are provided for item 1 for each MDA based on current civil service staffing and pay levels, adjusted by an annual increment. For 2009, the base figure assumed no new recruitment. One of the difficulties is that the alignment of the data with actual staff numbers for MDAs is not complete, including for new/changed ministries and for subvented agencies, making it difficult to have a clear basis for overall staff numbers in each MDA, making it difficult to have a solid basis for setting realistic ceilings for payroll.6 In addition, the indicative ceilings for personal emoluments (PE) include a contingency amount, specifically for item 1; the lack of finalisation of public sector wage negotiations before the coming budget year leads to the use of contingency line to cover pay increases agreed during the budget year;7 Ceilings for other GoG discretionary items. Ceilings for item 2, which include allowances, relate to PE. Ceilings for items 3 and 4 cover the balance of overall likely resources and are based on last years actual expenditures and current years projected 3For example, the model is not based on detailed econometric analyses or based on a Social Accounting Matrix (SAM), both of which would be difficult to justify, given difficulties with data availability. 4Specifically item 1 = personal emoluments; item 2 = administration; item 3 = services; item 4 = investment. 5However, with the start of preparation of the revised ceilings for the next budget year(s) occurring towards the end of the first quarter or the beginning of the second quarter, of the current budget year, there is a limit to the degree of additional information available on the current budget. 6For evidence of this, one may look at the, at times large, fluctuations in monthly figures for staff numbers 7The first step in public sector wage negotiations is the setting of the minimum wage by GoG, which tends to occur during the first quarter of the coming budget year (i.e. the year to which the wage negotiations refer). - 6 - budget performance, with some adjustments for information on new policies (e.g. ring- fenced activities). In practice, ceilings for items 3 and particularly 4 represent minimum figures, based primarily on the previous years actual expenditures. 3.8 Ceilings for other resources, including external resources and IGFs. The ceilings for DP resources (which are mainly used for investment) are based on current programmed disbursements over the medium term, adjusted by an adjustment factor for realistic levels of disbursement by DP. For the 2009-2011 MTEF, ceilings for external resources were given only for 2009 and 2010. Projections for ceilings of Internally Generated Funds (IGFs) are based on projections of IGFs (both lodged and retained) from the Non-Tax Revenues Unit in MoF. 3.9 Inclusion of inflation. There is no explicit guidance to MDAs on how to incorporate inflationary expectations in their estimates; for example, there is no mention of inflation in the 2009-2011 Guidelines. Ceilings for item 1 include an adjustment, roughly in line with real economic growth, whilst those for item 2 also include a factor related to growth. Inflationary expectations are not built into the ceilings for items 3 and 4. It is probable that some MDAs reflect projections of increases in prices in their estimates (the Activate software enables this, at least for items 3 and 4), but the bases for these increases is not clear, and they could be quite different from GoGs official estimates. 3.10 Budget Guidelines. Following preparation of the ceilings, MoF issues Budget Guidelines to MDAs, describing the tables to be filled in, setting out technical issues for MDAs in preparing their budgets and including indicative MDA ceilings. Bottom-up process (the setting of intra-sectoral/intra-MDA budgetary allocations) 3.11 Strategic planning. The bottom-up process of budget preparation concerns the preparation of MDA budget submissions, which begins with a strategic planning process. The overarching national strategic plan is contained in the GPRSII, which is intended to provide the basis for MDAsstrategic goals and hence its medium-term expenditure policies. Many MDAs (including health, education, roads/transport and agriculture) have produced 5-year plans (fixed time period, rather than rolling), which set out their sector goals and strategic framework, their medium-term sub-sectoral policies in line with GPRSII, and descriptions of the monitoring and evaluation framework (including policy targets and indicators). In some sectors (e.g. health), these plans are referred to as 5-year Programmes of Work (POW), with annual updates (annual POWs) produced each year. However, depending on the sector, there can be relatively limited information and analyses on the medium-term budgetary implications of the policies, beyond the costs of implementing the stated policies and associated funding gaps. With an effective MTEF, the choice and timing of policy implementation would be prioritised to fit within the overall resource framework; this prioritisation and realistic budget planning process would address and eliminate the funding gap. 3.12 In conjunction with preparing the POW, some sectors/MDAs (e.g. health) prepare 5-year capital investment plans (CIP) or Strategic (Sector) Investment Plans (SIPS). These are intended to look specifically at the allocation of resources (both domestic and DP) for capital projects, with annual plans showing budgetary resources to specific projects and an overview of the past years investments. These plans also identify and quantify financing gaps. However, the degree to which the associated recurrent costs linked to planned new investments are accommodated in - 7 - planned budgets is not clear, nor it is clear the extent to which the projects are covered in the detailed MTEF volumes, nor indeed how the management and use of non-traditional sources of funds are being planned. 3.13 NDPC plays an important role in detailed sector planning through providing support for: translating broad GPRS policies and strategies into more detailed sectoral plans, costing policy interventions, and monitoring and evaluating sectoral policies. However, it is up to the MDA (in theory, through the budget/MTEF process) to translate these detailed sector plans into realistic budgets. NDPCs APR review process provides a structure for reporting back on progress in implementing GPRS policies. As discussed below, non-financial performance information generated as part of the APR process would be useful to incorporate into the budget process. 3.14 Preparation of annual/medium term budget estimates. As part of their sectoral strategic planning process, MDAs (e.g. health and education) hold sector or policy reviews to review progress on policy implementation and review budget priorities for the coming years, focussing mainly on the upcoming budget year. These reviews are intended to be centred on reviewing progress in sector policies and in implementing the GPRSII. 3.15 Following the policy reviews, and (intended to be) guided by the indicative ceilings provided in the Budget Guidelines, MDAs prepare updated budget requests by detailed input (e.g. fuel and lubricants, training materials, other materials and consumables),8 objective and activity (objectives and activities are shown for service and investment only), focussing primarily on the upcoming annual budget. In principle, the estimates are updated through a process of aggregating the budget requests of districts; in practice, given time constraints (see below), these changes are more likely to take place at the MDA headquarters. These updated estimates are intended to reflect policy priorities, the performance and pace of progress in meeting policy objectives, and new policy initiatives for the MDA; however, in practice, given that there are relatively few changes in the activities and emphasis in MDA volumes year-on- year and that many of the activities are concentrated on non-developmental matters, such as workshops, it is difficult to see the active translation of policy initiatives into budget allocations. At the same time, whilst the budget preparation software facilitates the compilation of the budget estimates and thus the detailed MDA volumes, at the same time, it can lead to a mechanical approach to budgeting; one which does not lend itself well to what should be a political process. In practice, the detailed activities and the costing of these activities were undertaken during the initial preparation stages of the MTEF, and largely marginal changes have taken place since.9 3.16 Revision of budget estimates. Following the completion of their initial budget requests, which as discussed below can be significantly higher than the indicative ceilings given in the Budget Guidelines, MDAs submit their initial budget requests to MoF. After the budget (technical) hearings, where MDAs defend their submissions, MDAs go through a process of re- prioritising their estimates to reflect changes in their overall parameters. Whilst this process is intended to be undertaken in a bottom-up fashion, as with the original estimates, time pressures mean that in practice the prioritisation takes place largely by staff in the main MDA headquarters; it is not clear how much active communication of the revised requests and of the details of reprioritisation there is with the lower levels during this time. At the same time, as the 8Inputs for item 1 are broken down into established posts and (new) recruitment; examples of inputs for item 2 include postal charges, and printed material and stationery. 9A review of detailed MTEF volumes over time confirms this. - 8 - objective is to bring the estimates within the final ceiling and finalise the budget estimates during a relatively short space of time, this puts further pressure on the ability of time-pressured headquarters staff to plan resources, to be implemented often in the districts, in the most policy efficient way. 3.17 Institutionally, MDAs Budget Committees, consisting of senior management from planning, budget and the main MDA sub-sectoral departments, were originally set up to provide cross-MDA input into budgetary decisions. As comments from participants at the MTEF training show (see below), these are not operating effectively. The result is less transparency across MDAs in the budget decision-making process. 3.18 Performance information. Performance information, in terms of supporting (informing) the budgeting process, is concentrated at the planning stage. Specifically, MDAs, with support from NDPC, compile a series of targets and indicators which may be used to monitor progress on achieving GPRS policies; progress on these indicators is discussed in the APR process and at budget policy hearings, as well as in the Budget Statement. There is limited information on performance provided alongside the budget submissions by MDA; in the detailed MTEF volumes, the information shown under "outputs" (e.g. "population of small ruminants increased by 50% by 2010"10) provides some targets for MDAs. However, how the medium-term budgetary resources will help achieve the targets is not clear; in some cases, links between targets and resources are not clear, such that targets which are listed as due to be completed in the first year (e.g. "procure ICT equipment by June 2008"11) show continued funding allocations through the medium term. At the same time, the link between the targets and the strategic objectives are not explained. Consultative process and interaction between top down and bottom-up processes 3.19 Policy hearings. Following the circulation of the Budget Guidelines, MoF, NDPC, and MDAs attend policy hearings and cross-sectoral meetings, which aim to discuss the updated resource framework and policies and (mainly annual) budget requirements. In recent years, the interactive process for the coming budget has kicked off with what is known as Policy Hearings, attended by MDA Chief Directors. At these hearings (which for the 2009 budget were held in July 2008) MoF provides an overview of the macro-fiscal framework for the coming budget year. Following an overview of progress against their sector M&E indicators, MDAs present their medium-term strategic objectives and policies, together with their revised budget for the current year and the proposed budget for the next year (the draft budget year). However, there is little discussion of the cost-effectiveness of past spending policies or of how planned budgets will be used to meet the stated policy goals (in other words, a realistic discussion of prioritisation within the overall (total) level of resources expected, amongst the myriad of policies discussed). Instead, the focus of the discussions is largely on MDAs item-based spending needs (i.e. the allocation of GoG discretionary resources across spending items) for the coming budget year, which in effect is largely taken up by personal emoluments. 10Ministry of Food and Agriculture, 2008-2010 MTEF. 11IBID. - 9 - 3.20 An inter-governmental forum is held with MMDAs on the fiscal implications of budget policies. Following their receipt of MDAs initial budget submissions MoF holds technical budget hearings for MDAs to discuss MDAs budget submissions. 3.21 Technical budget hearings. These hearings are intended to give MDAs the opportunity to discuss and explain ("defend") their budget submissions to MoF and NDPC in an effort to justify receiving additional budgetary resources from Cabinet. As indicated below, MDAs tend to make initial budget submissions which are above (in some cases, considerably) initial ceilings in the hope or anticipation of additional resources allocated following the technical hearings and Cabinet discussions. Participants in the hearings include MDA management officials, those from the Budget Division in MoF, and NDPC senior management. The discussions are said to focus relatively more on the need for greater resources by spending item than on the specific policies and outcomes to be achieved by the resources. At the same time, there is relatively little detailed, rigorous scrutiny of the analytical basis for the allocations (e.g. the justification for the implied staff numbers behind the request for the personal emoluments budget). Following the hearings and Cabinet discussions, MoF prepares a report on the hearings, which focuses mainly on revised MDA budget allocations. Cabinet approval and Parliamentary scrutiny 3.22 There are four main steps in the approval process, which occur following the submission of the draft estimates to Cabinet (i.e. towards the end of the budget formulation process): (i) Cabinet discusses, makes changes, and approves final aggregate MDA budget estimates presented by MoF;12 (ii) the Minister of Finance reads the Budget Speech (Budget Statement) to Parliament; (iii) the Parliamentary Finance Committee reviews the MDAsbudget estimates, and has access to the detailed MDA volumes.13 It produces a report to the House on the estimates, which mainly focuses on the policies and estimates in the Budget Statement; and (iv) Parliament approves the Appropriations Bill. 3.23 For the last three years (since the 2006 Budget), Parliament has approved the budget before the beginning of the new budget year, in December.14 Following the approval of the budget, MoF has in recent years prepared a summary of the budget for the wider public, in the form of a Citizens Guide to the Budget Statement. 3.24 The steps in the budget preparation are set out below in Box II.3 and diagrammatically in Diagram 1. 12The Cabinet reviews the initial MDAsbudgetary allocations, which are intended to be based on initial ceilings, but are not in practice, and approves final MDA ceilings, which are circulated to MDAs; as will be discussed below, these are considerably different to the indicative ceilings included in the Budget Guidelines. The basis of these final ceilings is discussed below. Following the final ceilings, MDAs prepare their adjusted budgets, the Budget Statement is finalised, and the Appropriations Bill is prepared. 13The Parliamentary Finance Committee, together with select committee members from relevant sectors, reviews MDAs budgetary submissions. The focus of the review is the information in the Budget Statement and the MDA broad budgetary allocations in the draft appropriations bill. According to the reports on these reviews, the discussions focus on the allocations across spending items rather than how the allocations are used to achieve policy objectives. 14Approval of the 2009 budget was split due to the elections in December 2008, with approval for the first quarter of 2009 approved in November ("Expenditure in Advance of Appropriation"), and preparation and approval of the budget for the rest of the year to be approved by the new Government and Parliament, during the first quarter of 2009. - 10 - Diagram 1: Current Budget Preparation Process15 Cabinet-apprd Budget to Parliament (Nov) Cabinet/ Parliament Appropriation Bill approved by Parliamt (Dec) Prepn of Budget Updating of Budget Statement MoFEP macro-fiscal Guidelines, indic. ceilings issued to framework MDAs Technical Budget Policy Hearings, Hearings - Final MoFEP, budget NDPC, MDAs ceilings, estimates MDAs MDA policy Preparation of MDAs reviews draft budget submissions Jan-March April-June July-September October-December Source: Betley,2009 based on discussions with MoF officials. Box II.3: Budget preparation timetable Budget Step Timing1 Request for input into the budget by civil society and general public January Updating of macro-fiscal framework (MoF) January-March Preparation and circulation of Budget Guidelines (MoF) May MDAs policy review (MDAs) April-June Policy hearings ­ MDAs, NDPC, and MoF July MDAs prepare budget estimates July-September MDAs submit budget requests to MoF September Budget hearings ("technical hearings") on MDA requests (MoF) September-October MDAs revise their budget estimates, as needed September-October Submission of draft budget (detailed estimates) to Cabinet October Cabinet recommends changes to MDAs budget allocations October MDAs finalise their detailed budget estimates October Cabinet approves draft budget October MDAs prepare their detailed MTEF volumes October Preparation of Budget Statement (including MDA budget policies and estimates) August-October Submission of Budget Statement (including MDA budget policies and November estimates) to Parliament Parliamentary Finance Committee discusses and approves MDAs budget November estimates Parliament debates and approves Appropriation Bill December Budget information is disseminated to the public, including the preparation of a January-February Citizens Budget Note: 1. Based on 2008 and 2009 Budget processes Source: 2008-2010 and 2009-2011 Budget Guidelines, MoFEP. 15The diagram by necessity does not show all of the steps, nor all of the stakeholders involved. - 11 - 3.25 The role of stakeholders in the preparation of the budget is summarised in Box II.4. Box II.4: Roles of Stakeholders in Budget Preparation Stakeholder Specific role in MTEF/budget process MoF: Policy Analysis and Research Division, Update macro-fiscal and MTEF resource framework Budget Division Prepare Budget Guidelines to guide the budget formulation process Liaise with MDAs on their budget preparation Scrutinise MDAs budget requests MDAs Review MDA policies with regard to GPRSII Liaise with districts on budget priorities and detailed estimates Co-ordinate/Prepare detailed budget estimates Defend budget estimates Re-prioritise estimates as required NDPC Liaise with MoF and MDAs on GRPS implementation progress Sector/thematic working groups Review progress on implementing GPRS and consider priorities for upcoming budget period Civil society/general public Provide input to the budget process through MoF Development partners Provide updated estimates of external finance to MoF Participate in sector reviews and working groups where appropriate Cabinet Review and approve budget estimates Parliamentary Finance Committee Review Budget Statement (budget policies) and estimates. Prepare report on findings of review Parliament Approve Appropriation Bill Source: Betley,2009 based on discussions with MoF officials. B. DOCUMENTATIONANDSTAFFTRAINING 3.26 There appears to be limited documentation of current MTEF procedures available to new or existing staff. Whilst an MTEF manual was produced in 1999, it has not been updated, with new staff expected to learn on the job.16 The manual is reportedly being updated currently. 3.27 MoF is aware of the need for greater capacity building on the MTEF for staff. With the support of GTZ, it has embarked on a programme of stakeholder training on the MTEF for staff in MDAs, which is expected also to encompass training specific to individual sectors (e.g. health). 16Mission team members were not given access to the manual, as it was deemed too old. - 12 - 4. ASSESSMENT OF CURRENT PRACTICE 4.1 Having described how the MTEF/budget is prepared, this section provides an assessment of the current practices of preparing and implementing the MTEF, against each of the criteria of a strategic budget. These criteria include the following basic characteristics:17 (i) a medium- term perspective to budgetary planning; (ii) a realistic and credible macro-fiscal resource framework; (iii) setting out the overall likely level of all public sector resources over the medium term; (iv) the explicit linkage of sectoral budgetary allocations to Government policies; and (v) more efficient planning and utilisation of resources. Institutional and capacity issues are reviewed at the end of this section which concludes with a summary of the impact of the MTEF on PFM performance. 4.2 The assessment of current practice in the following paragraphs is based on each of these criteria, and is supported by a set of data analyses, contained in the Technical Annex (Annex 1). A. MEDIUM-TERM PERSPECTIVETO BUDGETARY PLANNING 4.3 As government policies and programme initiatives may take several years to implement, it is important that the MTEF provides a budgetary framework which enables MDAs to plan their spending over the medium term (usually, 3 years). GoGs MTEF performs well on this measure, with its aggregate macro-fiscal framework having been presented within a comprehensive multi-year time horizon since the early 2000s. This involves three-year projections (relevant budget year plus two forward years) of domestic revenues (including IGFs) and external grants, domestic and external financing (including budget support and programme and project credits and loans), and statutory and discretionary expenditures (including Statutory Funds and the use of HIPC and MDRI resources). 4.4 The macro-fiscal framework is reasonably comprehensive of all budgetary resources, and GoG continues to improve the level of comprehensiveness over time. In the 2008 Budget and 2008-2010 MTEF, an analysis of tax expenditures (in the form of import exemptions) was included for the first time. Projected expenditures for the Statutory Funds are included. As indicated in the PEFA analysis,18 the coverage of external finance in the budget is reasonably comprehensive (coverage of at least 90% for external assistance, at least for the coming budget year). 4.5 Whilst the MTEFs aggregate macro-fiscal framework has a multi-year perspective, the MTEF at a more detailed level (i.e. detailed expenditures at MDA level) does not. In fact, at the level of the MDA, budget information is focussed primarily on the coming budget year (see Annex 1). The presentation of the detailed MDA objective and activity-based allocations does not satisfy the requirement of a multi-year time horizon for the planning of public expenditures since government policies and programmes may take several years to implement but these estimates of the costs of multi-year programmes are not shown. 17Based on international practice. Specific references are available on request. 18See PI-7(ii), PEFA Performance Report for Ghana, 2006 - 13 - B. REALISTICANDCREDIBLE MACRO-FISCAL RESOURCE FRAMEWORK 4.6 An effective MTEF provides MDAs with predictable and stable projections of expected budget resources, from all sources (both domestic and external), to enable them to plan their expenditures realistically over the medium term and thus encourage more effective and efficient utilisation of resources in implementing policies. This sub-section assesses the relative credibility of GoGs macro-fiscal MTEF framework in recent years, in terms of: the accuracy of revenue projections; and the credibility of spending ceilings provided to MDAs, focussing specifically on ceilings for the annual budget. Sub-section (ii-b) below addresses the question of the reliability of medium-term revenue projections and forward expenditure allocations. Resource projections 4.7 Analyses of revenue projections (both domestic revenues and external grants) suggest that resource projections overall are reasonably accurate for the coming budget year (see analyses in Annex 1). In particular, they are sufficiently robust, particularly in aggregate, not to warrant changes in ceilings during budget preparation, (i.e. there is insufficient new information provided during the year [i.e. between May and August] to undermine the overall ceilings provided in the Budget Guidelines). In terms of overall receipts, over the past three years, total revenues and grants have been within 3% of projections in the relevant Budget Statement/MTEF. Indeed, over the last 5 years, there has been only one year in which outturns differed by more than 6% of the Governments projections.19 There has been greater variability amongst individual categories of revenues, particularly for VAT, which affects projections of resources for Statutory Funds. Nonetheless, lower-than-projected receipts of VAT on imports, and excise and import duties recently, due to global economic factors, have been compensated for by higher-than-projected receipts of direct tax receipts and VAT on domestic goods. Annex 1 contains more detailed analyses of revenue projections. Ceiling analysis 4.8 One of the requirements of MTEF is the credibility of ceilings provided to MDAs. Setting annual and medium term ceilings is intended to enforce spending limits and to encourage MDAs to prioritise their own expenditures within a hard budget constraint. If MDAs believe that there is the possibility of negotiation on their final budget amounts, then they will not focus during detailed budget preparation on prioritising their planned expenditures within a realistic resource framework. 4.9 In practice, an analysis of the data on initial and final ceilings over the past 3 years (see Annex 1) indicates that the ceilings in the Budget Guidelines are unreliable as a guide to MDAs in their preparations of budget submissions; there are significant differences between indicative and final MTEF ceilings, both by spending item and across MDAs (Table II.1). It is notable that there are wide variations between indicative and final ceilings across MDAs, suggesting that some MDAs are more or less successful at obtaining (in some cases, significantly) greater resources during budget negotiations. Further details are available in the technical annex (see Annex 1). 19This is reflected in the relatively high scores for the PEFA PI-3 indicator for the period 2003-2007 - 14 - 4.10 Given the realistic expectation of additional resources during the budget process, MDAs have the incentive to submit bids which are higher than those given by the indicative ceilings, and indeed this could be seen to be intentional in giving minimum ceilings by MoF. Overall, based on data for the 2008 budget process, budget submissions by MDAs were around one- quarter higher than their ceilings, with some very significant differences (Table II.2). Thus, in preparing their budget submissions, MDAs do not adhere to a hard budget constraint on their ceilings, as would be expected in an MTEF budget process. 4.11 Whilst MoFsintention is to provide an opportunity for MDAs to justify their requests for additional funds, in practice, it serves potentially to undermine MDAsefficient budget planning and prioritisation process and potentially leads to results that have more to do with relative bargaining strength amongst MDAs than realistic expenditure plans for achieving policy objectives (focussing on likely results) and the cost-effectiveness of past spending. This risk would be lessened if specific and detailed expertise to challenge MDA budget proposals (on the relative cost effectiveness, the likely results in terms of policy achievements, and the value-for- money of budget requests) were in place in MoF. Nonetheless, the setting of indicative ceilings which are ignored by MDAs during budget preparation has the effect of potentially undermining two measures of PFM performance, the strategic allocation of resources in line with government policies, and efficient service delivery. Table II.1: Difference between indicative and final MTEF ceilings by spending item (originally denominated Cedis Billion) 2005 2006 2007 Indicative Final Diff as % Indic Final Diff as % Indic Final Diff as % ceilings ceilings of Indica- ceilings ceilings of ceilings ceilings of (appropr.) tive Indicative Indicative Personal emoluments 6,992 7,246 3.6% 8,147 9,241 13.4% 9,915 12,534 26.4% Administ- ration 1,381 1,514 9.6% 1,693 2,089 23.4% 1,968 2,562 30.2% Service 995 598 -39.9% 665 712 7.1% 774 831 7.4% Investment 1,257 1,347 7.2% 1,631 1,592 -2.4% 2,651 2,273 -14.3% Total 10,624 10,705 0.8% 12,137 13,634 12.3% 15,308 18,201 18.9% Source: MoF, CAGD. 4.12 The second rationale for setting indicative ceilings conservatively is due to caution regarding likely revenue for the coming budget. However, as indicated above, adjustments to revenue projections in practice are unlikely to be significant between the second and third quarters of the year. Thus, given the reasonably accurate aggregate resource projections, it would appear that there is relatively little value-added to an additional one quarters worth of data on actual receipts (i.e. in practice, there is a difference of 3-4 months between the setting of indicative and final ceilings). Indeed, if revenue projections are so unreliable even in the very short term, it begs the question of whether or not there is sufficient stability in the macro-fiscal framework beyond the very short term to warrant an MTEF (whose focus by definition is on the - 15 - medium term and whose successful implementation relies on a stable macro-economic environment).20 Table II.2: Comparison of MDAs' Indicative Ceilings and Budget Submissions 2008 Budget (GH¢ `000) 2008 2008 Diff as % of Largest MDAs Indicative ceiling MDA submission indicative ceiling Min. of Interior 18,946 74,824 294.9% Audit Service 4,020 10,910 171.4% Min. of Finance and Economic Planning 12,493 26,597 112.9% Min. of Manpower Youth & Employment 4,988 9,967 99.8% Judicial Service 9,893 16,675 68.6% Min. of Education, Science and Sports 618,136 1,017,553 64.6% Office of Gov't Machinery 9,116 14,085 54.5% Min. of Land, Forestry and Mines 8,928 13,344 49.4% Min. of Transportation 7,245 9,947 37.3% Office of Parliament 7,109 9,618 35.3% Min. of Food & Agriculture 20,801 28,100 35.1% Min. of National Security 21,393 27,212 27.2% Min. of Water Resources, Works and Housing 4,936 5,835 18.2% Min. of Health 252,788 281,000 11.2% Min. of Local Gov't, Rural Dev't and Env't 33,576 35,449 5.6% Revenue Agencies 72,846 73,901 1.4% Remaining MDAs 348,069 149,520 -57.0% Total for all MDAs 1,455,283 1,804,537 24.0% Source: MoF, CAGD. C. COMPREHENSIVENESS -SETTINGOUT THE OVERALL LIKELYLEVEL OF ALL PUBLIC SECTOR RESOURCES OVER THE MEDIUM TERM 4.13 MTEF should be comprehensive of all public resources and all public spending in order to link expenditures to government policies effectively and to present the overall trade-offs between expenditure alternatives. An analysis of the MTEF indicates that the prioritisation process is not comprehensive of all public spending representing less than 50% of total public expenditures (see Annex 1). In addition, as will be noted below, not only does the MTEF cover a relatively limited proportion of total expenditures, but also the spending items that it does cover are those that are the least robust in terms of implementation performance. 4.14 Specifically, GoGs MTEF currently excludes a significant share of public expenditure from the decisions on strategic resource trade-offs within overall resource constraints, and this potentially negatively affects the linkage between policies and resources (the allocative efficiency measure of public spending). As indicated above, whilst the aggregate macro-fiscal framework (the top-down part of the strategic budget process) is reasonably comprehensive and includes all sources of funds (Consolidated Fund, IGFs, statutory funds, HIPC, MDRI and resources from development partners), the detailed allocations to activities (in the MDA MTEF volumes) are not. In fact, the data in the detailed volumes only include details of expenditures on services and investment for sources of funds covering GoG discretionary, donor funds, and 20This is confirmed by studies of international experience; specific references are available on request. - 16 - IGFs (the first 3 columns of Appendix 6 of the Budget Statement). The process for allocating and prioritising personal emoluments and administration expenditure (items 1 and 2, respectively),21 expenditures of the Statutory Funds, MDAs internally-generated funds, HIPC funds, and some DP project resources occurs outside of the MTEF planning process. Similarly, separate processes govern the allocation of domestic and externally-financed investment. These differences result in a significant fragmentation of the budget process. For the 2008 budget, the detailed volumes cover less than half (around 42%) of total non-interest appropriated public expenditures (see Table II.3). Table II.3: Expenditures included in the MTEF prioritisation process: 2008 Budget (GH¢ Million) Expenditures Share of total expenditures Total 5,060 100% of which prioritised by activity-basis in MTEF: GoG discretionary resources allocated to services 109 2.1% GoG discretionary resources allocated to investment 712 14.1% Allocation of IGFs (retained)2 299 5.9% Allocation of resources by development partners3 1,000 19.8% Sub-Total of above 2,120 41.9% Note: 1. Total expenditures (Appendix 6 of 2008 Budget Statement) include spending from all sources of funds, including domestic revenues, IGFs, HIPC and MDRI, and Statutory Funds. 2. Allocated to services and investment, in accordance with the FAA.22 3. Mainly allocated to investment. Source: MoF. Reliability of MTEF Parameters (Revenues and Expenditures) over the medium-term 4.15 A key objective of MTEF is to provide predictability of funding to MDAs over the medium term. In other words, MDAs should be able to plan their expenditures two to three years ahead based on the indicative fiscal parameters, including the expenditure allocations, given in the MTEF. With advanced-level medium-term budgets (as opposed to MTEFs), projected allocations for outer MTEF years are taken as the basis for preparing the next years budget, with any deviations from these forward parameters for the coming budget explained in the Budget Guidelines.23 As will be discussed below, this requires a stable macro-economy and high-level commitment to enforcing budget/MTEF ceilings; in practice, relatively few countries prepare medium-term budgets. 4.16 Medium-term indicative resource projections. As indicated above, revenue projections are less reliable over the medium-term than for the annual budget, with some categories (particularly, VAT) falling short of projections in recent years. An analysis of revenue projections contained in recent MTEFs compared to actual revenue outturns indicates that MDAs may be wise to be cautious in planning expenditures beyond the immediate budget year (Table II.4). MoF will likely continue its efforts to improve the accuracy of its medium term revenues projections over time; however, success will likely take time, particularly given the uncertain macro-economic environment in the near term. However, as indicated above, there is greater 21Items 1 and 2 are shown by cost centre in the detailed volumes but not by objective and activity. 22However, there is anecdotal evidence of some IGF resources being used to top-up salaries, particularly for subvented agencies. This practice is being addressed. 23This is the basis for the PEFA PI-12 indicator. - 17 - relative certainty of the macro-economic environment fuelling projections for the annual budget (with the possible exception of the coming year to 18 months). Table II.4: Accuracy of Revenue Projections over Time1 (originally denominated Cedis Billion) 2006 2006 2007 2007 2007 Projections Actual Projections Projections Actual Source of projections 2006 MTEF 2006 MTEF 2007 MTEF Direct 7,269 7,183 8,269 8,877 9,404 Personal 2,255 3,111 2,570 3,823 4,015 Self-employed 428 356 488 454 519 Companies 3,789 3,013 4,318 3,822 3,998 Other direct 797 702 894 778 872 Indirect 12,060 10,662 13,228 14,125 13,073 VAT 7,032 5,881 7,803 8,876 8,342 Petroleum 4,148 4,071 4,422 4,226 4,033 Excise 880 710 1,003 1,023 698 Inat'l trade 4,859 5,418 5,870 6,136 5,766 Total Tax Revenue 24,187 23,263 27,367 29,138 28,243 Other revenue (inc. NHIF) 1,514 1,383 1,725 1,824 2,577 NTR 711 923 810 3,175 3,383 Other 0 0 0 0 0 Grants 5,099 6,349 7,400 8,945 8,572 Total Revenue & Grants 31,511 31,918 37,303 43,082 42,775 Note: 1. Excludes tax exemptions Source: MoF, CAGD. 4.17 Medium-term indicative expenditure allocations. Giving a more uncertain revenue environment, it is not surprising that there are significant differences between MTEF outer year projections for MDA expenditures and subsequent ceilings, both overall and for individual MDAs (see Annex 1 for details of analysis). D. EXPLICIT LINKAGE OFSECTORAL BUDGETARYALLOCATIONSTO GOVERNMENT POLICIES (ALLOCATIVE EFFICIENCY)24 4.18 Both the MTEF and the annual budget are intended to ensure that sufficient and appropriately targeted resources are provided to implement Governments documented policies, such as the GPRSII, and the MDGs. This requires: (i) for inter-sectoral allocations, sometimes referred to as the "top-down" process: That Cabinet approves a transparent allocation of resources to sectors and MDAs in line with government priorities, based on analyses of the overall resource and sector policy frameworks; 24This refers to the second of the three PFM performance measures. - 18 - That trade-offs amongst competing policy choices (different sectoral and MDA allocations over time) be clear to Cabinet when reviewing the decisions on allocations between sectors/MDAs; That these allocations are approved early in the budgeting process, in order to be communicated to MDAs before they begin preparing their detailed budget submissions (this provides high-level commitment to the MTEF parameters and helps strengthen accountability); That the basis for the approved sectoral/MDA allocations be clear to sectors/MDAs; That the reasons for any changes to indicative ceilings (across years) be transparent and based on changes in policy priorities or on revised projections of the resource framework, which are then reflected in transparent changes to allocations based on policy priorities. 4.19 It also requires: (ii) for intra-sectoral/MDA allocations, sometimes referred to as the "bottom-up" process: That expenditure allocations (trade-offs) within the sector (intra-sectoral spending policy choices) be clear; That intra-sectoral budgetary allocations be based on clear and achievable policies, realistic estimates of the resources necessary to implement these policies, and analyses of the cost-effectiveness (value-for-money) of past expenditures and policies (in effect, that the detailed planning process be operationalised and reflected in both budget submissions and final allocations); That intra-sectoral allocations be comprehensive of all spending in the sector (i.e. that spending by all MDAs in the sector is co-ordinated and in line with overall sectoral policy priorities and that all spending in the sector is included in the sectoral analyses, including e.g. Statutory Funds); That there be a feedback loop between the sectoral analysis (bottom-up process) and the process for allocating subsequent inter-sectoral allocations (top-down process); in other words, that the sectoral analysis is communicated effectively to, and taken into account, by the MoF and Cabinet in their discussions of, and decisions on, budget allocations across sectors and MDAs (i.e. providing an analytical basis for setting subsequent MDA ceilings). Assessment of the process of allocating inter-sectoral resources (top-down process) 4.20 The current GoG MTEF preparation process (determining both inter-sectoral and intra- sectoral/MDA allocations) does not facilitate a clear link between the priority policy areas and sectoral budgetary allocations; it is not clear from examining the budget/MTEF how resources are being used to achieve the governments priority policies. The format of the current MTEF (with detailed expenditures set out in copious detail in the MTEF volumes) does not present strategic information in such a way that Cabinet can see clearly, and decide amongst, relevant policy trade-offs (based on expenditures to sectors or other broad policy areas); it is difficult for them to do so given the fact that the analytical framework to facilitate their decision-making on expenditure ceilings is relatively limited (and thus could be expanded), and that the detailed - 19 - MTEF documents are very hard to follow. This has implications for high-level commitment to the annual and forward expenditure ceilings. 4.21 In other words, the broad strategic assumptions behind the medium-term sectoral allocations are not set out in a way which makes the strategic spending trade-offs clear. There does not appear to be a comprehensive and systematic analysis and overview of the strategic policy assumptions and key spending choices across sectors and, within sectors, MDAs as the basis for setting the ceilings. The Budget Guidelines, which contain the indicative ceilings sent out at the beginning of the budget preparation period, do not provide detail on the overall strategic and policy context for the MDA ceilings, nor is there a separate comprehensive and systematic analysis of the strategic policy assumptions and key spending choices across sectors and, within sectors, MDAs as the basis for setting the ceilings. That is, there is not a separate MTEF policy document (i.e. separate from the annual budget documentation in the form of the Annual Budget Statement), presented to policymakers at the beginning of the budget preparation process which sets the policy framework for both the MTEF and annual budgets, which is standard for many MTEFs. 4.22 Hence, the basis for the approved sectoral/MDA allocations is not clear to MDAs; the analytical justification for expenditure allocations tends not to be communicated as a matter of course to stakeholders, including underlying assumptions for the macro-fiscal (aggregate) resource framework and the allocations of indicative ceilings both across MDAs and across years, and the final approved ceilings.25 In particular, the central policy trade-offs amongst priority sector areas are not analyzed or discussed. Analyses show significant variations across MDAs in differences between their indicative and final ceilings, indicating relatively greater success for some MDAs in obtaining additional resources during budget negotiations. Minutes of Cabinet discussions on final MDA budget allocations are not public. However, an analysis of the movements over time in GoG budgetary allocations across sectors indicate a mixed picture in terms of links with stated policies (see technical analysis in Annex 1).26 In terms of allocations by item, final ceilings for one or other of service and/or investment were below the initial amounts in 2005-2007 in favour of higher ceilings for personal emoluments and/or administration. 4.23 Whilst the Budget Statement sets out projected allocations by MDA and source of funds, it does so according to the administrative classification (MDAs) grouped under the three GPRS key thematic areas. However, since MDAs may include a number of disparate policy areas which may not naturally share a common strategy (e.g. education, science and sport), the presentation of allocations by MDA would not necessarily represent clear trade-offs across shared policy areas. At the same time, the three GPRS thematic areas provide an aggregation of different policy areas which, given their different objectives, could be more operational by disaggregating them. An example is good governance, which covers different issues such as defence, justice, local government, as well as public administration. Thus, it would seem useful to show a broad overview of allocations in between the very broad detail of the thematic areas and the more detail MDA level. In many countries implementing strategic budgeting processes, 25This was a comment frequently made by MDAs, including during the 2008 Health Summit and during the MTEF training. 26An analysis of expenditure shares is necessarily a rough proxy for relative policy priorities since policy measures may include legislative or other changes which require fewer resources. - 20 - allocations are shown along lines similar to COFOG27 functions, which may or may not cut across MDAs. 4.24 In the analyses behind the strategic projections, it is important to show both past and future movements in relative allocations over time so that policymakers may see the allocations in context. The aggregate (top-down) information in the Budget Statement includes MDA allocations for the coming three years but does not show recent past trends. Whilst this would make the presentation less compact, it would help policymakers place the medium-term projections in context. This usually involves the most recent (or two) years of actual expenditures, the current budget year (revised estimates), the coming budget year, and projections for the medium term (two years after the coming budget year). 4.25 Finally, high-level decisions on, and commitment to, sectoral and MDA ceilings early in the budget process would help to provide high-level backing to expenditure ceilings and credibility to the medium-term budget process. However, at present, meaningful (final) ceilings are not approved by Cabinet before the Budget Guidelines (with the ceilings) are circulated. Assessment of the process of allocating intra-sectoral resources (bottom-up process) 4.26 In practice, the links between the NDPC-supported strategic/overall planning process and the allocation of resources is weak. Without the setting out of budgetary resources within a strategic policy framework, it is not clear that intra-sectoral budgetary allocations are based on achievable policies, realistic estimates of the resources necessary to implement these policies, and analyses of the cost-effectiveness (value-for-money) of past expenditures and policies. A too-detailed budget format focuses on micro-spending decisions rather than on what resources are intended to accomplish. The focus in the MDA volumes in practice is on small items deemed as "strategic" but which in fact do not appear to be so (e.g. the prevalence of conferences and seminars). Whilst the inability to report expenditures on the basis of MTEF activities makes it difficult to undertake a definitive analysis of spending on such activities, a concerted examination of the 2007 MTEF by MDA indicates a substantial proportion of MDA cost centre activity focused on these activities. It is difficult to see how many of these assist the sector to move forward in terms of reforms and the achievement of results. 4.27 As with the inter-sectoral allocations, overall sectoral expenditures are not planned or presented within a sectoral framework; thus, it is not easy to see how resources are being used to meet (sectoral) policy objectives. Using the information in the current MTEF, it is very difficult to analyse how all expenditures (including all resources and all spending items), working together (i.e. across all spending items), will be used to meet key policy and service delivery objectives in a given sector. Whilst it is the case that the budget allocations for the coming budget year are disaggregated by function (as defined in terms of GFS2001) in a summary table in the detailed MDA volumes, this information is not used elsewhere to highlight movement of resources across sectors over time. Some sectors, particularly those benefiting from significant development partner support, undertake such analyses as part of their annual joint sector reviews, which is intended to form the basis of their MTEF/annual budget preparation process. These kinds of analyses would be useful to undertake and use more widely in discussions with stakeholders (including Cabinet and Parliament) as part of the budget process. 27UNs Classification of the Functions of Government. - 21 - 4.28 The very detailed presentation of MDA budgets means that expenditure allocations (trade-offs) within the sector (intra-sectoral spending policy choices) are not clear. It is difficult to follow the trail of expenditures (see Annex 1), particularly since changes to MDA figures at the last minute may lead to internal inconsistency between the summary MDA tables and the detailed allocations in the rest of the volume. The lack of textual descriptions of the strategic assumptions makes it difficult to place the budgetary framework in a strategic context. MDAs and others report that the MDA documents are too detailed to be useful as a strategic planning document (the level of detail hinders making choices explicit across policy areas) and, without a summary, it is difficult to see the expenditure trade-offs being made. The detailed MDA volumes are in practice more annual operational plans than strategic (i.e. medium term and explicitly based on policy priorities) expenditure strategy documents. Inter-action between top-down and bottom-up processes 4.29 There is limited linkage in practice between the top-down and bottom-up budget processes, in terms of the regular interaction between the detailed MDA/sectoral work undertaken by MDAs on the one side, and the setting of subsequent MDA ceilings on the other side. The lack of a true forward perspective at the individual sector/MDA level and weak links between sector plans and annual budgets (detailed information on how resources will be used to meet policy objectives and their relative cost-effectiveness) make it difficult for Cabinet to have a clear basis for the setting of sectoral ceilings. 4.30 Scrutiny of MDAs budget submissions. There appears to be limited detailed/analytical scrutiny of MDAs budget submissions, in terms of how resources are being used to achieve priority policy objectives, and the effectiveness of past spending, and even less (effectively none) on their medium term plans by MoF, Cabinet and Parliament. At present, the focus of budget preparation (including interaction between MDA and MoF) and consideration by both Cabinet and Parliament is on the coming budget year. It can be argued that unpredictability in the policy or the macroeconomic context over the medium term makes this a sensible strategy; however, this attention paid to allocations across line (spending) items in budget discussions and debate, including between MDAs and MoF, and with Cabinet and Parliament discourages a strategic approach to budgets. 4.31 The technical hearings provide an important opportunity to strengthen this inter-active process. However, as indicated above, the discussions currently focus relatively more on the need for greater resources by spending item, rather than on sector expenditure performance and the cost-effectiveness of past and planned expenditures (and of the results of spending). There is, as yet, relatively little ex-post analysis of the outcomes of budget hearings, although there appears to be a desire within MoF to do more. - 22 - E. MORE EFFICIENT PLANNINGANDUTILIZATION OFRESOURCES (EFFICIENT SERVICE DELIVERY)28 4.32 The achievement of value-for-money for budgetary resources depends on how efficiently these resources are planned and utilised (implemented); in other words, the budget planning and execution processes are important. Important issues include how comprehensive and co- ordinated budgets are, the appropriate mix of inputs (spending items), and how closely budgets are implemented as planned through the provision of timely and predictable releases to those delivering services and implementing projects, and the processes governing procurement. Efficiency of budgeting ­ fragmentation of budgetary resources 4.33 Fragmentation in budget planning potentially undermines both the efficiency (value-for money) and the policy basis (strategic allocation) of spending. Increasing concentration of GoG discretionary expenditures on PE-related spending means that, effectively, spending from the Consolidated Fund is being used primarily to pay salaries and related expenditures29, whilst MDAs are having to rely on other sources of funding, including IGFs, external resources, Statutory Funds, resources from debt relief and non-traditional funds, e.g. private sector partnerships, for funding services and investment. The separation of the decision-making processes for allocating resource use by different sources of budget funds (different funding pots) may reduce value-for-money for public resources since the spending decisions made by each of the sources of funds are unlikely to be the same as those which would have been had all resources been considered together (an example is the earmarking of DP funding, which has the effect of reducing MDAsflexibility in using their resources most efficiently). At the same time, the separation of the decision-making processes is unlikely to lead to the best (most efficient) combination of inputs to implement activities. At the same time, differences in the funding flows of different sources of funds may lead to further fragmentation and potential inefficiencies in the implementation of strategic programmes (budgeting for results). 4.34 As the prioritisation of expenditures during the MTEF process focuses on "discretionary" expenditures, an argument is sometimes made that personal emoluments and related expenditures are not discretionary in the short term. However, they are in the medium term, which is the MTEFs focus. MDAs should be analysing how their combinations of inputs (personnel, supplies and investment) may provide services at less overall cost and provide better value for money. 4.35 Finally, it is worth underlining the point that earmarking external finance (particularly in terms of earmarked sector budget support)30 hinders MDAs ability to prioritise their resources according to their policy priorities. 4.36 Thus, this fragmentation of the budget formulation process potentially undermines efficient inter-sectoral and intra-sectoral prioritisation of resources since it is not possible to analyse how all expenditures (including all resources and all spending items), working together, 28This refers to the third PFM performance measure, also known as technical efficiency. 29Items 1 and 2 constituted 80% of GoG discretionary spending in 2007. They represent nearly 90% of the 2009 January- March Expenditure in Advance of Appropriation. 30It is to be noted here that earmarking sector budget support is not consistent with the OECD-DAC definition of sector budget support. - 23 - will be used to meet key policy and service delivery objectives. Thus, the MTEF has in practice been of limited assistance in helping Cabinet make the trade-offs between sectoral allocations and between major categories of spending, such as the payroll, non-wage operating expenses, and investment. The way the budget is planned makes it difficult to ensure: (i) that resources are linked to GoGs policy objectives; (ii) policies and programmes are financially sustainable, and (iii) that the resources are spent as intended to meet these objectives. Efficiency of MTEF/budget preparation process 4.37 Inefficiencies in budget preparation (including the fragmentation of budgets) can lead to inefficiencies in the provision of public services; conversely, improvements in putting together budgets can lead to better uses of resources. The way that MDA budgets are currently prepared is very wasteful of staff time and does not promote efficient allocation of resources in line with policies. As already mentioned, detailed planning at the deconcentrated level is undermined by significant subsequent changes in the MDAs ceilings after these initial budgets have been prepared. Time constraints for reflecting these changes in MDA budget submissions mean that detailed bottom-up estimates (from sub-district level through district and regions to headquarters) prepared by the disaggregated levels may be revised without consultation or notification either by headquarters or by MoF. This may serve potentially to reduce ownership of the final budgetary decisions. This process (and the weaknesses described) is repeated when resources do not flow to MDAs as expected. Payroll planning 4.38 Given the importance of the wage bill in annual budgets, weaknesses in payroll planning potentially undermine overall efficient budgetary planning. Analyses indicate that responsibility for the relatively poor results of budget credibility by MDA (see analysis below) is due to systematic differences between the approved payroll budget (planned) and actual payroll spending during the year, with reallocations from planned contingency resources. Realistic planning of personal emoluments expenditure appears to be weak, linked to poor underlying data on the public sector and a lack of a de facto hard budget constraint on PE. Payroll planning during budget preparation appears to be affected by the fact that civil service pay negotiations are completed after the budget has been approved, poor information on actual and projected staff numbers, and weak justification for MDAs budget submissions; together with weaknesses in payroll controls by MoF,31 these contribute to retrospective approvals of increases in staff levels, and in-year unanticipated increases in payroll.32 Given analytical capacity constraints in both MDAs and MoF and weak data on the existing nominal roll by MDA, MDAspayroll estimates may not be scrutinised as effectively as is desired. At the same time, there is an expectation amongst MDAs that payroll is a statutory expense (at least in the short term) and must be met as presented. Thus, the incentives on both MDAs and MoF would tend to lead to upward pressure on payroll in the absence of a hard budget constraint. 31Partial analyses of original payroll budget submissions by some MDAs suggest that actual expenditures for these MDAs were close to their original payroll estimates (which were then subsequently reduced during the budget finalisation process); the closeness between original request and actual spending could indicate a lack of a hard budget constraint in practice for payroll. 32The completion of civil service pay negotiations during the year (i.e. after the budget has been approved) leads to unanticipated increases in payroll, and the use of the contingency budget item, which should not be its role. - 24 - Capital budgeting 4.39 It is a relative strength of the MTEF process in Ghana that recurrent/investment expenditures are shown side-by-side. However, without a medium-term perspective the interaction between the two, in terms of ensuring that recurrent expenditures associated with new investments are reflected in forward projections, is not as strong as it could be. At the same time, the fragmentation of the investment budget between domestically and externally financed resources (the former, primarily for small-scale items, and the latter for more development- oriented projects), reflected in differences in the way investments are planned and managed depending on the nature of the investment and the funding source, potentially undermines the strategic basis for investment as a whole. Whilst externally-financed investment tends to be multi-year in nature and reasonably well-costed, investment expenditures by most MDAs (with the exception of works/housing, and roads/transport) tend to be for small-scale capital items (e.g. vehicles, office equipment, etc) and undertaken largely on an annual basis. During the budget year, large variations between investment allocations and outturns potentially further inhibit MDAs from being strategic in their investment plans. A separate annex provides more detail on how the MTEF could address the challenges of plans to increase public investment significantly, including the use of non-traditional sources of funding (e.g. the private sector). Efficiency of budget implementation 4.40 The efficiency of budget implementation has an impact on the efficiency of the provision of public services. Four issues affecting efficient budget implementation are discussed below: (i) budget credibility; (ii) in-year predictability of resources for MDAs to implement their budgets; and (iii) budget monitoring during the year. Detailed analyses are available in the technical annex to this report (see Annex 1). 4.41 Budget credibility remains a challenge. The credibility of the overall budget remains weak, particularly in terms of deviations across MDAs and by spending item (see Budget Deviation Index in Table II.5 and a detailed analysis in Annex 1). Based on these data, the updated scores for the PEFA indicators PI-1 and PI-2 (focussing on the period 2005-2007) would be B and D, respectively. Weak credibility appears to be the result of weaknesses in payroll planning, as well as poor in-year predictability of resource flows to MDAs, relative to MDAs spending plans, particularly for the items of services and investment. - 25 - Table II.5: Budget Deviations by MDA, 2003-20071 Budget Year Total expenditure deviation2 Total expenditure variance3 Variance in excess of total deviation4 2003 2.3% 15.8% 13.6% 2004 12.4% 25.9% 13.4% 2005 8.0% 33.3% 25.3% 2006 6.0% 36.3% 30.3% 20075 3.2% 21.0% 17.7% 1/Notes: 1. Data cover both discretionary and statutory expenditures, exclude contingency, and measure: (i) individual deviations between originally budgeted expenditures and actual expenditure outturns for the 20 largest budget heads; and (ii) total deviations for the remaining budget heads as a group. 2. Figures refer to the absolute value of the difference between actual outturns and original budgeted amount, expressed as % of original budgeted amount. For budgeted amounts, the original appropriated amounts were used, excluding supplementary budgets. 3. Figures refer to the sum of the absolute value of deviations as a proportion of the total (original) budgeted allocation for primary expenditure. 4. Percentage difference between expenditure deviations and expenditure variations (previous two columns). 5. CAGD preliminary data, pending confirmation by the Auditor-General. It is to be noted that data on total expenditures differ from those compiled by MoFEP. Sources: MoFEP, CAGD. 4.42 Weaknesses in in-year predictability for MDAs of resource availability. Successful strategic budget implementation requires that MDAs have predictability in their receipt of funds through the year, in other words, that releases from Treasury are in line with MDAsexpenditure plans and take place in a timely fashion. An analysis of the implementation chain (from appropriations to MDAs cash plans, cash ceilings, releases and expenditures) indicates that MDAs continue to face challenges in implementing their budgets in a predictable manner, and this can contribute to increased budget deviations. There are significant differences between MDAsown cash plans and the timing and amounts of Treasury releases, both by MDA and by spending item, particularly for services and investment.33 Specifically, in terms of the gap between cash plans and cash ceilings, there were significantly greater differences on services and investment (items 3 and 4) than on personal emoluments and administration. This implies that there is greater discretion applied to these items when allocating resources to MDAs from Treasury during the year. Thus, given delays in the release of funds, and strong central budget control, budgetary spending in practice is effectively set by small decisions on inputs. 4.43 Limited ability to report on budget as planned. For successful monitoring of any strategic budget, it is important that budget managers can review budget performance (both financial and non-financial performance) over time; this requires the ability to report on the budget as planned. However, the inability to report during budget implementation on the same basis as the very detailed MTEF outputs and activities hinders the linkage to overall objectives; in other words, there is currently not a mechanism to link activities to resources during budget implementation. Reporting is done down to sub-item level (equivalent to a cost centre), such as Feeder Roads Department. The existing classification system and the budget management system (BMS) used by CAGD does not include activity codes (shown as objective and output in the MDA volumes),34 so reporting on expenditures may not be done on the basis of planned activities, such 33Further details on the analysis may be found in Annex 1. 34Although it appears that such codes are being prepared. - 26 - as those already recorded by the NETS system.35 Finally, there is limited focus on non-financial budget performance (the cost-effectiveness of how resources are used to meet policy objectives). F. INSTITUTIONAL/CAPACITYISSUES 4.44 Ultimately, the weaknesses indicated above (such that the budget process has significant elements of line-item, annual, incremental budgeting) are a reflection of capacity constraints at both MoF and MDA level. This is hardly surprisingly given that international experience suggests change from line-item incremental budgeting to a strategic budgeting process, which links allocations to policies, performance and ultimately results, requires a substantial change in thinking about how the budget is structured, supported by a concerted change management process. Whilst there have been improvements in the process of MDAs review of their sector policies, nonetheless analytical capacity constraints mean that many medium-term sector strategies are not fully costed and realistically incorporated into the forward budget estimates, particularly with multi-year investment expenditures. At the same time, in-year reductions in allocations in one year have an impact on the reliability of the estimates for forward years. 4.45 Particularly for MDAs for which services are planned and delivered at the district level, the very detailed, bottom-up approach to budgeting requires significantly stronger analytical capacities at the lowest levels than are likely to be available in the medium term. There are often considerable inefficiencies in the ways that resources are being used, such that existing resource allocations provide little guidance as to the true costs of providing services efficiently. These problems are exacerbated by the fact that financial management often provides limited detail of the unit costs of providing a particular service or running an institution; whilst the information may be available, locating it requires considerable analysis. 4.46 Strengthening the link between objectives, policies and resource allocations at a more aggregated level is an important objective for improving the MTEF and ensuring better linkage with GPRS policy priorities. It is important that the allocations of resources are realistic, comprehensive of all resources, and reflect overall financial and physical capacity constraints. Improving MDAs ability to analyse the cost-effectiveness of alternative options for meeting policy objectives, determine what activities will be required, and how much these activities are likely to cost is critical in this regard. 4.47 In order to make the MTEF at sector/MDA level more meaningful, it would be useful to develop progressively the format, content and analysis of MTEF strategies to provide clearer guidance for the preparation of budget estimates. Broad descriptions of the strategies underlying the allocations would provide a better basis for analysing the appropriateness of these allocations and hence the justification for MDA budget requests. At the same time, it would be useful to try to simplify the process by analysing broader-based programmes rather than detailed objectives and activities. GoGsand MoFsemphasis on improving the relevance of budget hearings may provide an opportunity to strengthen this message. Improving the analytical basis for budgeting involves a longer-term process of building up detailed information on the costs and benefits of Government services and activities. Broad programmes (as distinct from objectives) may be useful to build this information in such a way that it is clearly linked to government objectives, 35The system is based on the National Expenditure Tracking System (NETS) to facilitate the generation of routine monthly expenditure reports based on accounting data. - 27 - thus allowing setting priorities among those on an informed basis, but that is not so detailed that the overall strategic direction is lost. 4.48 Enhancing analytical capacities in MDAs, as well as in MoF, will assist in this process. Specific training on budget policy analysis for Budget Division staff in MoF can facilitate the evaluation of MTEF and expenditure proposals from MDAs. Similarly, training of MoF and MDA staff in budget planning and evaluation techniques (both direct training and training-of- trainers) could be undertaken. It would be useful for MoF and NDPC to develop joint programmes for strengthening capabilities in MDAs to analyze public expenditure implications and priorities arising from sector policies and strategies, and to integrate such analyses into MTEF proposals and budget requests. 4.49 Indeed, stakeholders acknowledge that there are weaknesses in such capacities at these levels. Whilst the recent MTEF training has been viewed as useful and interesting by participants, nonetheless some areas have been seen as not relevant because they refer to systems and practices which are not part of the MTEF process in Ghana. Box II.5 summarizes some of the comments from the MTEF training.36 Box II.5: Summary of comments from recent MTEF training The following represents a selection of comments by participants at the recently completed MTEF training, sponsored by GTZ, on what areas of the MTEF could be strengthened. Ceiling setting not transparent; basis for determining ceilings is not communicated to MDAs Policy review doesntfeed into budget ceiling Lack of consultation with MDAs in setting ceilings Frequent changes in ceilings, particularly for item 1 Basis for changes to ceilings not communicated to MDAs Adherence to ceilings for second and third years is limited Ceilings are limited (funds less than those required to carry out GPRS) => MDAs cannot carry out their policies Lack of prioritisation by MDAs => shopping lists Displacement of items 3 and 4 to meet obligations for 1 and 2 Timing of announcement of final ceilings not known No feedback from Cabinet reviews on budget Lack of capacity to link policies and strategies Selected recommendations: Credible ceilings, including donor funds Cabinet to issue directives to guide sector working groups Improve costing of activities Improve budget hearings Revive sector working groups Introduce inter-sectoral meetings Strengthen institutional arrangements (Budget Committees) Need consultative process for setting ceilings Timely release of funds Sanctions for submissions of budget estimates which are late and/or not within the ceilings Capacity building Source: Summary of filled-in questionnaires as part of background papers for: MacKenzie, et al, Consultants Final Report: Deepening the MTEF Process in Ghana, July 2008. 36See MacKenzie, et al, "ConsultantsFinal Report: Deepening the MTEF Process in Ghana", July 2008. - 28 - 4.50 Beyond the systems and procedures, the underlying institutional incentives and broader work practices are important factors in determining the behaviour of budget stakeholders. In this regard, it may be argued that a number of underlying systemic issues have undermined the incentives for strategic budgeting and that stakeholdersbehaviour is a rational response to these underlying incentives. A summary of those incentives which potentially inhibit the process of strategic budgeting includes (discussed in more detail above): (i) the lack of high-level ownership of the indicative budget ceilings communicated to MDAs (at the beginning of the budget process), which may result in changes in MoF-set budgetary parameters towards the end of the budget preparation process; (ii) weaknesses (non-comprehensiveness) in payroll/wage bill planning by MDAs and in detailed MoF scrutiny of payroll budget requests during budget negotiations; (iii) weaknesses in predictability of in-year releases of resources; (iv) the lack of high-level or central level focus on the coming budget year; and (v) attention on the performance of the budget by spending item37, all of which serve to render detailed work on medium-term budgeting by objectives extremely wasteful of staff time. In this situation, the most rational choice for stakeholders is to plan their budgets incrementally and by line item. G. SUMMARY ­ IMPACT OFMTEFON PFMPERFORMANCE 4.51 As a way of concluding the assessment, this section looks at the implications of the MTEF process on the standard three measures of PFM performance, namely: (i) aggregate fiscal discipline; (ii) strategic allocation of resources (allocative efficiency); and (iii) efficient service delivery (value-for-money of spending, also known as technical efficiency). Aggregate fiscal discipline 4.52 The fact that budget preparation takes place within a transparent medium-term macro- fiscal framework helps to anchor the aggregate budget parameters. However, this process could be strengthened by: (i) the issuance of Cabinet-approved budget ceilings before the beginning of the MTEF/budget process; and (ii) MDAs adherence to these ceilings in their budget submissions. At the same time, improvements in overall expenditure controls, through improvements in Treasury/MoFsrole in commitment control, have helped to ensure that budget authorisations (through releases) provide a ceiling on overall expenditures. Whilst this is true at the aggregate level, resources flows for spending on priority objectives and activities at the MDA level are less credible, with significant differences from budget plans. Strategic allocation of resources (Allocative efficiency) 4.53 It is effective interaction between the top-down (inter-sectoral allocations) and bottom-up (intra-sectoral allocations) processes, based on meeting key policy priorities, which drives the strategic allocation of resources (allocative efficiency). In other words, the process of rolling over the MTEF each year should help provide a renewed strategic focus and enable new information, including changes in fiscal parameters, new priorities or progress in undertaking reforms, to be taken into account in subsequent spending decisions. This should be accompanied 37In-year reports on budget performance focus on spending items, as this is what appears to be requested by Government and Parliament. - 29 - by transparency in the setting of expenditure ceilings, which should have high-level endorsement, in order to facilitate widespread scrutiny of expenditure decisions. 4.54 However, whilst the strategic policy and sectoral objectives set out in the medium-term GPRS have provided a policy basis for guiding inter- (and intra-) sectoral allocations, including external finance, inefficient practices in Ghanas current MTEF and budget preparation and implementation processes potentially hamper the efficient linking of resources to these key policy priorities. These include the fragmentation of the MTEF and earmarking of resources (both domestic and external), the lack of an overall discussion of strategic medium-term expenditure choices as an integral part of the MTEF process (i.e. to guide the annual budget process), the lack of effective feedback and communication from MDAs to MoF and vice-versa in the rolling over of the MTEF, and difficulties with providing credible and predictable resources to MDAs during budget implementation. In addition, changes in budget parameters (e.g. ceilings) during budget preparation require a cumbersome re-prioritisation process, which in practice, due to pressure of time, is undertaken at headquarters without recourse to, or even communication with, lower levels. Efficient service delivery (Technical efficiency) 4.55 The emphasis of the budget preparation process, including the budget reviews, budget discussions and negotiations, and scrutiny by Parliament, on expenditures by spending items precludes discussions on spending efficiency and an examination of the effectiveness of public spending (i.e. the value for money) over time (that is, for what purpose are resources being used and how cost-effective are particular spending initiatives on achieving key policy outcomes [in other words, what are the results of spending]).38 The focus on vertical aggregation of budgets (which promotes a silo approach to budgeting) doesntencourage the identification of efficiency savings, including horizontally across activities implemented in districts or sub-districts. Frequent changes in ceilings during budget preparation may undermine detailed (and time- consuming) work on prioritisation. 4.56 At the same time, weaknesses in budget implementation may also adversely affect the efficiency of delivering services. During budget implementation, poor predictability of resource flows may lead to ad hoc prioritisation of resources. Higher-than-expected payroll requirements may require adjustments in other spending items, particularly service and investment, which are likely to have a significant impact on the efficiency of resources used to deliver basic services. 4.57 Gaps in analytical capacities to undertake the analysis required to assess the cost implications of alternative ways of achieving policy objectives, as well as relatively limited capacities for monitoring and evaluating spending, may also affect the search for greater cost effectiveness. In this regard, the data collected by the Public Expenditure Tracking Surveys (PETS) are essential in examining inefficiencies in the delivery of basic services. 38For example, the impact and cost-effectiveness of specific measures to increase agricultural mechanisation on agricultural output (noted in the Budget Statements each year since at least 2005 . - 30 - 5. CONCLUSIONS AND RECOMMENDATIONS A. CONCLUDINGREMARKS 5.1 The main objective behind a strategic focus for the budget is to direct actual public expenditures (thus encompassing both expenditures and revenues from all public sources of funds) to Governments policy priorities in order to improve the provision of efficient and effective public services. Building a strong strategic budgeting process normally requires: a stable macroeconomic framework; strong political leadership; commitment and active engagement (at the level of Cabinet); a credible and predictable budget (budget discipline); sufficient analytical and other institutional capacities and supporting PFM reforms, including downstream reforms to budget implementation. 5.2 At present, the MTEF appears to be neither strategic nor medium term. In practice, budget formulation takes the form of annual incremenftal line item budgeting. The MDAs MTEF volumes are focussed on presenting detailed items with little or no link to the overall strategic policy context. The emphasis in budget discussions tends to be focussed on implementing the budget according to line items, rather than on how resources are used. Whilst these expenditure items are presented under individual (specific and detailed) objectives and activities, the fact that they exclude a significant part of public resources necessary for assessing policy trade-offs to provide efficient service delivery serves in effect to render them meaningless for overall strategic direction. 5.3 The effect of these weaknesses is that prioritisation takes place by default (by item, rather than by policy objective or activity). This undercuts the strategic basis of the MTEF and encourages a non-strategic, incremental approach to budgeting. In effect, at present, there are 3 separate budget processes with separate processes of prioritisation, namely: (i) medium term fiscal framework: the aggregate (top-down) macro-fiscal framework; (ii) MDAs budget submissions, effectively incremental and by line-item; and (iii) in-year budget changes. 5.4 There appears to be little active political engagement at the beginning of the process in setting out broad strategic (medium-term) policy direction. At the same time, during budget preparation, the budgetary and Parliamentary discussions appear to be almost exclusively focussed on the budget estimates for the coming budget year (the estimates to be appropriated), rather than within the wider medium-term context. At the same time, reviews of budgetary performance tend to be based on budgetary performance by line item, rather than on the implications for wider service performance. 5.5 Given institutional constraints of the wider PFM cycle, this situation is understandable and stakeholder behaviour may be considered rational. Whilst the macroeconomic situation has until recently been relatively stable, weaknesses in ensuring credibility and predictability of implemented budgets undermine the strategic focus of the budget and serve to narrow the focus to meeting line item (particularly salary) requirements. 5.6 As a result, the budget process has become routine and the detailed bottom-up budget preparation work by the districts (i.e. the deconcentrated departments) becomes effectively meaningless as any detailed work on prioritisation is undermined by late changes in budgetary parameters for MDAs. As a result, an examination of the detailed MDA MTEF volumes (both - 31 - over years and across MDAs) shows that there are few changes over time in the parameters underlying the estimates. This is particularly true in terms of routine service and (non- developmental) investment. 5.7 Thus, whilst the ultimate objective of strategic budgets is the efficient deployment of resources to deliver services efficiently and in line with government policies, there appears to be a missing middle between the discussions of MDA policies and strategies, which are linked to GPRSII, and the discussions about expenditures and budgetary allocations. Indeed, there appears to be two detailed processes ­ the first is policies/strategies are set out in detail (GPRS), whilst the second concentrates on detailed budgets (according to spending items). Discussions of the link between the two do not appear to focus on a concerted attempt to analyse and explain how resources have been used in the recent past (the effectiveness and impact of previous spending decisions) and thus how resources will be used as a result in future to address policy challenges set out in the GRPS and in MDAs strategies. B. RECOMMENDATIONS 5.8 A strategic budgeting process has been implemented in some form in Ghana for nearly a decade, and GoG has recognised the need for refreshment and strengthening of the process in order to improve the results focus of the budget process. With a new government and the preparation of GoGs long-term development strategy coming up, it may be considered timely to review measures to strengthen budget formulation and provide an effective medium-term framework for kicking off the new governmentspolicy agenda. However, it must be noted that international experience indicates that fiscal stability is important for refreshing or re- invigorating an MTEF process. In the short term, difficult fiscal conditions, including measures to contain inflation, are likely to make it more challenging to do so. 5.9 GoG can take steps to help make budgeting more strategic through the MTEF but it needs to address the underlying credibility of the budget and weaknesses in annual budget planning first. 5.10 The recommendations set out below and the measures in the Action Plan are based on the objective of improving the results-orientation of the budget (the link between spending and results) and thereby the efficiency of spending (both allocative and technical). Specifically, the aim of the measures is to improve value-for-money (efficiency) of public spending and ensure that resources are used to achieve policy objectives. In other words, the measures are intended to ensure that the budgets focus is on results through improving accountability for, and transparency of, budget decisions. 5.11 The approach recommended is incremental and based on the existing strengths and challenges of the current budget/PFM system implemented by GoG. Improving the effectiveness of the strategic budgeting process will take time and needs to be accompanied by improvements in wider PFM, particularly in terms of budget implementation. Step-by-step changes will be required. 5.12 In brief, in order to re-invigorate the budget/MTEF process, three key steps are recommended:f - 32 - 1. Strengthen and signal political commitment early on to budget parameters through enhancing Cabinets role in the MTEF/budget process and engaging it early on in the setting of sectoral/MDA ceilings. 2. Ensure effective planning and management of payroll through better information on short term payroll requirements, completing public sector budget negotiations prior to the beginning of the year and incorporating the results in the annual budget, and enforcing a hard budget constraint on payroll. 3. Strengthen the links in practice between sectoral policies and the use of all budgetary resources (thereby improving budget comprehensiveness) by introducing two additional strategic policy documents into the process, namely, (i) a Budget Framework Paper prepared at the beginning of the process which sets out the strategic framework and trade-offs behind the proposed sector/MDA expenditure ceilings; and (ii) sector/supra-MDA medium-term expenditure strategies covering all spending, which would feed into the setting of sector and MDA ceilings. 5.13 Improving budget credibility (actual expenditures compared to appropriated budgets) would also strengthen the MTEF process. 5.14 Specifically, in the immediate/short term, it is recommended that efforts to refresh the budget process concentrate on (details for each of these measures are set out in the paragraphs below: Introducing a strategic step before the annual budget preparation process begins, such as the introduction of a Budget Framework Paper and high-level approval of ceilings (in the short term, this could be a report to which the Budget Guidelines with the ceilings could be annexed); Implementing institutional changes to ensure high-level commitment to the MTEF; Strengthening links at the sectoral level between policies and budgetary allocations, including providing incentives for reviewing and improving the cost-effectiveness of resource use and including information on performance; Restructuring the budget format to simplify the presentation, provide a better link with policy areas, and make the inter-sectoral expenditure trade-offs more transparent; Ensuring the comprehensiveness of the MTEF/Budget process; Improving payroll/wage bill planning and enforcing existing payroll controls; Improving the planning and management of public investment; Training on MTEF and prioritising expenditures ­ linking budgets to policies; Strengthening M&E: incentives for reporting on linking budgets to policies; Improving predictability and communication in the budget implementation process. 5.15 In the medium and longer term, the emphasis of reform measures can turn to: Implementing institutional changes to ensure high-level commitment to the medium term budget process, including tasking a Cabinet sub-committee with the role of overseeing the budget process; Gradually introducing more performance information into the budget process; Providing greater discretion for budget managers during the budget formulation stage, so that resources can moved from personal emoluments to other expenditure items as - 33 - long as the overall resource envelope for the MDA is not breached. At the budget execution stage this reallocation of resources would not be permitted; Building analytical capacities; Deepening M&E. 5.16 Each of the immediate/short term reform is discussed in turn below. 5.17 Introduce a strategic step in the budget process. A quick win in improving the strategic basis for the MTEF would be for MoF to prepare a strategic policy document, at the beginning of the budget process, which sets out the framework and context for the upcoming MTEF, including a medium-term fiscal (revenue and expenditure) policy context for the budget/MTEF and sector expenditure strategies. This would enable the Cabinet39 to assess strategically the proposed medium-term ceilings. As this type of analysis would be useful at the beginning of the budget preparation process, it is suggested to produce an expanded analytical paper alongside the Budget Guidelines which would set out an economic/fiscal context (including expenditure) related to the proposed ceilings. The current information provided in the Budget Guidelines could be attached, and it would include the Cabinet-approved ceilings. Once approved by Cabinet, this document could be circulated widely, including to MDAs and potentially Parliament. 5.18 As part of this analysis, it would be very useful to provide ceilings according to a category above MDA (supra-MDA), such as those covering areas for each minister (roughly representing sectors). It would also be useful to include analyses and expenditure projections by cross-cutting (cross-sectoral) policies. This would involve moving towards viewing trade-offs as across and within broad programme areas (under which service delivery and new policy objectives may be made) and away from trade-offs in terms of economic items. It would involve a shift in thinking by MDAs and the ability to prioritise and address expenditure trade-offs between/amongst policy areas containing all sources of funds together. The overall MTEF strategy document (see above) should include the analytical context behind the (indicative) ceilings for all sources of funds. Eventually (much later), it would be useful to consider providing greater flexibility in the allocation of resources, with the focus on providing block allocations to MDAs in return for being accountable for achieving agreed results, thereby moving away from a focus on line items. In this way, greater discretion for accounting officers would be matched by their greater accountability for results. 5.19 Institutional changes to ensure high-level commitment to the MTEF. As indicated above, in order to signal high-level commitment to the MTEF and budgetary ceilings, it is recommended to ask Cabinet to approve the MTEF/budget ceilings at the beginning of the budget process (i.e. before the Budget Guidelines are circulated). 5.20 In addition, in order to strengthen political (Cabinet-level) commitment to (and understanding of) the MTEF, it is recommended that a Cabinet sub-committee on the budget be established to oversee the MTEF/budget process. Its role would be to review the MTEF macro/fiscal resource framework, as well as sectoral polices and spending plans (MTEF sector expenditure strategies), and to make recommendations to the full Cabinet on medium-term sector/MDA ceilings in line with government priorities. It would also monitor the budget 39And thereby also the President. - 34 - process over the medium term to ensure a close linkage between the MTEF and subsequent annual budgets. The sub-committee could be supported by a technical working group of budget officials from MoF who would provide specialised advice on scrutinising MDA budget submissions. 5.21 At the sectoral level, it is recommended that a sectoral budget review committee be (re)- established to co-ordinate sectoral policy and review effectiveness of budgets and resource use against these policies. They would also review and approve the sectors MTEF expenditure strategy. These committees would comprise senior officials from across the sector, including planning and budget departments. The sector budget committee may be supported by a technical level sector working group. 5.22 Strengthen links at the sectoral level between policies and budgetary allocations. The main focus of the bottom up analytical process of linking budget resources to policies should become more strategic, with the preparation of sector expenditure strategies, which would set out sectoral expenditure policies, specific intra-sectoral priorities and the associated expenditure implications over the medium term, including expenditures from all sources of funds and covering all inputs, within the aggregate resource constraints provided by MoF. The intra- sectoral allocations could be shown according to broad programmes/objectives, for example, by service level for the education sector. For example, the Annual Sector Operational Plans, such as those produced by MoESS, provide a useful start in this regard and could be expanded to cover the medium term and include analyses of the impact of recent spending decisions.40 As part of this process, MDAs should plan and prioritise their intra-sectoral and intra-MDA expenditures according to different scenarios (e.g. in the event of particular percentage reductions in revenues during the year). It would also aim to review the cost effectiveness of on-going policies and identify potential areas of efficiency savings. 5.23 Restructure the budget format. This process would involve moving towards viewing trade-offs across and within broad programme areas for all line items (for which service delivery and new policy objectives may be made) and away from trade-offs in terms of line items (e.g. making budget choices between items 1 vs items 3 and 4). This would involve a shift in thinking by MDAs and the ability to prioritise and address expenditure trade-offs between/amongst policy areas covering all sources of funds together. If desired, detailed activity-based operational plans could be prepared by MDAs and, if necessary, used by CAGD, although in time greater management discretion should be provided to MDAs to manage their own resources within broad parameters; expenditure control would then focus on expected performance. 5.24 Simplicity in programmes works most effectively. It is important that information requested from sector ministries is not overly detailed, time-consuming or unlikely to be credible. E.g. three-year allocations for detailed activities are unhelpful when the first priority should be to get broad resource allocations more in line with sector strategies. 5.25 In the medium-term, once there is a more effective strategic budgeting framework in place, sectors and MDAs can being to introduce new elements (such as performance indicators). International experience41 would caution against introducing such indicators too quickly (see below). 40See annex on a suggested outline and contents of such an expenditure strategy. 41See annex on lessons from international MTEF experiences. - 35 - 5.26 Increase the comprehensiveness of the budget framework through the incorporation of items 1 and 2 into the overall budget planning process. This should not be done at the detailed activity level (as the analysis required is too resource-intensive) but rather at the level of broader programmes. In addition, improvements may be made to the comprehensiveness of projections of DP funding, including the coverage of less traditional sources. As part of their sector expenditure frameworks, MDAs should plan across all sources of funds and improve or introduce sector-wide budget planning. 5.27 Improve payroll planning. Enforcing existing payroll controls would go some way to improving the implementation and credibility of the budget. Provide incentives and/or sanctions to MDAs to provide full information on short term payroll requirements. Complete public sector budget negotiations prior to the beginning of the year and incorporate information in the annual budget. During budget implementation, MoF should enforce hard payroll budget constraint. Introduce payroll modelling software in an effort to improve annual and medium term projected payroll requirements. Over time, there is the need for MoF to develop and implement a plan for bringing payroll into a more sustainable state. 5.28 Improve the planning and management of public investment. Recommendations for strengthening public investment include: moving towards a common process to apply to both domestically and externally financed investments; strengthening capacities for costing, appraisal and evaluation; and overcoming delays in procurement procedures. In addition, all sources of funds should be incorporated into more comprehensive planning of investment (covering developmental and non-developmental), including requirements for on-going operational expenditures. At the MDA level, realistic planning and management of investment programme should take place within a realistic medium term fiscal framework, with MDAs looking more carefully at trade-offs between recurrent and investment spending. Finally, attention will need to be paid to strengthening analytical capacities for costing, appraisal and evaluation of investment proposals, including for value-for-money (particularly important when using non-traditional sources of financing, such as private sector partners). 5.29 Training on MTEF and prioritising expenditures. Continued training on strategic budgeting would be useful. Whilst building a better understanding of MTEF principles and good practices amongst MoF and other MDA staff is important, it cannot change the underlying systemic weaknesses in budget processes. Strengthening the process will take significant time and will require broader PFM reforms, particularly in terms of enforcing budget discipline, to be effective first. Although the planned MTEF training programme is valuable, given the wider systemic PFM challenges, it is unlikely in itself to change the strategic budgeting process significantly in the short term. 5.30 Strengthen incentives for reporting on the use and cost effectiveness of budget resources linking budgets to policies. Incentives to MDAs to identify efficiency savings could be provided, for example, through the provision of additional discretionary budget funding during budget negotiations (but only if the MDAs original submission was within the Cabinet-approved ceiling). 5.31 Improve budget credibility through increased predictability and communication in budget implementation. The two most important steps in helping improve the credibility of the budget are to enforce existing commitment controls, and to the streamline the budget implementation process through reductions in unnecessary steps. It would also be useful to provide regular communication during the year between MoF and MDAs on the expected - 36 - provision of cash ceilings and releases. Over time, more discretion for MDA budget managers could be facilitated by gradually shifting responsibility for expenditure control to their Accounting Officers (e.g. Chief Director); this would need to be supported by measures to increase accountability and enforce a system of appropriate sanctions for these Accounting Officers. 5.32 To support a stronger strategic budgeting process, development partners should provide as full projections as possible on likely medium-term disbursements of funding and ensure that these projections are provided in conjunction with GoGs budget timetable, rather than in line with DPs own national timetable.42 At the same time, DPs should ensure that their support is consistent with GoG budgetary priorities as set out in the proposed Budget Framework Paper and sector expenditure strategies. Finally, further fragmentation of DP funds through earmarking otherwise flexible sources of finance (e.g. sector budget support) should be discouraged. 5.33 A summary of the recommendations to strengthen strategic budgeting is presented in Table II.6. The list is intended to provide a menu of choices for addressing the challenges set out in the rest of the report. Diagram II.2 (an update of Diagram II.1 above) presents the proposed revised budget process. Diagram II.2: Proposed budget preparation process43 Cabinet Cabinet-apprd approval of budget to MTEF Paper Parliament (Nov) Cabinet/ and Budget Parliament Ceilings Appropriation Bill approved by Parliamt (Dec) Updating of macro-fiscal Budget Prepn of Budget framework Guidelines, Statement MoFEP ceilings issued to Prepn of MDAs Technical MTEF Paper Budget Policy (Budget Hearings, Hearings Framework Final Paper) budget MDA policy ceilings, estimates reviews MDAs Preparation of MDAs Sector expend draft budget submissions strategies prepn Jan-March April-June July-September October-December 42This is less of an issue with MDBS DPs. 43 To be read in conjunction with Diagram 1 above. The diagram by necessity does not show all of the steps, nor all of the stakeholders involved. - 37 - Table II.6: Summary of ERPFM Recommentations on Strengthening Strategic Budgeting/MTEF PFM Challenges identified by the Recommendation Sequencing ERPFM (Quick Win/ Short Term/ Medium Term) 1. Strengthen and signal political commitment early on in the MTEF/Budget process Insufficient scrutiny of budget submissions Organise study visits for MoF staff and QW by MoF, Cabinet and Parliament. Parliamentarians (Finance Committee members and clerks) to 1 or 2 relevant countries (e.g. South Africa) with stronger budget hearing processes. Identify and provide training to a group of inter- disciplinary specialists to work with MoF, the ST, MT proposed Cabinet Budget sub-committee (see recommendation below) and the Parliamentary Finance Committee to act as a regular review team for MTEF and budget requests through to appropriations. Relatively limited high-level (political) Establish/strengthen high-level working group ST commitment to the MTEF (as opposed to (e.g. Cabinet sub-committee on the budget) to annual budget). oversee the MTEF/budget process. Their TORs would involve reviewing the MTEF macro/fiscal resource framework, reviewing sectoral polices and spending plans and making recommendations to full Cabinet on medium- term sector/MDA ceilings in line with government priorities, and ensuring a close linkage between the MTEF and subsequent annual budgets. As part of the MTEF/budget process, a regular high-level workshop could be convened to discuss options for the MTEF budget framework prior to the start of the annual budget process. Budget process doesnt start with high-level Revise budget calendar to introduce ST decisions on budget allocations (ceilings). strategic/decision step, requiring Cabinet approval of ceilings before they are issued to MDAs (by May). Ceilings in outer years not credible. The recommendation below on introducing an ST, MT MTEF Strategy Paper (Budget Framework Paper), together with the recommendation to increase political commitment (i.e. at the Cabinet-level) to the defined budget ceilings should help increase the credibility of outer year budget ceilings. On the technical side, based on an analysis over time of how closely the outer year budget ceilings are to final budget ceilings (as defined in the appropriations law), review and revise the methodology for setting outer-year budget ceilings - 38 - PFM Challenges identified by the Recommendation Sequencing ERPFM (Quick Win/ Short Term/ Medium Term) Differences between indicative and final Require Cabinet approval of budget ceilings and QW, ST ceilings enforce budget ceilings Consider imposing sanctions on MDAs whose MTEF/budget submissions are above the Cabinet-approved MTEF/budget ceilings. MTEF process is in need of refreshment. Consider re-launch (rebranding) the MTEF, ST linked to new governments plans. Communicate the new process (with strong political backing) widely within government and civil society. 2. Increase transparency in the MTEF process Little consultation and communication on MTEF Strategy Paper (see recommendation ST, MT basis of ceiling setting. below) is intended to set out the justification for the setting of the expenditure ceilings and to be communicated to stakeholders. Poor interaction between MDAs and MoF Improve regular communication between MoF ST, MT in rolling over of MTEF. Weak feedback and MDAs (e.g. through the sectoral budget loop from MDAs to subsequent ceiling review committees ­ see below) on the setting. interaction between information in the sector expenditure strategies and the setting of (subsequent) MTEF/budget ceilings. Lack of overall strategic document kicking Introduce strategic MTEF/BFP analytical QW off MTEF process. document at the beginning of the annual MTEF/budget process. Medium-term revenue projections could be This process is on-going through the Tax Policy ST strengthened. Unit in MoF. Reporting on MTEF as planned Ensure that changes to the MTEF format (e.g. QW simplified programme structure) are reflected in the CAGD reporting formats. Report external financed project expenditures by MDA in CAGD reports to improve the comprehensiveness of information on budget implementation. MTEF volumes too detailed. Simplify MTEF through allocating MTEF ST resources according to sectoral/MDA ceilings and broad sectoral/MDA programmes so that the emphasis of the MTEF is on broad strategic choices. Existing detailed volumes may be used as operational plans for MDAsown use. - 39 - PFM Challenges identified by the Recommendation Sequencing ERPFM (Quick Win/ Short Term/ Medium Term) 3. Strengthen the links in practice between sectoral policies and the use of budgetary resources MTEF not comprehensive of all public MoF to make policy decision to ensure MTEF QW spending. fully reflects all public expenditures for 2010- 2012 MTEF, including Statutory Funds and remaining spending items (personal emoluments and administration). Reflect this change in 2010-2012 Budget Guidelines. Change to accompany the simplification of the MTEF (see above).44 Lack of institutional structure (sector (Re-)Introduce sectoral budget review ST, MT working groups, budget committees) to committees, which would involve senior officials review policies by sector. from across the sector, including planning and budget departments. Their role would be to co- ordinate sectoral policy and review effectiveness of budgets and resource use against these policies. They would also review and approve the sectors MTEF expenditure strategy. The sector budget committee may be supported by a technical level sector working group to provide analytical support. 44Full activity-based costing requires significant analytical capacities; this would require even more time and detail for MDAs and, given the current challenges with prioritising items 3 and 4, may prove counterproductive. - 40 - PFM Challenges identified by the Recommendation Sequencing ERPFM (Quick Win/ Short Term/ Medium Term) Weak links between sector policies and the Convene working group of officials from MoF, ST, MT outcomes and cost effectiveness of spending NDPC and key spending MDAs (e.g. MoESS, (missing middle). MoH) to discuss refocusing budget process on Emphasis of budget process and throughout the results of overall spending in the the budget cycle) is on spending items, sector/MDA. rather than what the resources are being Re-focus 2010 policy hearings on the used for. performance and cost effectiveness of on-going expenditure policies and the implications for the medium term budget. Introduce sector expenditure strategy as part of MTEF preparation (pilot expenditure strategies could be introduced for 2010-2012 and could be reviewed during policy hearings). These strategies would set out the medium term policies and priorities for the sector and the associated budgetary implications within the overall macro- fiscal resource framework (realistic Cabinet- approved sector ceilings). Incentivise MDAs to identify efficiency savings through for example providing additional discretionary budget funding during budget negotiations (but only if original MDA submission was within the Cabinet-approved ceiling). MDAs get bogged down in detailed budget The introduction of a sector expenditure strategy ST, MT submissions, which have become routine is intended to provide a less detailed and more and not necessarily linked to meeting policy strategic way of allocating resources to objectives. expenditure priorities (see above). Decentralised/silo approach to determine As above, the introduction of a sector ST, MT detailed MDA budgets ­ by MDA, within expenditure strategy is intended to provide a less MDAs. By spending item, by administrative detailed and more strategic way of allocating units within MDAs. resources to expenditure priorities. When budget parameters change, requires cumbersome prioritisation process, which in practice gets done at headquarters without recourse, communication with lower levels Little incentives for efficiency gains. As part of budget submissions, requests for (ST), MT additional resources (e.g. from an amount held back from allocating to ceilings for this purpose) could be made contingent on the identification of efficiency savings (see above). Line ministries to review the current balance of inputs to improve the efficiency and effectiveness of delivering services. - 41 - PFM Challenges identified by the Recommendation Sequencing ERPFM (Quick Win/ Short Term/ Medium Term) Limited analytical capacities to undertake On-site training for budget officers in MDAs, ST, MT detailed expenditure analysis to identify tailored to reviewing the cost effectiveness of expenditure efficiency savings and the most existing sector/MDA spending. A group of cost-effective options for achieving policy trained officers could sit on the sectoral budget objectives. review committee. Monitoring & Evaluation (M&E) not Capacity building for strengthening analytical ST, MT effectively integrated in budget cycle. capacities to undertake expenditure analysis should also include strengthening M&E capacities. Results of M&E should feed into the process of reviewing the cost effectiveness of on- going spending policies. 4. Improve budget credibility Poor payroll/wage bill planning. Enforce existing payroll controls. Provide QW, ST, MT incentives and/or sanctions to MDAs to provide full information on short term payroll requirements. Complete public sector budget negotiations prior to the beginning of the year and incorporate information in the annual budget. During budget implementation, MoF to enforce hard payroll budget constraint. Introduce payroll modelling software in an effort to improve annual and medium term projected payroll requirements. Insufficient discretion for budget managers Over time, provide more discretion for MDA MT at MDA level during budget budget mangers through gradually shifting implementation. responsibility for expenditure control to their Accounting Officers (e.g. Chief Director). Weaknesses in budget implementation such Improve regular communication during the year ST, MT that MDA budgets are not credible. between MoF and MDAs on the expected provision of cash ceilings and releases. Enforce existing commitment controls; ensure greater discipline in following the process of cash plans, ceilings, and releases. Streamline budget implementation process through reductions in unnecessary steps. Note: Further details on any of these recommendations may be made available. - 42 - A more results-orientated focus on budget preparation 5.34 Within GoG, and particularly, MoF, there is a desire to incorporate more performance elements in the budgeting process, such that future budgetary allocations could be measured in terms of progress in meeting sector/MDA performance targets. Currently, the extent of performance information is limited in the current budget process. As indicated above, "outputs" in the detailed MTEF volumes represent a kind of non-financial performance targets. The Annual Performance Review (APR) process provides information on individual sector/MDA targets and M&E indicators. However, these targets and indicators do not play a central role in the setting and scrutiny of MDA budgets. Implications for GoG 5.35 For GoG, the desire is to focus better on what it is achieving, and is able to achieve, with its limited resources. Full performance budgeting would not be appropriate for the GoG budget process in the short/medium term since it would require: sophisticated monitoring and statistical systems with good quality data; strong/efficient PFM cycle; strong analytical and scrutiny capacity, at MDAs, MoF, Cabinet and Parliament; sufficient political will, and finally, it is resource intensive. In Ghana, there is not the degree of sophistication of the monitoring and statistical systems, the extent of necessary analytical capacities and those to scrutinise the budget information for decision-making purposes, and the necessary efficiency of the broader PFM cycle (including credible and predictable budgets) necessary to introduce full performance budgeting. 5.36 What may be done, gradually, is to incorporate performance information into the budget dialogue, whilst building up some of these accompanying systems. Thus, the budget would be allocated on a broader strategic basis (e.g. programmes), and some (limited) performance information would be included in the information accompanying the proposed allocations. In practice, this means bringing some of the performance information from the APR process into the documents accompanying the budget and the budget discussions. However, experience from elsewhere would suggest that care (and time) needs to be taken when defining appropriate (and an appropriate number of) performance indicators. The increased use of performance information could be piloted in particular sectors; health and education are sectors in which PIs tend to be used more frequently. Over time (in the long term), as capacities and monitoring systems are further developed, the GoG budget process can move toward more performance- informed budgeting, where PIs take on a more central role in the budget discussions and negotiations. 5.37 Over time, it may be possible to incentivise MDAs (through allocating a limited amount of additional resources with greater flexibility/discretion for use ­ in this way, MDAs would compete for their use) to improve their use of PIs in their budget submissions and their justification to MoF during budget discussions. Rewarding or sanctioning good/poor performance by MDAs is more problematic, particularly given the complex nature of performance and the quality of monitoring information required, and should be avoided. - 43 - 5.38 All budgets which incorporate performance information need to make the distinction between on-going activities (including basic service provision) and reforms/new proposals; not all expenditures are intended to address new policies. The recommendation above of introducing a programme format for budgets provides a way of moving from the current detailed process towards a more strategic approach which can incorporate performance information. Box II.6: Overview of Performance Budgeting45 Governments around the world have increasingly been focussing on improving the results-orientation of their budget. The objective of these experiences is to improve the value-for-money for public spending and ultimately the efficient delivery of public services. Often, these reforms have been accompanied by increased flexibility for budget mangers in return for stronger accountability for the results. As with the concept of MTEF, there is no single agreed definition/model of performance budgeting: the broad definition is that it is a budget which links the funds allocated to measurable results. Nonetheless, it is to be noted that simply including performance information in budgets is a long way from performance budgeting. The latter also involves integrating performance information into the budget decision process. The following typology classifies the different types of performance budgeting: Presentational performance budgeting ­ non-financial performance information (PI) is included in budget or other documents. Information can relate to targets, or results, or both. PI is intended as background information for accountability and dialogue with Parliament and the public on public policy issues, and is not intended to play a role in decision-making. This is used in countries such as South Africa and Namibia, where budget allocation decisions are made/shown by programme ­ arguably, the most appropriate in Ghanas case. Performance-informed budgeting: resources are indirectly linked to proposed future performance or to past performance. PI is important in the budget decision-making process but does not determine the amount of resources allocated. PI is s used along with other information in the decision-making process. This is the approach undertaken by the US and Canada, for example. Direct performance budgeting: involves allocating resources based on results achieved (e.g. number of students who graduate with a Masters degree will determine the following years funding for the university running the programme. The format of the budget would be based on e.g. the MoHs allocating its resources based on vaccinating a certain number of patients. This form of performance budgeting is only practiced in a few OECD countries (e.g. Australia, UK, Sweden, Denmark, New Zealand, and the Netherlands). In practice, relatively few countries (and these are concentrated in the OECD) practice true performance budgeting (as opposed to incorporating performance information in their budget process, as described above). Other countries do not have the sufficient monitoring and data systems and analytical/scrutiny capacity to use it as intended. The broad lesson from these experiences suggests that countries should gradually (over time) integrate performance information into the budget process, but should avoid government-wide systems that tightly link performance results to resource allocations until adequate monitoring and information systems and capacity are in place. Source: OECD 45For more information, see Performance Budgeting: Policy Brief, OECD 2006. - 44 - Annex 1 Technical Appendix 1. This annex provides more analytical detail to that provided in the main assessment of MTEF current practices and processes found in the main body of the report. It does so against a fuller set of criteria for a strategic budget (Box A1.1), centred on the following basic characteristics: (i) a medium-term perspective to budgetary planning; (ii) a realistic and credible macro-fiscal resource framework, setting out the overall likely level of public sector resources over the medium term; (iii) the explicit linkage of sectoral budgetary allocations to Government policies; and (iv) more efficient utilisation of resources. The analysis below should be read in conjunction with the main report. Box A1.1: Key Features of Strategic Budgets46 THE MAINFEATURES OF ANMTEF/STRATEGICBUDGETING PROCESS INCLUDE: a multi-year time horizon for the planning of public expenditures recognising that government policies and programme initiatives may take several years to implement; inclusion of all public resources including both domestically generated revenues and domestic and external assistance financing; comprehensiveness in terms of coverage of all public expenditure both recurrent and investment and including spending from the central budget, from local budgets and from extra-budgetary funds; a focus on overall expenditures of a sector, covering all activities and organisations in the sector as a whole; it allows the Government to take decisions on strategic budget choices before detailed preparation of the annual budget is undertaken, thereby considerably enhancing the efficiency of the budget process; it provides sectors with predictable and stable projections of expected budget resources; it forms an integral element of the annual budget cycle, being rolled forward annually to maintain the forward perspective for expenditure planning; it can lead to greater transparency in the planning of public expenditure programmes through an MTEF Budget Framework Paper (or equivalent) being presented to Parliament as a government policy statement and by being made publicly available. Source: Strategic Budgeting and Sector Support, Paper for EC Guidelines (revised), February 2007. Multi-year time horizon and inclusion of all public resources 2. The aggregate macro-fiscal framework has been presented within a comprehensive multi-year time horizon since the early 2000s. This involves three-year projections (relevant budget year plus two forward years) of domestic revenues (including IGFs) and external grants, domestic and external financing (including budget support and programme and project credits and loans), and statutory and discretionary expenditures (including Statutory Funds and the use of HIPC and MDRI resources). 3. The macro-fiscal framework is reasonably comprehensive of all budgetary resources. In the 2008 Budget and 2008-2010 MTEF, an analysis of tax expenditures (in the form of import exemptions) was included for the first time. Projected expenditures for the Statutory Funds are included. As indicated in the PEFA analysis,47 the coverage of external finance in 46Based on international practice. Further specific references available on request. 47See PI-7, PEFA Performance Report for Ghana, 2006 - 45 - the budget is reasonably comprehensive (coverage of at least 90% for external assistance, at least for the coming budget year). 4. Table A1.1 sets out a summary of the types of information provided in the detailed MDA MTEF volumes. For each table type (shown in the first column), the contents of the table are shown (column 2), the number of budget years for which the information is provided, e.g. annual budget or medium-term (column 3), the links between the figures in the relevant table and data in other tables (column 4), and any inconsistencies or other comments (column 5). 5. Table A1.1 demonstrates that, whilst the MTEF aggregate macro-fiscal framework has a multi-year perspective, the MTEF at a more detailed level (i.e. detailed expenditures at MDA level) does not. In fact, at the level of the MDA, budget information is focussed primarily on the coming budget year (Table A1.1, third column). The presentation of the detailed MDA objective and activity-based allocations does not satisfy the requirement of a multi-year time horizon for the planning of public expenditures since government policies and programmes may take several years to implement but these estimates of the costs of multi-year programmes are not shown. Table A1.1: Tracking Expenditure Information in Individual MDA MTEF Volumes No. years' Table Content coverage Figures match: Remarks By function: cannot break Agg. Totals (GoG 1,2,3,4, IGF, SF, donor) function down by MDA to Total expenditures by match (w/ sl. diff.) overall figs in Budget double-check with MDA Table 1 function 2008 Statement (Appendix 6) totals Narrative list of Table 2 objectives N/A For Tables 1 and 2, grand totals are shown without SFs. Can see all expenditures (GoG 1,2,3,4 Agg. Totals (GoG 1,2,3,4, IGF, SF, donor) + IGF+SF+donor) by MoFA Expenditures by match overall figs in Budget Statement thematic area (the 3) for Table 3 admin. structure 2008 (Appendix 6) MDA (MoFA) Not clear from the tables what these represent. They Disaggregated do not seem to link up with approved budgets by Data do not appear to match figures in other elements in the Table 4 MMDA 2008 Table 2 volume. Grand total (GoG, IGF, donor included but not shown separately) for the MDA Cannot see all expenditure GPRSII Theme, Key (MoFA) matches the agg totals used to meet GPRS Focus Area, Policy (GoG+IGF+donor) in Appendix 6 less thematic/focus area and Table 5 Objective 2008-2010 Items 1 and 2 policy objective GPRSII Theme, Key Cannot see all expenditure Focus Area, Policy Details for items 3 and 4 and broken down used to meet GPRS Objective by item 3 by GoG, donor, IGF of aggregated figs in thematic/focus area and Table 6 and 4 2008-2010 Table 5 policy objective Details for items 3 and 4 (including GoG, donor, IGF) broken down by objective, Items 1 and 2 are not output and activity. Assume the totals for all assigned to an objective Tables 7- Detailed tables by cost cost centres are the same as the summary in (objective 0000), output or end centre 2008 Table 6 described above activity Note: The analysis is based on the 2008 MTEF volume for the Ministry of Food and Agriculture. 6. Whilst a positive aspect of the process is that the MTEF covers both recurrent and investment spending and thus there is not a separation of recurrent/investment processes, as in other countries, however, in practice, there is a de facto separation as developmental - 46 - investment spending tends to be funded disproportionately by development partner sources and the planning and management of many of these resources does not follow the same procedures as for domestic investment resources.48 Provides sectors with predictable and stable projections of expected budget resources 7. A key objective of an MTEF is to enable MDAs to plan their expenditures realistically over the medium term, thereby encouraging more effective and efficient utilisation of resources in implementing policies; this requires that MDAs are provided with predictable and stable projections of their likely resources over this period. As MDAs plan on the basis of their expenditure ceilings, an analysis has been undertaken of the credibility of MDAs indicative ceilings. Budget ceilings provided with the Guidelines are not credible (not the final ones) 8. One of the requirements of an MTEF is the credibility of ceilings provided to MDAs. Setting medium term ceilings is intended to enforce spending limits and to encourage MDAs to prioritise their own expenditures within a hard budget constraint. If MDAs believe that there is the possibility of negotiation on their final budget amounts, then they will not focus on prioritising their planned expenditures within a realistic resource framework. 9. In Ghana, an analysis of MDA ceilings and their eventual appropriations over the past three years shows significant differences between the indicative ceilings provided in the Budget Guidelines at the beginning of the budget preparation process by MDAs and the "final" ceilings, as indicated by the appropriated budgets. The first analysis shows ceilings by spending item (Table A1.2), whilst the second looks at the difference across MDAs (Figures A1.1-A1.3). Total differences were as high as just under 20%, as in 2007. In 2005, the final ceiling total was close to the indicative amount. For individual spending items, the differences between the initial and final ceilings were significant. In most cases, the final ceilings were greater than the initial ones, although final ceilings for one or other of service and/or investment were below the initial amounts in all three years in favour of higher ceilings for personal emoluments and/or administration. Table A1.2: Difference between indicative and final MTEF ceilings by spending item (originally denominated Cedis Billion) 2005 2006 2007 Indicative Final ceilings Diff as % Indic Final Diff as % of Indic Final Diff as % of ceilings (appropr.) of Indica- ceilings ceilings Indicative ceilings ceilings Indicative tive Personal emoluments 6,992 7,246 3.6% 8,147 9,241 13.4% 9,915 12,534 26.4% Administ-ration 1,381 1,514 9.6% 1,693 2,089 23.4% 1,968 2,562 30.2% Service 995 598 -39.9% 665 712 7.1% 774 831 7.4% Investment 1,257 1,347 7.2% 1,631 1,592 -2.4% 2,651 2,273 -14.3% Remaining MDAs Total 10,624 10,705 0.8% 12,137 13,634 12.3% 15,308 18,201 18.9% Source: MoFEP, CAGD. 48See 2007 ERPFM report. - 47 - 10. There are wide variations between indicative and final ceilings across MDAs, suggesting that some MDAs are more or less successful at obtaining (in some cases, significantly) greater resources during budget negotiations. In particular, differences between the two amounts ranged from -90% (National Commission on Culture) to +45% (Ministry of Defence and Ministry of the Interior). The data do not reveal any particular pattern apart from the fact that public administration MDAs were more likely to receive relatively higher amounts in the final ceilings than other MDAs, whilst public infrastructure (i.e. Roads and Transport) were likely to receive less. Figure A1.1: Differences between indicative and final ceilings by MDA (largest differences), 2005 % difference between indicative and final ceilings 40% 30% 20% 10% 0% -10% -20% -30% -40% -50% Source: MoFEP, CAGD. Figure A1.2: Differences between indicative and final ceilings by MDA (largest differences), 2006 % difference between indicative and final ceilings 60% 40% 20% 0% -20% -40% -60% -80% -100% -120% Source: MoFEP, CAGD. Figure A1.3: Differences between indicative and final ceilings by MDA (largest differences), 2007 % difference between indicative and final ceilings 50% 40% 30% 20% 10% 0% -10% -20% Source: MoFEP, CAGD. - 48 - 11. Indicative ceilings appear to be conservatively set due to caution regarding likely revenue for the coming budget and may be influenced by the lack of high-level commitment to the ceilings early on. However, regarding the former, as will be argued below, adjustments to revenue projections in practice are unlikely to be significant between the second and third quarters of the year, and, regarding the latter, early political commitment to budget parameters would strengthen the MTEF/budget process significantly. Consequently, MDAs may not follow indicative ceilings when they prepare their budget submissions 12. MDAs do not appear to follow the indicative ceilings given in the Budget Guidelines (Table A1.3). Overall, based on data for the 2008 budget process, budget submissions by MDAs were around one-quarter higher than their ceilings, with some very significant differences. Thus, there is a lack of realism in MDAsexpenditure plans with regard to adherence to a hard budget constraint. However, given that the initial ceilings changed quite significantly before the final allocations, it would appear that, in not following the indicative ceilings provided by MoFEP in the Budget Guidelines, MDAs were acting rationally. Table A1.3: Comparison of MDAs' Indicative Ceilings and Budget Submissions ­ 2008 Budget (GH¢ `000) Diff as % 2008 of 2008 Indicative MDA indicative Largest MDAs ceiling submission ceiling Min. of Education, Science and Sports 618,136 1,017,553 64.6% Min. of Health 252,788 281,000 11.2% Revenue Agencies 72,846 73,901 1.4% Min. of Local Gov't, Rural Dev't and Env't 33,576 35,449 5.6% Min. of National Security 21,393 27,212 27.2% Min. of Food & Agriculture 20,801 28,100 35.1% Min. of Interior 18,946 74,824 294.9% Min. of Finance and Economic Planning 12,493 26,597 112.9% Judicial Service 9,893 16,675 68.6% Office of Gov't Machinery 9,116 14,085 54.5% Min. of Land, Forestry and Mines 8,928 13,344 49.4% Min. of Transportation 7,245 9,947 37.3% Office of Parliament 7,109 9,618 35.3% Min. of Manpower Youth & Employment 4,988 9,967 99.8% Min. of Water Resources, Works and Housing 4,936 5,835 18.2% Audit Service 4,020 10,910 171.4% Remaining MDAs 348,069 149,520 -57.0% Total for all MDAs 1,455,283 1,804,537 24.0% Source: MoFEP. Ceilings over the medium-term are merely indicative 13. A key objective of an MTEF is to provide credibility of funding to MDAs over the medium term. In other words, MDAs should be able to plan their expenditures two to three years ahead based on the indicative fiscal parameters, including the expenditure allocations, given in the MTEF. With more advanced MTEFs, projected allocations for outer MTEF years are taken as the basis for preparing the next yearsbudget, with any deviations from these forward parameters for the coming - 49 - budget explained in the Budget Guidelines.49 As will be discussed below, this requires a stable macro-economy and high-level commitment to enforcing budget/MTEF ceilings. 14. For Ghana, the data show that there are significant differences between MTEF outer year projections for MDA expenditures and subsequent ceilings (Table A1.3). For 2008 allocations as a whole (total), for example, the projections for 2008 GoG discretionary expenditures in the 2006- 2008 MTEF under-projected the final ceilings by just under 7%, whilst in the 2007-2009 MTEF the projections were over-projected by nearly 10%. This is not surprising given uncertainties in the macro-economic environment (see discussion below) and a lack of commitment to indicative ceilings (cf. above discussion). 15. How credible are the individual MDAs (as opposed to total) forward projections? The variance amongst MDAs shows greater uncertainty, with overall variance in MDAs forward projections of around 20% from their final 2008 ceilings. 16. Finally, Table A1.5 also gives an indication of how credible the overall macro-fiscal framework is. Table A1.4: Accuracy of MTEF Forward Projections (bn originally denominated Cedis) Actual 2008 Actual 2008 2008 Allocation 2008 Allocation 2008 Allocation vs expected vs expected (2008-2010 MTEF) (2007-2009 MTEF) (2006-2008 MTEF) ( 2007) (2006) Total GoG Discretionary 30,435 33,736 28,476.0 -9.8% 6.9% Variance amongst MDAs2 N/A 5,914.7 6,095.9 17.5% 21.4% IGFs 2,992 15,0003 1,810 -80.1%3 65.3% HIPC/MDRI 1,832 0 2,114 N/A -13.3% Statutory Funds 7,654 7,387 6,705 3.6% 14.2% External finance 10,002 10,398 9,647 -3.8% 3.7% TOTAL 52,915 66,521 48,752 -20.5% 8.5% Notes: 1. Data exclude the additional funds from capital markets and divestiture intended for investment. 2. Data represent the sum of the absolute value of differences in MDA allocations between the 2008 budget and the projected allocations in the 2007-2009 and 2006-2008 MTEFs, respectively. 3. Data need to be confirmed. Source: MoFEP. Accurate medium-term macro-fiscal (revenue) projections are a challenge 17. It may be argued that an uncertain macro-economic environment potentially hampers the credibility of MTEF expenditure ceilings, particularly for the later (outer) years. Indeed, studies of lessons from MTEF experiences50 around the world indicate that an important criterion for success of the introduction of an MTEF is macro stability. If indicative medium-term expenditure ceilings are based on inaccurate macro-fiscal projections, or projections are difficult to make because of a volatile or uncertain macro-fiscal environment, then MDAs would be rational to treat with scepticism medium-term expenditure ceilings based on these. This uncertainty would necessarily undermine the effectiveness of an MTEF. 18. For Ghana, an analysis of revenue projections from the last three years compared to actual revenue outturns indicates that MDAs may be wise to be cautious in planning expenditures beyond 49This is the basis for the PEFA PI-12 indicator. 50See, for example, EC (2007), World Bank (2004), Holmes (2003), and Le Houerrou (2002). - 50 - the immediate budget year (Table A1.5 and Figure A1.4). Whilst projections of overall revenues and grants have been relatively close to actual receipts over the past several years, particularly with the forthcoming budget year, some categories of revenue show significant unpredictability. 19. In terms of overall receipts, over the past three years, total revenues and grants have been within 3% of projections in the relevant Budget Statement/MTEF. Indeed, over the last 5 years, there has been only one year in which outturns differed by more than 6% of the Governments projections. This relative closeness of projections is the same for an analysis of the accuracy of domestic revenue (i.e. excluding external grants, whose receipts depend on factors outside of the governments control). Indeed, this is reflected in the relatively high scores for the PEFA PI-3 indicator for the period 2003-2007. 20. The picture changes when one looks at individual categories of revenues, where there has been significantly greater variability in accuracy of projections amongst revenue categories. Amongst direct taxes, receipts from personal income tax have been consistently higher than anticipated (in some cases significantly, such as 38% higher in 2006 than estimated in the 2006-2008 MTEF/2006 Budget Statement), whilst company (profit) tax receipts have been lower than projected. 21. By contrast, projections of indirect taxes were over-projected during the last three years, recording an 11% shortfall in 2006 and a more than 7% underperformance in 2007. Unsurprisingly, the largest category of indirect taxes, VAT, has been the most variable of the sub-categories, and hence is responsible for the relative under-performance of the sub-sector, driven in recent years by lower-than-expected VAT receipts on imported goods. With strong economic growth and increased levels of imports, receipts would have been higher but for import exemptions of 1.9 trillion cedis, or 15% of indirect taxes. Exemptions are granted to assist economically-important areas, e.g. imports for VALCO and VRA; in 2006, exemptions were also awarded on oil imports in order to assist energy companies respond to the power shortages.51 22. Amongst other revenue sources, actual receipts of non-tax revenues have been greater than projected in recent years, given the success of the Non-Tax Revenue Unit in MoF, established in 2002; this has had significant success in increasing the reporting of IGFs, as well as the amount collected for both lodgement and retention. External grants have been relatively hard to predict. Recent years (e.g. 2007) have seen shortfalls in grants received. Finally, delays in planned divestitures of government assets have made overall revenue targets difficult to meet. 23. When data on 2008 revised projections become available, it will be possible to confirm if these trends have continued. Early figures indicate that the achievement of the Governments revenue targets for the year will prove a challenge. 24. It is not surprising that the accuracy projections get better as the projections get closer to the year in question; the 2007 projections in the 2007-2009 MTEF, for example, were closer to actual revenue receipts than the projections in the 2006-2008 MTEF, both overall and for individual revenue categories. 51Import exemptions were reported for the first time in the Budget Statement for 2007. - 51 - Table A1.5: Accuracy of Revenue Projections over Time1 (bn originally denominated Cedis) 2006 2006 2007 2007 2007 Projections Actual Projections Projections Actual Source of projections 2006 MTEF 2006 MTEF 2007 MTEF Direct 7,269 7,183 8,269 8,877 9,404 Personal 2,255 3,111 2,570 3,823 4,015 Self-employed 428 356 488 454 519 Companies 3,789 3,013 4,318 3,822 3,998 Other direct 797 702 894 778 872 Indirect 12,060 10,662 13,228 14,125 13,073 VAT 7,032 5,881 7,803 8,876 8,342 Petroleum 4,148 4,071 4,422 4,226 4,033 Excise 880 710 1,003 1,023 698 Inat'l trade 4,859 5,418 5,870 6,136 5,766 Total Tax Revenue 24,187 23,263 27,367 29,138 28,243 Other revenue (inc. NHIF) 1,514 1,383 1,725 1,824 2,577 NTR 711 923 810 3,175 3,383 Other 0 0 0 0 0 Grants 5,099 6,349 7,400 8,945 8,572 Total Revenue & Grants 31,511 31,918 37,303 43,082 42,775 Note: 1. Excludes tax exemptions Source: MoFEP, CAGD. Figure A1.4: Accuracy of MTEF Revenue Projections 16,000 14,000 12,000 2006 Proj 1 10,000 2006 Actual 8,000 2006 Actual 6,000 2007 Proj 1 4,000 2,000 2007 Proj 2 0 2007 Actual Direct taxes Indirect taxes Inat'l taxes Non-tax rev Grants Source: MoFEP, CAGD. 25. In order to address this issue, it is necessary to understand whether these results, particularly in terms of underperformance of indirect taxes, are the result of overly optimistic revenue forecasts or unexpectedly poor (or good) revenue performance due to unforeseen circumstances (e.g. external shocks). As can be seen from the diagrams below (Figures A1.5 and A1.6), whilst total revenues have broadly followed a reasonably steady trend over the past few years, budget projections in the most recent years have been above this trend. It is likely that expected increases from measures to improve the efficiency of tax administration have been slower than expected to achieve. Nevertheless, recent global economic pressures undoubtedly will continue to have an effect on the macro-fiscal framework. 26. MoF continues to improve its analysis of revenue measures and thereby the accuracy and comprehensiveness of its revenue projections. To improve the analytical basis for these projections, including the likely fiscal impact of proposed tax policy changes, MoFEP has recently established a - 52 - Tax Policy Unit (TPU).52 The Unit is currently staffed with personnel from Policy Analysis Division and will be supplemented by Technical Support staff (recruited from outside of the civil service). 27. In conclusion, whilst there is an argument to be made for improving the accuracy of medium term revenues projections, and thereby exercising continuing caution on the detailed macro-fiscal framework for the outer years (years 2 and particularly 3), the argument is weaker for the necessity of making significant changes in overall ceilings during preparation of the annual budget (year 1) due to uncertain macroeconomic forecasts, as there is relatively little value-added to an additional one quarters worth of data on actual receipts (i.e. in practice, there is a difference of 3-4 months between the setting of indicative and final ceilings). Figure A1.5: Overview of Revenue Performance, 2004-2008 Trillion cedis 60 50 Grants 40 Non-tax 30 Other tax 20 Excise, trade Petroleum 10 VAT 0 Direct taxes 2004 2005 2006 2006 2007 2007 2008 Budget Budget Budget Source: MoFEP, CAGD. Figure A1.6: Trends in Actual Revenue Performance, 2004-2007 Trillion cedis 50 40 30 20 10 0 2004 2005 2006 2007 Actual expenditures Source: MoFEP, CAGD. Comprehensiveness of public expenditures 28. In order to link expenditures to government policies effectively and to present the overall trade-offs between expenditure alternatives, the prioritisation of resources needs to be comprehensive of all spending and based on all public resources. As indicated above, whilst the aggregate macro-fiscal framework (the top-down part of the strategic budget process) is reasonably 52Supported by the Swiss government and GTZ - 53 - comprehensive and includes all sources of funds (Consolidated Fund, IGFs, statutory funds, HIPC, MDRI and resources from development partners), the detailed allocations to activities (in the MDA MTEF volumes) are not. In fact, the data in the detailed volumes only include expenditures on services and investment for sources of funds covering GoG discretionary, donor funds, and IGFs (the first 3 columns of Appendix 6 of the Budget Statement). The process for allocating and prioritising personal emoluments and administration expenditure (items 1 and 2, respectively), expenditures of the Statutory Funds, MDAsinternally-generated funds, HIPC funds, and some DP project resources occurs outside of the MTEF planning process. Similarly, separate processes govern the allocation of domestic and externally-financed investment. These differences result in a significant fragmentation of the budget process. In other words, for the 2008 budget, the detailed volumes cover less than half (around 42%) of total non-interest appropriated public expenditures (see Table A1.6). Thus, the MTEF currently excludes a significant share of public expenditure from the decisions on strategic resource trade-offs within overall resource constraints. Table A1.6: Expenditures included in the MTEF prioritisation process: 2008 Budget (GH¢ Million) Expenditures Share of total expenditures Total 5,060 100% of which prioritised by activity-basis in MTEF: GoG discretionary resources allocated to services 109 2.1% GoG discretionary resources allocated to investment 712 14.1% Allocation of IGFs (retained)2 299 5.9% Allocation of resources by development partners3 1,000 19.8% Sub-Total of above 2,120 41.9% Note: 1. Total expenditures (Appendix 6 of 2008 Budget Statement) include spending from all sources of funds, including domestic revenues, IGFs, HIPC and MDRI, and Statutory Funds. 2. Allocated to services and investment, in accordance with the FAA.53 3. Mainly allocated to investment. Source: MoFEP. 29. This fragmentation of the budget formulation process potentially undermines efficient inter- sectoral and intra-sectoral prioritisation of resources since it is not possible to analyse how all expenditures (including all resources and all spending items), working together, will be used to meet key policy and service delivery objectives. Whilst the Government has been quite effective with aggregate fiscal management, the MTEF has in practice been of limited assistance in helping Cabinet make the trade-offs between sectoral allocations and between major categories of spending, such as the payroll, non-wage operating expenses, and investment. At present, the wage bill is around 11% of GDP and in the years ahead the Government expects to increase public investment. The way the budget is planned makes it difficult to ensure: (i) that resources are linked to GoGs policy objectives; (ii) policies and programmes are financially sustainable, and (iii) that the resources are spent as intended to meet these objectives. 30. As the prioritisation of expenditures during the MTEF process focuses on "discretionary" expenditures, an argument is sometimes made that personal emoluments and related expenditures are not discretionary in the short term. However, they are in the medium term, which is the MTEFs focus. MDAs should be analysing how their combinations of inputs (personnel, supplies and investment) may provide services at less overall cost and provide better value for money. 53However, there is anecdotal evidence of some IGF resources being used to top-up salaries, particularly for subvented agencies. This practice is being addressed. - 54 - 31. In addition, the same fragmentation argument against earmarking of domestic budgetary revenues may be made for the earmarking of external finance (particularly in terms of earmarked sector budget support), which hinders MDAs ability to prioritise their resources according to their policy priorities. Focus on overall expenditures of a sector 32. In order to ensure that the MTEF reflects governments policy priorities in allocating public sector resources, some governments prioritise their broad expenditure allocations (expenditure ceilings) according to sectors first and then administrative institutions (i.e. MDAs) second, on the grounds that: (i) the definition of sectors is relatively less changeable than the list of ministries and departments, since the latter are intended to facilitate a particular governments political needs and may change frequently; and (ii) government policies for a particular expenditure theme may cut cross a number of administration institutions (MDAs). For example, a policy to increase access to education at both primary (basic) and secondary levels would involve the following institutions: the Ministries of Education, Science and Sports, Ghana Education Service, GET Fund management, district education offices, DACF management, and other ministries involved in education provision. It is more difficult to set and monitor policies at the first instance across institutions; indeed, the policies for GPRSII are defined in broad sectoral terms. Thus, the analysis of spending decisions (choices) should be made for sectors as a whole. 33. However, using the information in the current MTEF, it is very difficult to analyse how all expenditures (including all resources and all spending items), working together (i.e. across all spending items), will be used to meet key policy and service delivery objectives in a given sector. Whilst it is the case that the budget allocations for the coming budget year are disaggregated by function (as defined in terms of GFS2001) in a summary table in the detailed MDA volumes, this information is not used elsewhere to highlight movement of resources across sectors over time. 34. Some sectors, particularly those benefiting from significant development partner support, undertake such analyses as part of their annual joint sector reviews, which is intended to form the basis of their MTEF/annual budget preparation process. These kinds of analyses would be useful to undertake and use more widely in discussions with stakeholders (including Cabinet and Parliament) as part of the budget process. Facilitating government to take decisions on strategic budget choices 35. As an MTEF is key tool for operationalising broader strategic policy initiatives (linking budget allocations to these policies), including GPRSII and the MDGs, a critical objective of the MTEF is to provide information to policy makers on these policy-budget linkages to enable them to make the relevant strategic budget choices. The aim of the MTEF is to make the choices between different sectors explicit for policymakers and to highlight the opportunity cost (i.e. the cost of not undertaking the next best alternative activity) of funding a particular activity. Thus, the trade-offs between priority-sectoral expenditure areas (in line with policy objectives) over time should be clear. This would imply that strategic choices (relative allocations) both amongst sectors and within sectors across the different services provided by these sectors be clear and that movements in these relative allocations be shown over time. 36. In order to do this effectively, the MTEF needs to provide for policymakers at both government and Cabinet levels at the beginning of the budget formulation process: (i) a presentation of an overview of the overall macro-fiscal strategy and the assumptions behind the medium-term - 55 - resource framework and projections (top down parameters); (ii) an overview of the broad sectoral expenditure policies and the basis for the setting of medium-term expenditure ceilings (by sector and sub-sector/institution); and (iii) simplicity in presentation so that the strategic choices are clearly set out. In turn, policymakers are required to make a high-level commitment to predictability in resource allocations (i.e. endorse and ensure the enforcement of budget ceilings early in the MTEF/budget preparation process) so that MDAs can plan ahead effectively. How effectively does the MTEF/Budget Process in Ghana Fulfil These Requirements? Clarity of macro-fiscal resource framework 37. In terms of the first requirement listed above, the provision of an analysis of the medium term macro-fiscal strategy and assumptions, the Budget Statement provides policymakers with an overview of the macro-fiscal projections, but the explanation (in the Budget Statement) is focused on the coming annual budget; there is no textual analysis of the medium term. However, the Budget Statement is discussed near the end of the budget formulation process, after the MDAs have prepared their detailed budget estimates. In other words, there is not a separate MTEF policy document (i.e. separate from the annual budget documentation in the form of the Annual Budget Statement), presented to policymakers at the beginning of the budget preparation process which sets the policy framework for both the MTEF and annual budgets, which is standard for many MTEFs. 38. At the same time, the broad strategic assumptions behind the medium-term sectoral allocations are not set out in a way which makes the strategic spending trade-offs clear. Whilst Cabinet makes changes to the budget allocations, given that, as discussed above, the indicative ceilings are often quite different to the final allocations, the basis of these decisions is not clear. In other words, there does not appear to be a comprehensive and systematic analysis and overview of the strategic policy assumptions and key spending choices across sectors and, within sectors, MDAs as the basis for setting the ceilings. The Budget Guidelines, which contain the indicative ceilings sent out at the beginning of the budget preparation period, do not provide detail on the overall strategic and policy context for the MDA ceilings. At sector/MDA level, a similar type of broad overview of strategic allocations and textual analysis of the assumptions behind the allocations is also not prepared for all sectors, although, as indicated above, some sectors do prepare such strategic analyses. In other MTEFs, this type of broad, strategic analysis is fed into the discussions of the high-level decisions on cross-sectoral allocations. Basis of strategic sectora//MDA ceilings are not clear 39. For the second requirement, setting out strategic choices across policy areas clearly in order to help guide policymakers, whilst the Budget Statement sets out projected allocations by MDA and source of funds, it does so according to the administrative classification (MDAs) grouped under the three GPRS key thematic areas. However, since MDAs may include a number of disparate policy areas which do not naturally share a common strategy (e.g. education, science and sport), the presentation of allocations by MDA would not necessarily represent clear trade-offs across shared policy areas. At the same time, the three GPRS thematic areas provide an aggregation of different policy areas which, given their different objectives, could be more operational by disaggregating them. An example is good governance, which covers different issues such as defence, justice, local government, as well as public administration. Thus, it would seem useful to show a broad overview of allocations in between the very broad detail of the thematic areas and the more detail MDA level. - 56 - In many countries implementing strategic budgeting processes, allocations are shown along lines similar to COFOG54 functions, which may or may not cut across MDAs. 40. In the analyses behind the strategic projections, it is important to show both past and future movements in relative allocations over time so that policymakers may see the allocations in context. The aggregate (top-down) information in the Budget Statement includes MDA allocations for the coming three years but does not show recent past trends. Whilst this would make the presentation less compact, it would help policymakers place the medium-term projections in context. This usually involves the most recent (or two) years of actual expenditures, the current budget year (revised estimates), the coming budget year, and projections for the medium term (two years after the coming budget year). 41. Thus, MTEF/budget preparation does not begin with an early high-level discussion of the central policy trade-offs amongst priority sector areas. In some MTEFs, senior policymakers (i.e. Cabinet) are presented with a number of options for broad resource allocations, with assumptions behind each scenario set out against the background of the key government strategic priorities, before the Budget Guidelines are issued. The ceilings in the Budget Guidelines are the result of Cabinets decision on which of the options to support. Simplicity in presentation 42. In terms of the third requirement, simplicity in presentation so that strategic trade-offs are clear, the extensive (considered by stakeholders to be excessive) detail of the allocations within MDAs (in some cases, running to several volumes) prevent a clear picture of intra-sectoral expenditure trade-offs. Indeed, MDAs and others report that the MDA documents are too detailed to be useful as a strategic planning document (the level of detail hinders making choices explicit across policy areas) and, without a summary, it is difficult to see the expenditure trade-offs being made. The detailed MDA volumes are in practice more annual operational plans than strategic (medium term and explicitly based on policy priorities) expenditure strategy documents. 43. As indicated above, it is difficult to follow the trail of expenditures (see Table A1.5 above), and the lack of summaries makes it difficult to see the whole picture. At the same time, the lack of textual analysis of the strategic assumptions 44. In addition, changes to MDA figures may be made at the last minute, such that there can be internal inconsistency between the summary MDA tables and the detailed allocations in the rest of the volume.55 High-level commitment to MTEF parameters by policymakers 45. As for the final requirement, a high-level discussion and decision on, and commitment to, sectoral and MDA ceilings early in the budget process would help to provide high-level backing to expenditure ceilings and credibility to the medium-term budget process. This involves two issues: (i) when during the budget preparation cycle does Cabinet get involved in decisions on resource allocations; and (ii) what information does it have to make those decisions. Regarding the first of these, at present, Cabinet is not involved actively in the decisions behind the ceilings which begin the MTEF and budget preparation process; their involvement in the overall resource allocations takes place later on when the draft budget is presented after the (initial) detailed budget analyses are 54UNs Classification of the Functions of Government 55This issue was confirmed by stakeholders and indeed an analysis of some of the 2008-2010 MTEF documents found examples of such inconsistencies - 57 - carried out by MDAs. By contrast, in many MTEFs, a Budget Framework Paper is prepared by MoFEP and discussed with Cabinet, and the sectoral and MDA ceilings are approved by Cabinet before the Budget Guidelines are issued. Can lead to greater transparency of spending decisions 46. Greater transparency of spending decisions can lead to increased accountability, both to Parliament and to civil society, and, at least in theory, to expenditure allocations which reflect the publics needs more closely. Through providing more information on the broad parameters behind expenditure decisions and detailing the purposes for which public spending will be used, the MTEF can help lead to greater accountability for public expenditures. 47. As indicated above, the MTEF in Ghana is primarily reflected in the medium-term projections of the macro-fiscal framework and the detailed MTEF volumes prepared by MDAs. As indicated above, the MDA volumes are too detailed and insufficiently comprehensive (covering less than 50% of total expenditures) to provide a useful overview of where (for what purpose) expenditures are being spent. At the same time, the lack of overall textual analyses of the strategic (medium term) allocation decisions makes it difficult to see how budgets are being used in practice to meet key policy priorities. Provides an integral element of the annual budget cycle 48. As part of the process for introducing a more strategic phase to the budget, governments need to ensure that attention is paid to the wider budget systems on which budget plans depend. The MTEF is an integral part of the budget/PFM cycle. Weaknesses in any of the links in the chain potentially undermine the ability of the MTEF to be used to achieve Government policy objectives. Firstly, if the broad strategic allocations in the MTEF are not translated into approved allocations in the promulgated annual budget, then the resources employed preparing a strategic budgeting phase are wasted. 49. Secondly, well-made budget plans which are ignored or changed significantly during budget implementation undermine the credibility of the budget process and any strategic decisions that are made. More sophisticated planning of expenditures is futile if there are not also mechanisms to ensure that budget implementation is in line with plans, and if good information is not available on budget performance. Specifically, if the supporting budget systems (e.g. the budget classification system) are unable to allocate resources to the priority policy areas in line with the strategic budget plans, then there could be a mismatch between the use of budgetary resources and the achievement of budget policies. If accounting and reporting systems do not allow the recording and monitoring of budget programmes, then central agencies, Governments and Parliaments will be unable to monitor if budget objectives are being met. 50. Finally, in terms of scrutinising the MTEF/budget, if the executive, the legislature, and civil society do not understand the budget in its strategic form, then it is more difficult to achieve an appropriate match between budget policies and resources. 51. Turning to specific factors during budget implementation which potentially affect strategic budgeting, it is important that MDAs are able to plan how they will use their allocated resources sufficiently credibly in advance, that is, that budgeted plans are matched by the timely provision of expected resources by Treasury. This is particularly important for infrastructure projects and other - 58 - public investment, which relies on adequate planning and timely flows of resources over more than one budget cycle (see below for fuller discussion of this issue). 52. This sub-section looks at four main issues, including: (i) the ability of MDAs to implement the MTEF as planned in terms of the credibility of the budget; (ii) the within-year predictability of resource flows for the annual budget, which has an impact on the ability of MDAs to sequence their spending as planned; (iii) the ability to report on the budget as planned; and (iv) the ability of MoF to maintain the forward perspective of budgeting through linking the budget information from MDAs with the setting of subsequent ceilings (top-down process). Budget credibility remains a challenge 53. The first question asked is whether or not annual budgets, both overall and for individual MDAs, are credible. As shown in the annual Budget Deviation Index (BDI), which analyses differences between the appropriated budget and final expenditures, there have been significant deviations particularly across budget heads (MDAs) during the years studied (Table A1.7). As indicated, between 2003 and 2007, overall deviations have averaged just over 6% of the original appropriation amount. By contrast, the variance across MDAs has averaged more than 25% over the same period. Based on these data, the updated scores for the PEFA indicators PI-1 and PI-2 (focussing on the period 2005-2007) would be B and D, respectively. Table A1.7: Budget Deviations by MDA, 2003-20071 Budget Year Total expenditure deviation2 Total expenditure variance3 Variance in excess of total deviation4 2003 2.3% 15.8% 13.6% 2004 12.4% 25.9% 13.4% 2005 8.0% 33.3% 25.3% 2006 6.0% 36.3% 30.3% 20075 3.2% 21.0% 17.7% 1/Notes: 1. Data cover both discretionary and statutory expenditures, exclude contingency, and measure: (i) individual deviations between originally budgeted expenditures and actual expenditure outturns for the 20 largest budget heads; and (ii) total deviations for the remaining budget heads as a group. 2. Figures refer to the absolute value of the difference between actual outturns and original budgeted amount, expressed as % of original budgeted amount. For budgeted amounts, the original appropriated amounts were used, excluding supplementary budgets. 3. Figures refer to the sum of the absolute value of deviations as a proportion of the total (original) budgeted allocation for primary expenditure. 4. Percentage difference between expenditure deviations and expenditure variations (previous two columns). 5. CAGD preliminary data, pending confirmation by the Auditor-General. It is to be noted that data on total expenditures differ from those compiled by MoFEP. Sources: MoFEP, CAGD. 54. There is an argument56 that in the interests of containing the overall deficit and in the light of unplanned net increases in salary items during the year, displacement (net decreases) of spending from other items may take place. However, based on data from the last three years (Table A1.8), there is not clear, systematic evidence of whether or not such virement took place. What is notable from the analysis is that both deviations and variance for personal emoluments were relatively smaller in all three years than for any other type of spending; thus, there was greater success in maintaining spending on personal emoluments close to budgeted (both overall and for MDAs) than on other items. In addition, in two out of the three years studied, variance was relatively greater for service and investment than for personal emoluments and administration. However, two factors potentially distort the clarity of the results: (i) supplementary budgets were passed in both 2006 and 2007; and (ii) differences between MoFEP and CAGD data on budget outturns, giving rise to 56See for example the 2006 ERPFM report (Volume 2). - 59 - questions about which is the more correct and particularly about the comprehensiveness of CAGD data. In 2005, when there was no supplementary budget, investment spending was the only item to experience lower levels of expenditure than budgeted. Table A1.8: Deviations between MDAs' Budgeted and Actual Expenditures by Spending Item, 2005-20071 2005 2006 2007 Deviation2 Variance3 Deviation2 Variance3 Deviation2 Variance3 Personal emoluments 12.5% 22.5% 27.5% 29.7% 10.3% 22.5% Administration 44.6% 92.0% 55.2% 58.7% 49.4% 70.1% Service 23.6% 59.6% 88.4% 99.1% 35.6% 74.7% Investment -36.1% 51.8% 51.1% 69.6% 45.3% 77.1% Notes: 1. Data cover both discretionary and statutory expenditures and measure: (i) individual deviations between originally budgeted expenditures and actual expenditure outturns for the 20 largest budget heads; and (ii) total deviations for the remaining budget heads as a group. 2. Figures refer to the absolute value of the difference between actual outturns and original budgeted amount, expressed as % of original budgeted amount. For budgeted amounts, the original appropriated amounts were used, excluding supplementary budgets. 3. Figures refer to the sum of the absolute value of deviations as a proportion of the total (original) budgeted allocation for primary expenditure Sources: MoFEP, CAGD. Poor in-year predictability for MDAs of resource availability 55. The second question for determining how successfully budgets may be implemented as planned is to what extent do MDAs receive budgetary resources in the timing and amount they are expecting? Successful strategic budget implementation requires that MDAs have predictability in their receipt of funds through the year, in other words that releases from Treasury are in line with their expenditure plans. Implementation of the annual budget is based on a two-way flow of information between MDAs and MoF and vice-versa. Based on their final appropriations, MDAs provide information to MoF on their expenditure requirements for the year broken down by item (personal emoluments, administration, service and investment) and by month. These cash plans are supposed to take into account the size and timing of resources required to carry out specific expenditure plans, as set out in their detailed MTEFs, to be carried out by the MDA. In addition, for investment, MDAs submit monthly procurement plans to MoF. 56. In turn, MoF puts together an annual expenditure plan across MDAs, intended to be based on these MDAs cash plans and taking into account likely the timing of revenue flows, broken down monthly and quarterly by expenditure item. During the budget year, MoF provides MDAs with quarterly spending authorisations, known as cash ceilings, on the basis of MoFsoverall expenditure plan, and updated data on expected revenue receipts. These represent the maximum amounts that will be authorised to MDAs in the coming quarter/month and are supposed to provide advance warning to MDAs of their maximum amount of funds for spending for the given quarter/month. In the final step in the expenditure authorisation process, the Treasury issues monthly warrants, also known as releases, to MDAs, which are paper-based notification of the amount of cash resources to be released from Treasury to MDAsaccounts. 57. However, whilst the relating of the issuance of cash ceilings and releases to MDAs to cash availability has ensured that the budget has been maintained in aggregate, MDAs continue to face challenges in implementing their budgets in a predictable manner, and this can contribute to increased budget deviations for MDAs. If budget implementation processes were working in an ideal manner, cash ceilings would be expected to be issued by item in accordance with the monthly expenditure plans prepared (and regularly updated and communicated to MoF) by MDAs, followed by a similar level of releases to decentralised treasures for the given month, followed by payment amounts made in a timely manner on those items for which cash ceilings and releases were issued. - 60 - 58. In other words, how well are MDAs cash plans reflected in their cash ceilings and cash releases from Treasury? Using data from 2007, a comparison of the two (cash plans and cash ceilings) across MDAs would suggest that there are significant differences. For each MDA, deviations were analysed between its cash plans (the MDAsspending requirement according to its budget plan) for each item and for each quarter and its corresponding cash ceiling (the MDAs spending authorisation provided by MoF). The data on MDA cash plans cover GoG discretionary expenditures only and exclude MDAs for whom there were no data recorded on their cash plans (although data were recorded for these MDAs cash ceilings, the lack of comparable data on cash plans made the exercise less meaningful in these cases).57 59. What are the overall differences? Total cash ceilings for the year were 11% higher than the amounts submitted by MDAs in their cash plans at the beginning of the year. Greater differences amongst MDAs were reflected in higher variance amongst MDAs, at 17% of originally-submitted cash plans. That overall cash ceilings were higher than cash plans reflects the fact that MDAscash plans were made at the beginning of the year and reflect the originally-approved (ex-supplementary) budget, whilst the cash ceilings reflect the final revised budget.58 It is not clear why some MDAs cash plans are not recorded in the IT system covering budget implementation (as opposed to their cash ceilings, which are reflected in the system). 60. Do the analyses show any differences by item? If authorisation for all of the four spending items were given proportionally (e.g. in the event of an expected 10% revenue shortfall, budget authorisation would be reduced by 10% relative to appropriations across all items), one would expect that not to be any differences across the items. In fact, there are significant differences in the gap between MDAscash plans and their cash ceilings across the four spending items (Table A1.8). In terms of the gap between cash plans and cash ceilings, there were significantly greater differences on services and investment (items 3 and 4) than on personal emoluments and administration. This implies that there is greater discretion applied to these items when allocating resources to MDAs from Treasury during the year. Table A1.9: Difference between MDAs' quarterly cash plans and cash ceilings, 2007 (GH¢ `000) MDAs' Cash Cash ceiling vs. Cash plans Ceilings plan (% difference) Personal emoluments 781,035 797,847 2.2% Administration 79,967 78,523 -1.8% Services 30,333 19,337 -36.2% Investment 148,718 259,784 74.7% TOTAL 1,040,053 1,155,491 11.1% Source: MoFEP. 61. The same pattern (i.e. significantly greater differences between cash plans and ceilings for service and investment than for PE and administration) appears when one looks at the sum of the gaps across MDAs, with an overall variance across all four items of 17%. This translates to an average variance of 7% for items 1 and 2 and an average of 70% for items 3 and 4. Indeed, the 57These data cover just under half of total 2007 GoG discretionary expenditures; although these reduce the overall number of data points in the analysis, they reduce the likelihood of distorted results and are nonetheless are sufficient to reveal significant patterns. there were a surprising number of MDAs for whom no data were recorded, including MoFEP. 58Distortion of the data by significant additional resources provided to certain MDAs has been avoided by the fact that the ones which were the largest recipients of resources have been excluded due to lack of data recording of their cash plans - 61 - variances range from 5.5% for personal emoluments to 105% for investment spending (Figure A1.7).59 Figure A1.7: Total variance between MDAs' cash plans and cash ceilings ­ 2007 120% 100% 80% 60% 40% 20% 0% Item 1 Item 2 Item 3 Item 4 Source: MoFEP. 62. Do the analyses show any differences in the gap between MDAs cash plans and their MoFEP-authorised cash ceilings by quarter (see Table A1.10)? The MDAscash plans data show a distribution of cash requirements of just under one-quarter each for the first and last quarters, and just over one-quarter each for the middle two quarters. This effect is even stronger when one disaggregates the data by spending item, with greater distribution of spending expected in the middle two quarters. This makes sense, given the need for MDAs to plan their expenditures to allow for more time-consuming procurement procedures to undertake spending on services and investment. By contrast, the distribution of cash ceilings during the year is more heavily tilted towards the end of the year, which reflects the pattern of revenue collections (proportionally largest in the fourth quarter) as well as a desire to complete unused budget authorisations before the end of the year. 63. In terms of differences between cash plans and cash ceilings, based on the available data, the largest (positive) gap between the two measures appears in the last quarter of the year. This pattern may indicate a greater caution in issuing authorisations during the year, pending expected revenue receipts, as well as the additional expenditures approved during the third quarter of the year, which would have been authorised for spending in the fourth quarter. Table A1.10: Difference between MDAs' quarterly cash plans and cash ceilings, 2007 (GH¢ `000) MDAs' Cash Cash ceiling vs. Cash plans ceilings plan (% difference) Quarter 1 253,049 232,306 -8.2% Quarter 2 261,326 276,055 5.6% Quarter 3 270,822 258,461 -4.6% Quarter 4 254,856 388,668 52.5% TOTAL 1,040,053 1,155,491 11.1% Source: MoFEP. 59This measures shows the sum of the absolute value of differences between MDAscash plans and cash ceilings as a proportion of the cash plan amount. - 62 - 64. It does not appear to be the case that uncertainty in revenue flows has caused MoFEP to limit cash ceilings for MDAs relative to their cash plans, as the proportion of revenues received in the first three quarters of the year was greater than cash ceilings authorised (see Table A1.11). Table A1.11: Quarterly distribution of MDAs' cash plans, cash ceilings and revenue receipts, 2007 MDAs' Cash Total revenue cash plans ceilings receipts Quarter 1 24.3% 20.1% 22.7% Quarter 2 25.1% 23.9% 23.8% Quarter 3 26.0% 22.4% 25.2% Quarter 4 24.5% 33.6% 28.4% TOTAL 100.0% 100.0% 100.0% Source: MoFEP. 65. Caution should be used in making definitive conclusions from one years data. At the same time, data on cash ceilings do not necessarily translate uniquely to Treasury releases (the next stage in budget implementation) and ultimately to actual expenditures (the final stage). 66. In contrast to the two sets of analyses above, which looked at differences between MDAs overall planned budgets and their actual expenditures (overall budget deviations), and differences between MDAscash plans for the year and their budget authorisations during the year, this section looks at the rest of the budget implementation process, specifically, adding the stages of budget releases and actual expenditures. The 2007 ERPFM report examined the evolution of spending across all of the stages of budget implementation. It found that, for 2006 (see Table A1.12), recorded releases for investment spending were significantly lower as a percent of the appropriated budget than for other spending items, reflecting in part the specific (and more in-depth) procedures required for undertaking procurement (and thereby for requesting releases). It may also reflect greater flexibility in undertaking budget implementation procedures in some cases, particularly for services and investment. The lower percentage of the investment budget spent likely reflects the length of time required to undertake procurement, as well as an unrealistic increase in the investment budget. Table A1.12: Analysis of 2006 Budget Implementation by Spending Item1 (bn originally denominated Cedis) Appropriated2 Cash Releases Expenditure Releases as Expenditures Ceilings % of budget as % of budget Personal emoluments 10,011 10,066 10,066 11,415 100.5% 114.0% Administration 2,239 1,987 1,987 2,867 88.8% 128.1% Services 1,242 807 1,241 1,277 99.9% 102.8% Investment 5,506 712 1,106 2,836 20.1% 51.5% TOTAL 18,998 13,572 14,401 18,394 75.8% 96.8% Notes: 1.Data show GoG discretionary expenditures only and exclude HIPC. 2. Including original and Supplementary Appropriation Acts. Source: MoFEP, CAGD. 67. The updated analysis for 2007 shows that the same pattern holds (Table A1.13). - 63 - Table A1.13: Analysis of 2007 Budget Implementation by Spending Item1 (bn originally denominated Cedis) Appropriated2 Cash Releases Expenditure3 Releases as Expenditures Ceilings % of budget as % of budget Personal emoluments 13,176 13,141 14,675 14,188 111.4% 109.3% Administration 3,159 2,762 3,535 4,099 111.9% 147.2% Services 1,492 714 931 1,551 62.4% 88.5% Investment 6,903 3,622 4,525 5,111 65.6% 52.8% TOTAL 24,730 20,239 23,666 24,949 95.7% 97.2% Notes: 1.Data show GoG discretionary expenditures only and exclude HIPC. 2. Including original and Supplementary Appropriation Acts. 3. Data from CAGD Report to Auditor-General on the Public Accounts (Consolidated Fund); data differ somewhat from MoFEP.. Source: MoFEP, CAGD. Ability to report on the budget as planned in the MTEF is limited 68. For successful monitoring of any strategic budget, it is important that budget managers can review budget performance over time; this requires the ability to report on the budget as planned. However, the inability to report during budget implementation on the same basis as the very detailed MTEF outputs and activities hinders the linkage to overall objectives; in other words, there is currently not a mechanism to link activities to resources during budget implementation. Reporting is done down to sub-item level (equivalent to a cost centre), such as Feeder Roads Department. The existing classification system and the budget management system (BMS) used by CAGD does not include activity codes (shown as objective and output in the MDA volumes),60 so reporting on expenditures may not be done on the basis of planned activities, such as those already recorded by the NETS system.61 Linkage between bottom-up and top-down processes could be strengthened 69. The inter-action between the overall resource limits (top-down framework) through MoFEP and the detailed expenditure requests by MDAs (the bottom-up process) is essential for ensuring that subsequent MDA ceilings reflect the intra-sectoral as well as inter-sectoral strategic priorities. This defends on how effectively the analyses are used to inform medium-term inter-sectoral allocations (the interaction between the "bottom-up" and "top-down" processes); in other words, the regular interaction between the detailed sectoral analytical work undertaken by MDAs and the subsequent setting of budget ceilings over the medium term is important. In Ghana, the overwhelming focus on the annual budget during the budget process precludes effective interaction and thus an impact on subsequent ceiling setting. 60Although it appears that such codes are being prepared. 61The system is based on the National Expenditure Tracking System (NETS) to facilitate the generation of routine monthly expenditure reports based on accounting data. - 64 - Attachment 1 Strengthening Strategic Budgets ­ Recommended Elements of an Action Plan Introduction 1. The following recommended measures of an Action Plan are mainly technical in nature. However, for any of these reforms to be effective, they require sustained political commitment and a strong signal from the top (Cabinet/Minister of Finance), including at each level (e.g. Chief Directors in MDAs) that the budget process is undergoing fundamental change. 2. The recommendations are based on the principle of aiming to improve the results- orientation of the budget (the link between spending and results) and thereby the efficiency of spending (both allocative and technical). Specifically, the aim of the measures is to improve value-for-money (efficiency) of public spending and ensure that resources are used to achieve policy objectives. In other words, the measures are intended to ensure that the budgetsfocus is on results through improving accountability for, and transparency of, budget decisions. 3. The measures listed below are not based on recommending any particular type of strategic budgeting process (e.g. "zero-based budgeting" or "performance budgeting") or the methodology used in any particular country. The approach recommended is incremental and based on the existing strengths and challenges of the current budget/PFM system implemented by GoG. Sequencing 4. The sequencing of these measures is important. The sequencing adopts a realistic approach, based on: (i) addressing immediate challenges in the annual budgeting process first; (ii) whilst, at the same time, addressing wider PFM issues, specifically, strengthening budget implementation so that the budget may be implemented as planned; (iii) gradually strengthening strategic and performance elements of the budget process; and (iv) on a regular basis, addressing the need for greater analytical capacities at both central (e.g. MoF, NDPC), and MDA levels.62 5. Measures listed for Year 1 would be implemented already in the 2010 budget. Those measures indicated for years 2 and 3 could be introduced initially with the 2011 budget and then consolidated with the 2012 budget. Recommended measures from year 4 are those which are longer term in nature. 6. In order to place the measures in context, the Action Plan should be read in conjunction with the recommendations set out in the report. The suggested measures below represent a menu of options to address the challenges outlined in the main report; it is not expected (nor would it be realistic to expect) that all options would necessarily be implemented. 62The Action Plan focuses deliberately on improving the budgeting process of central government, since this was the basis for the review, and strategic improvements to sub-national government, where analytical capacities are that much more limited than central government, will take longer and are likely to follow changes at central level. - 65 - Box Att. 1.1: Suggested Action Plan by Budget Step63 Suggestions for Improving Budget Year 1 Years 2-3 Years 4- Step 2010 Budget 2011-2012 Budgets From 2013 Budget onwards Macro-fiscal framework MoF and revenue agencies to continue to MoF and revenue agencies to continue to MoF and revenue agencies to continue to maintain conservative/realistic focus on maintain conservative/realistic focus on maintain conservative/realistic focus on revenue projections (as is currently being revenue projections revenue projections done) MoF (including Tax Policy Unit) to continue MoF to continue to develop further the to develop its capacities for undertaking domestic macro-fiscal model for macro, analyses of new revenue measures economic and fiscal (particularly, revenue) Revenue agencies to continue to focus on projections increasing revenue generation through MoF (including Tax Policy Unit) to continue improvements in enforcement, joint working, to develop its capacities for undertaking and overall administration/management. analyses of new revenue measures Revenue agencies to continue to focus on increasing revenue generation through improvements in enforcement, joint working, and overall administration/management. 63The Action Plan is intended to be read in conjunction with the recommendations set out in the previous section. - 66 - Suggestions for Improving Budget Year 1 Years 2-3 Years 4- Step 2010 Budget 2011-2012 Budgets From 2013 Budget onwards Preparation of ceilings MoF to prepare strategic macro-fiscal (Budget Implement the new pay negotiation calendar: MoF to incentivise MDAs to identify scope Framework) paper (as part of Budget complete public sector budget negotiations for savings on payroll. Incentives could Guidelines ­ see below), which sets out the prior to the beginning of the year and include greater flexibility in the use of (some analytical basis (policy justification) for the incorporate information in the annual budget. of) their resources, and/or a guarantee of their ceilings. During budget preparation, MoF to enforce MDA planned allocation (particularly for priority programme areas). GoG to discuss and agree (with the unions) on hard budget constraint on payroll expenditure a new calendar/timetable for public sector pay (including contracted out staff). Provide MoF to implement MT strategy for overall negotiations with the aim of completing incentives and/or sanctions to MDAs to payroll management: in line with the civil payroll negotiations before the finalisation of provide full information on short term payroll service reform, and an appropriate upper the coming budget. requirements. ceiling on wage bill as % of GDP, MoF to MoF to introduce specific payroll hearings for propose sustainable resource framework for Introduce payroll modelling software in an MDAs to justify their staffing and associated total payroll (over time, growth rates of the effort to improve annual and medium term payroll requests. wagebill would be lower than projected projected payroll requirements. increases in GDP). This could include GoG to conduct a payroll census: improve the performance-related pay and implementation MoF to work with MDAs to improve data on quality of information on the staff strength for of measures to support the development of actual staffing levels by MDA. MDAs, including subvented agencies. MoF to staff capacities. Enhance measures for follow up on data. recruitment and retention of appropriate key During budget implementation, MoF to skills. enforce existing payroll controls. It could also include the carrying out of Improve planning of capital investment (see functional reviews, focussing on priority below). (core) functions Budget Guidelines MoF to prepare a new strategic, analytical In line with the MTEF/Budget Framework Continue implementing changes to the Budget MTEF/Budget Framework Paper for Cabinet Paper, MoF to commit to set binding spending Guidelines from previous years to explain the setting of the ceilings. as ceilings for MDAs in conjunction with justification as an expanded version of (or to Cabinet at the beginning of the budget accompany) the Budget Guidelines (existing preparation process (following Cabinet specific information in the BG could be approval of the ceilings prior to the circulation included in the Budget Framework Paper); this of the Budget Guidelines). Paper would be sent to Cabinet as a pre-cursor to circulating the ceilings to MDAs. During budget preparation, MoF to enforce MDA spending ceilings strictly. MoF to enforce MDA compliance with provisions in Budget Guidelines (see below). - 67 - Suggestions for Improving Budget Year 1 Years 2-3 Years 4- Step 2010 Budget 2011-2012 Budgets From 2013 Budget onwards Policy hearings Senior working group of officials from MoF, Re-focus 2010 policy hearings on the MDAs to prepare budget performance reports NDPC and key spending MDAs (e.g. MoESS, performance and cost effectiveness of on- in advance of the policy hearings. MoH) to discuss refocusing budget process on going (sectoral) expenditure policies and the the results of overall spending in implications for not just the coming annual MoF to incentivize (reward) those MDAs sectors/MDAs (also discussed below). budget, but also the medium term. These who prepare good performance reports in line with MoF guidelines. MoF and NDPC to prepare outline for report should also include performance of externally- which sectors/MDAs would be required to financed programmes and projects. Focus MoF to check/follow-up on whether MDAs prepare in preparation for the policy hearings. should be on what the resources are/have been budget submissions reflect the information in These reports would focus more on the results used for (not on the financial performance of the reports. of spending in terms of policies (non-financial spending items). Consider making policy hearings public performance). They would then form the basis for MoF scrutiny of MDAs budget submissions. Sector strategic planning Convene working group of officials from Consider introducing sectoral approach to Extend sector expenditure strategies (sector MoF, NDPC and key spending MDAs (e.g. strategic budgetary allocations (build on work Budget Framework Papers). These would set MoESS, MoH) to discuss refocusing budget that health, education are doing). Introduce out the strategic and fiscal analysis behind the process on the results of overall spending in sector expenditure strategy as part of MTEF intra-sectoral allocations. sectors/MDAs. preparation (pilot expenditure strategies could This working group could consider (re-) be introduced for 2010-2012 and could be introducing sectoral budget review committees reviewed during policy hearings). These to focus primarily on the results of budgetary strategies would set out the medium term spending. The sector budget committee may policies and priorities for the sector and the be supported by a technical level sector associated budgetary implications within the working group to provide analytical support. overall macro-fiscal resource framework (realistic Cabinet-approved sector ceilings). - 68 - Suggestions for Improving Budget Year 1 Years 2-3 Years 4- Step 2010 Budget 2011-2012 Budgets From 2013 Budget onwards Preparation of MDA budget MoF to commit to preparing a more simplified MoF to incentivize (reward/sanction) those Over time, as part of the budget submissions, submissions: improving link MTEF/budget format (by broad programme who prepare budget submissions in line with MDA should be encouraged to identify between MDA/sector strategies and and sub-programme) requirements of Budget Guidelines (including efficiency savings. budgets (improving results- Working group of officials (see above) to in terms of ensuring their ceilings are within On-site training for budget officers in MDAs, orientation of the budget) within discuss the introduction of sector/supra-MDA their Cabinet-approved ceilings) tailored to reviewing the cost effectiveness of Cabinet-endorsed ceilings. expenditure strategies to provide the strategic Capacity building for strengthening analytical existing sector/MDA spending. A group of and fiscal analysis behind the intra-sectoral capacities to undertake expenditure analysis trained officers could sit on the sectoral budget allocations (this would aim to fill the missing should also include strengthening M&E review committee. middle described in the report). capacities. Results of M&E should feed into Consider making distinction in budget MDAs to be encouraged to. the process of reviewing the cost effectiveness of on-going spending policies. between new and existing policy measures. MDAs to be encouraged to review existing Over time, as capacities and monitoring performance information (e.g. those in the Incorporate performance information systems are further developed, MoF may APR process) to identify small number of gradually into the budget process, whilst consider increasing the role of performance appropriate types of performance information building up systems for monitoring and good- information in budget preparation, discussions which could be monitored (and for which quality data collection. Increased use of and negotiations. there are adequate, good quality data) to performance information could be piloted in incorporate in the budget process (over time). particular sectors; health and education are MoF to implement changes to classification sectors in which PIs tend to be used more structure. frequently. Over time (much later), reviews of past MoF to develop changes to the classification performance could be incorporated in budget structure in order to reflect these programmes.discussions; sets out sector MT policy objectives outputs and planned activities; and corresponding expenditures for the medium term in line with MTEF/budget ceilings. Accountability for results (over time). - 69 - Suggestions for Improving Budget Year 1 Years 2-3 Years 4- Step 2010 Budget 2011-2012 Budgets From 2013 Budget onwards Technical hearings MoF to develop enhanced capacity for budget MoF to monitor/check/follow-up on whether Incentivise (reward) MDAs (but only if scrutiny: identify and provide training MDAs budget submissions follow the Budget original MDA submission was within the (including study tours to relevant countries) to Guidelines. Instil discipline for MDAs. Cabinet-approved ceiling) to identify a group of inter-disciplinary specialists to Reward/sanction those who do/do not follow efficiency savings and include in budget (e.g. develop specialised expertise in their sectoral the analytical requirements in the Budget those MDAs who have good proposals could area to work with MoF, the proposed Cabinet Guidelines. be rewarded with a small amount of additional Budget sub-committee and the Parliamentary resources which could be used at their Finance Committee in order to act as a regular MoF to begin to monitor/check/follow-up on discretion towards policy priorities; review team for MTEF and budget requests how well MDAs followed their budget plans accountability would be based on results through to appropriations. in previous years. achieved). Over time, MDAs to review the current balance of inputs to improve the efficiency and effectiveness of delivering services Over time, as capacities and monitoring systems are further developed, MoF may consider increasing the role of performance information in technical budget hearings. Cabinet guidance for the budget Establish/strengthen high-level working group Cabinet to implement the new budget To increase transparency and improve (e.g. Cabinet sub-committee on the budget) to preparation calendar, covering: (i) reviewing accountability, Cabinet to consider publishing oversee the MTEF/budget process. the MTEF/Budget Framework Paper; and (ii) its report of its discussion of the Budget Cabinet to commit to change in budget approving budget ceilings before the issuance Framework Paper and the approval of the process such that it agrees to: (i) review the of the Budget Guidelines to MDAs. ceilings. MTEF/budget strategy (Budget Framework) paper and discuss options (scenarios) for the MTEF/budget framework prior to the start of the annual budget process; and (ii) approve ceilings before they are issued to MDAs (by May). MoF to consider organising (encourage DPs to support, e.g. through budget networks, CABRI) for new government high-level workshop to bring together Cabinet-level colleagues to share lessons and discuss what benefits re-invigorating the strategic budget process can bring. - 70 - Suggestions for Improving Budget Year 1 Years 2-3 Years 4- Step 2010 Budget 2011-2012 Budgets From 2013 Budget onwards Parliamentary scrutiny/oversight of PFC to develop enhanced capacity for budget Consider providing MTEF/Budget Framework the budget scrutiny: GoG to organise study visits for MoF Paper to Parliament for information (not for staff and Parliamentarians (Finance approval) ­ this would help provide additional Committee members and clerks) to 1 or 2 traction for their scrutiny of the annual budget relevant countries (e.g. South Africa) with strong budget hearing processes. In-year budget monitoring Improve regular communication between MoF MDAs to develop (limited number of) MoF to require MDAs to provide (limited) and MDAs on financial budget performance. appropriate non-financial budget performance non-financial performance information on the information use of their resources throughout the year. Consider rewards for MDAs who provide appropriate performance information. Improve regular communication between MoF and MDAs (e.g. through the sectoral budget review committees ­ see above) on the interaction between information in the sector expenditure strategies and the setting of (subsequent) MTEF/budget ceilings Budget implementation ­ improved Enforce payroll controls (ex-ante Financial Improve regular communication (two-way In terms of improved links between credibility and predictability Clearance) ­ impose sanctions for non- information flow) during the year between budgets/spending and results, MoF to consider compliance. MoF and MDAs on the expected provision of requiring greater accountability for results by Improve predictability of budget releases; cash ceilings and releases. MDA Accounting Officers (AO) in exchange enforce existing expenditure controls; ensure Streamline budget implementation process for greater discretion for funds for these AOs greater discipline in following the process of through reductions in unnecessary steps. (e.g. with fewer ex ante controls) cash plans, ceilings, and releases. Introduce Refer to separate ERPFM recommendations appropriate rewards/sanctions (e.g. MoF to Ensure that changes to the MTEF format (e.g. on budget implementation commit to provide MDAs with a certain simplified programme structure) are reflected meaningful percentage of their budgets in line in the CAGD reporting formats. with MDAs cash plans). Refer to separate ERPFM recommendations Report external financed project expenditures on budget implementation. by MDA in CAGD reports in order to improve the comprehensiveness of information on budget implementation. Refer to separate ERPFM recommendations on budget implementation. Cross-cutting issues: - 71 - Suggestions for Improving Budget Year 1 Years 2-3 Years 4- Step 2010 Budget 2011-2012 Budgets From 2013 Budget onwards Comprehensiveness of the budget MoF to ensure MTEF fully reflects all public Reflect this change in 2010-2012 Budget Monitor comprehensiveness of MDAs expenditures for 2010-2012 MTEF, including Guidelines. spending plans. Statutory Funds and remaining spending items Eventually (much later), consider providing (personal emoluments and administration). Specifically, MDAs to incorporate personal emoluments and administration into their greater flexibility in mix of inputs ­ moving Change to accompany the simplification of the overall budget planning process, not at away from focus on line items. Greater MTEF/budget document (see above). detailed activity level but rather at the level of discretion for accounting officers ­ greater broader programmes. accountability for results MoF and MDAs to work to continue to improve the coverage of expected DP funding, including the coverage of less traditional sources. Investment planning MoF to implement measures to improve Incorporate all sources of funds into more MoF to scrutinise investment budget credibility for MDAs and improve comprehensive planning of investment submissions/proposals within MDAs overall predictability of resource flows, particularly (covering developmental and non- spending to ensure appropriate balance of for domestic investment. developmental), including requirements for inputs (investment and recurrent) for efficient on-going operational expenditures. spending, that investment has been planned Introduce common procedures (planning, comprehensively with a realistic MT resource appraisal and management) for investment framework, that investment proposals are projects for all sources of funds: will require realistically planned and implemented in strengthening and institutionalising (and accordance with these plans, and that there is documenting) GoG procedures and standards sufficient provision for on-going operational for selection, costing, appraisal, and spending requirements (including evaluation. maintenance) Strengthening analytical capacities for costing, appraisal and evaluation of investment proposals, including for value-for-money. - 72 - Attachment 2 Typology of MTEF Development over Time 1. This attachment sets out a typology of the different stages of development of a strategic budget, beginning with an annual, incremental budget process. The development of an effective medium-term strategic element to the budget process can take time, as it involves a process of institutional change and requires sufficiently developed policy and analytical capacities to reach more advanced levels of analyses. It is noted that an MTEF can be simpler or more complex. In practice, there exists a range of levels of development of an MTEF process, from that based on relatively simple analyses, to much more advanced forms. This is summarised in the boxes below (see Boxes Att.2.1-Att.2.2) and further explained in the paragraphs below. 2. It is important to note that from stage 1 onward, each development level involves all the basic elements of an MTEF (that is, a macro-fiscal framework, the setting out of sector ceilings and the development of sector expenditure strategies, elements which are themselves at different stages of development). This suggests that it is not helpful to talk about whether and when an MTEF is "in place". Rather, establishing an MTEF should always be considered to be an evolving process. 3. The sequencing set out below assumes that a country has controlled aggregate fiscal discipline before detailed sector work takes place. At the same time, the Ministry of Finances central budget disciplining role needs to be well established before meaningful sector work can take place.64 It also reflects the fact that detailed sector analyses are wasteful of staff time if basic expenditure control does not exist and if reliable expenditure information is not available. 64This is not a hard-and-fast rule as sector expenditure analyses can help sector ministries to prioritise their intra-sectoral expenditures and thereby assist with sector reforms. In general, however, if the Ministry of Finance has not established fiscal discipline in relation to sector ministries then there is the risk of detailed and well-planned expenditure plans being undermined at the Ministry of Finance. - 73 - Box Att.2.1: Typology of MTEF development In stage 0 (the stage of preparing to introduce an MTEF process), the basics of the legal and institutional framework (e.g. amendments to the Organic Budget Law, establishment of a Government MTEF Steering Group65, responsibility for co- ordinating the MTEF given to a department in the Ministry of Finance such as the Budget Department, etc.) have been established or are under way. In some countries the legal and institutional framework is the foundation for the beginning of a two-stage strategic budgeting process. What is important in the legislation is reference to a two-stage process (rather than necessarily specifically introducing an MTEF per se) whereby there is a stage prior to the start of the annual budget process where Government reviews the fiscal framework. Whilst not having the legal and institutional framework in place does not necessarily prevent countries from beginning such a process, it appears to be an important legitimising first step. In stage 1 (basic MTEF process), once the basic legal and institutional framework has been established, countries starting implementation of a two-stage strategic budgeting process begin with the development of a basic multi-year macro/fiscal framework, largely based around an IMF framework, and reflecting domestic revenues. There is limited policy analysis of the overall resource framework. No sector expenditure strategies are in place or there is an initial sector review, setting out the sectors policy objectives and classifying sectoral activities into broad programme areas to meet the objectives. Sector ceilings are based on overall policy priorities and limited analyses of the main economic categories of expenditures. In stage 2 (intermediate MTEF process), countries undertake more advanced analyses for the macro/fiscal framework and sector expenditure strategies. Macro/fiscal projections have been prepared on the basis of a simple domestic model and more detailed macro/fiscal policy analysis and the framework includes external grants and budget support resources. The setting of overall expenditure ceilings is based on some cross-sectoral analyses (overall government priorities e.g. poverty reduction, investment/recurrent ratio etc.). On the sector side, sector ministries (perhaps initially at the pilot stage) have begun to analyse their existing activities by broad programme and activity in order to prioritise existing activities and identify those which are no longer deemed to be priorities. The budget allocation mechanism has been changed to enable the budget to reflect more clearly government policy priorities by programme/activity, as reflected in sector expenditure strategies, in addition to the items of expenditure. Initial changes to the budget classification to reflect sectoral and ministry programmes are being put in place. When a country is at stage 3 (advanced MTEF process), the macro/fiscal framework includes all resources, including external project finance, extra-budgetary finance and off-budget resources. Macro/fiscal projections are based on a more sophisticated forecasting model and on detailed policy analyses of proposed revenue policy changes and macro feedbacks. Alternative macro/fiscal scenarios may be given in an MTEF Options Paper or equivalent. Sector ceilings are based on detailed analyses of cross-sectoral issues. With respect to sector expenditure strategies, initial or detailed costings of existing activities (both recurrent and capital) have been carried out, areas of cost savings and improvements to programme efficiency and operations have been identified, and broad expenditure implications of the new or restructured programme activities have been calculated. In this level, new prioritised activities to achieve policy objectives have been identified, a clear mechanism has been established for prioritising against policy objectives, and a mechanism is in place for allocating resources based on cutting or phasing out non-priority activities and funding resources to the highest priority and most cost- efficient activities. At sector level the output at this level is a detailed, fully-costed and prioritised medium-term sector expenditure strategy, with medium term expenditure plans which contain detailed expenditure (recurrent and investment) implications for programmes/activities and which are consistent with sector ceilings. In the most advanced cases, performance measurement indicators and analyses of sector performance are carried out66, and sector budgets have been restructured to realise cost and efficiency savings. Finally, a mechanism to cost out and prioritise new proposals and integrate them into the MTEF has been established. At this level, changes to the budget classification and Chart of Accounts are fully in place, and budgets are executed in accordance with programmes and activities linked to policy priorities. Source:Betley, Background paper for World Bank study on Budget Management and PRSPs, 2004. 65 While it is understood that the same agencies should be involved in both the strategic and detailed estimate preparation phases of an integrated MTEF/budget process (i.e. that the MTEF process should precisely not be the responsibility of separate groups or bodies), it is often useful to initiate the process of developing an MTEF by entrusting a steering body with the task of ensuring that this new way of doing business is harmoniously mainstreamed with existing processes and that those are adjusted as and when need be. 66 There is a continuum with respect to including performance elements in the budget process, from merely introducing performance indicators alongside budgets, for sectors to monitor and report on in their sector expenditure strategies, to a process of linking budget allocations not to broad programmes of activities but rather to specific outputs or performance targets. See P. Penroses PFM training course (Unit 6: Aggregate Budgeting and Performance), September 2006. Countries at level 3 in the development of an MTEF process are likely to start at the simpler end of the continuum. The more advanced along this continuum a country moves the greater are the analytical capacity requirements. - 74 - Box Att. 2.2: Summary of stages of development of an MTEF process Stage of Stage 0 Stage 1 Stage 2 Stage 3 development of Pre-MTEF Basic MTEF Intermediate MTEF Advanced MTEF MTEF MTEF Elements Macro-fiscal Establishment Based on IMF- Simple domestic Inclusive of all resources; framework (top- of basic legal type framework; model; stronger more sophisticated model; down) and focus on macro/ fiscal alternative scenarios may institutional domestic analysis; inclusion of inform decision-making parameters. revenues external finance Setting of sector Overall policy Some analysis of Deeper/more ceilings (top- priorities, inter-sectoral comprehensive analysis of down) little/no analysis priorities incl. basic inter-sectoral priorities policy priorities and cross-cutting issues Sector expenditure At most, initial Intra-sectoral broadly Costing of activities and strategies sector review prioritised strategic/ stricter prioritisation (bottom-up) programme mechanisms inform intra- framework in place sectoral strategic expenditure framework. Source: Betley, Background paper for World Bank study on Budget Management and PRSPs, 2004. - 75 - Carlos B. Cavalcanti C:\Documents and Settings\wb22579\My Documents\Ghana\Ghana-2009 External Review of Public Financial Management-draft Volume II-March 9 2009.docC:\Documents and Settings\wb22579\My Documents\Ghana\Ghana-2009 External Review of Public Financial Management-draft Volume II-March 6 2009-clean.doc 09/03/2009 18:58:0008/03/2009 16:01:00 - 76 -