Report No. 47767-BD Bangladesh Public Expenditure and Institutional Review Towards a Better Quality of Public Expenditure (In Two Volumes) Volume I: Main Report June 2010 Poverty Reduction and Economic Management Sector Unit South Asia Region Document of the World Bank CURRENCY AND EQUIVALENTS Currency Unit = Bangladeshi Taka (Tk) US$1 = Tk. 69 (March 12, 2009) GOVERNMENT FISCAL YEAR July 1 – June 30 Vice President: Isabel Guerrero Country Director: Xian Zhu Sector Director: Ernesto May Sector Manager: Joel Hellman Team Leader: Alma Kanani ii TABLE OF CONTENTS ACKNOWLEDGMENTS ...................................................................................................................................... VII LIST OF ACRONYMS AND ABBREVIATIONS ....................................................................................................... VIII EXECUTIVE SUMMARY ........................................................................................................................................1 PART I ........................................................................................................................................................... 14 PUBLIC FINANCIAL MANAGEMENT OUTCOMES ................................................................................................. 14 CHAPTER ONE: MAINTAINING AGGREGATE FISCAL DISCIPLINE .......................................................................... 15 1.1 MACRO-FISCAL OUTCOMES ........................................................................................................................15 1.2 MAINTAINING FISCAL SUSTAINABILITY ...........................................................................................................20 1.2.1 The Fiscal Performance of the Public Sector ................................................................................. 20 1.2.2 Central Government Long-run Sustainability ................................................................................. 21 1.2.3 Central Government: The Extended Accounting Framework ......................................................... 21 1.2.4 Consolidated Public Sector: Extended Accounting Framework ..................................................... 22 1.3 DEBT DYNAMICS AND SUSTAINABILITY ...........................................................................................................23 1.3.1 External Debt Sustainability ........................................................................................................... 23 1.3.2 Public Debt Sustainability .............................................................................................................. 24 1.4 KEY RISKS AND PRIORITIES IN THE MEDIUM TERM ............................................................................................24 1.5 IMPROVING REVENUE PERFORMANCE: THE CORNERSTONE OF THE MEDIUM-TERM FISCAL STRATEGY ...........................27 1.5.1 Modernizing the Value Added Tax System ...................................................................................... 30 1.5.2 Income Tax: The Need to Start Again ............................................................................................. 31 1.5.3 Modernizing Revenue Administration ............................................................................................. 32 1.6 KEY CONCLUSIONS AND RECOMMENDATIONS ..................................................................................................33 CHAPTER TWO: ALLOCATING RESOURCES TO BROAD STRATEGIC PRIORITIES: WHAT IS THE BUDGET FINANCING? ........................................................................................................................................................... 35 2.1 GOVERNMENT’S STRATEGIC PRIORITIES .........................................................................................................35 2.1.1 Social Sectors and Rural Development Remain in the Lead ........................................................... 38 2.1.2 Infrastructure Spending Overall Has Declined............................................................................... 39 2.1.3 General Administration Spending Remains Significant .................................................................. 39 2.2 THE ANNUAL DEVELOPMENT PROGRAM – INVESTMENT POLICY AND PERFORMANCE.................................................40 2.2.1 Priorities of the Investment Program .............................................................................................. 40 2.2.2 Bangladesh’s Two Parallel Budgets ............................................................................................... 41 2.2.3 Performance of the ADP Portfolio .................................................................................................. 43 2.2.4 Budget Deviation Indexes ............................................................................................................... 44 2.3 KEY CONCLUSIONS....................................................................................................................................45 CHAPTER THREE: ARE THE INTENDED OUTCOMES BEING ACHIEVED AT THE SECTOR LEVEL: OPERATIONAL EFFICIENCY .......................................................................................................................................... 47 3.1 SUMMARY OF CROSS–CUTTING CHALLENGES ...................................................................................................47 3.2 EDUCATION SECTOR: POLICY, FINANCING AND EXPENDITURE OUTCOMES ..............................................................49 3.3 HEALTH SECTOR: POLICY, FINANCING AND EXPENDITURE OUTCOMES ...................................................................53 3.4 TRANSPORT SECTOR: POLICY, FINANCING AND OUTCOMES ................................................................................56 3.5 THE ENERGY SECTOR: POLICY, FINANCING AND OUTCOMES ................................................................................60 3.6 THE AGRICULTURAL SECTOR: POLICY, FINANCING, AND OUTCOMES ......................................................................64 3.7 CONCLUSIONS AND RECOMMENDATIONS .......................................................................................................66 iii PART II ........................................................................................................................................................... 68 INSTITUTIONAL INCENTIVES AND POLITICAL ECONOMY CONSTRAINTS TO BETTER PFM OUTCOMES .................. 68 CHAPTER FOUR: INSTITUTIONAL CHANNELS FOR INFLUENCING BUDGET OUTCOMES ........................................ 69 4.1 INSTITUTIONS FOR AGGREGATE FISCAL DISCIPLINE ............................................................................................69 4.2 INSTITUTIONS FOR ALLOCATIVE EFFICIENCY .....................................................................................................74 4.3 INSTITUTIONS FOR OPERATIONAL EFFICIENCY ..................................................................................................77 4.4 KEY CONCLUSIONS....................................................................................................................................81 PART III ........................................................................................................................................................... 82 LOOKING AHEAD—THE MEDIUM-TERM HORIZON ............................................................................................. 82 CHAPTER FIVE: INTRODUCING A PERFORMANCE ORIENTATION IN BUDGETING ................................................. 83 5.1 INTRODUCTION OF THE MTBF .....................................................................................................................83 5.1.1 Objectives........................................................................................................................................ 83 5.1.2 The Key Elements of the MTBF ...................................................................................................... 85 5.2 ACHIEVEMENTS AND IMPACT OF MTBF ON INCENTIVES .....................................................................................88 5.3 BUILDING A SUSTAINABLE MTBF: THE ANCHOR FOR PFM REFORMS ....................................................................91 5.3.1 Strengthening Political Engagement in the MTBF/Budget Process ............................................... 91 5.3.2 Strengthening the Management Role and Flexibility of Line Ministries–A Priority for Improving Operational Efficiency .................................................................................................................................. 92 5.4 ADDRESSING THE BUDGET INTEGRATION AGENDA – THE CORNERSTONE OF IMPROVING EXPENDITURE OUTCOMES...........94 5.4.1 Short-Term Priority Actions............................................................................................................ 95 5.4.2 Longer-Term Challenges in Developing an Integrated Budget ...................................................... 95 5.4.3 Institutional Roles and Capacity Building Requirements ............................................................... 98 5.5 STRENGTHENING THE DEMAND SIDE FOR BETTER BUDGET OUTCOMES ................................................................ 101 5.5.1 Involving Parliament and Civil Society ........................................................................................ 101 5.5.2 Improving the Presentation of the Budget and Increasing Transparency..................................... 102 5.5.3 Strengthening the Role of the C&AG ............................................................................................ 103 CHAPTER SIX: INCREASING DEMAND FOR BETTER BUDGET OUTCOMES THROUGH FISCAL DECENTRALIZATION 104 6.1 SUBNATIONAL GOVERNMENT STRUCTURES IN BANGLADESH ............................................................................. 104 6.2 SUBNATIONAL GOVERNMENT FINANCE ........................................................................................................ 106 6.3 EXPENDITURE ASSIGNMENT – SUBNATIONAL GOVERNMENT FUNCTIONS .............................................................. 111 6.4 PLANNING AND BUDGETING AT LOCAL LEVEL ................................................................................................. 115 6.5 KEY POLICY DIRECTIONS ........................................................................................................................... 116 APPENDIX ONE: THE STRUCTURE OF THE PUBLIC SECTOR IN BANGLADESH ...................................................... 119 APPENDIX TWO: CONSOLIDATED PUBLIC SECTOR DEFICIT – CONCEPTS AND DEFINITIONS ............................... 121 APPENDIX THREE: FRAMEWORK FOR LONG-TERM SUSTAINABILITY ASSESSMENT ........................................... 123 APPENDIX FOUR: BUDGETARY PROCEDURES AND INSTITUTIONS IN BANGLADESH .......................................... 129 APPENDIX FIVE: PEFA MATRIX (FROM DFID’S IMPACT ASSESSMENT OF PFM REFORMS IN BANGLADESH, 2007) ......................................................................................................................................................... 153 REFERENCES.................................................................................................................................................... 165 iv LIST OF BOXES Box 1-1: Macroeconomic Assumptions Underlying the Debt Sustainability Analysis (DSA) ...................... 26 Box 4-1: External Audit in Bangladesh ........................................................................................................ 73 Box 4-2: Political Developments and the Way to Bureaucratic Centralization .......................................... 74 Box 5-1: GoB Objectives in Introducing the MTBF ...................................................................................... 85 Box 5-2: Coverage of the MTBF .................................................................................................................. 86 Box 5-3: Evaluation and Finalization of Ministry Budget Frameworks ....................................................... 87 Box 5-4: Line Ministry Reactions to the MTBF – the Local Government Division................................................... 91 Box 5-5: Political Engagement in the Budget Process – the South African Example .................................. 92 Box 5-6: Cadre System Issues...................................................................................................................... 98 LIST OF FIGURES Figure 1-1: The Fiscal Deficit Has Been Declining (% of GDP) ..................................................................... 17 Figure 1-2: Foreign Financing Has Been Declining (% of GDP).................................................................... 17 Figure 1-3: Positive Growth Performance (% of GDP) ............................................................................... 17 Figure 1-4: Increase in Central Government Debt Reversed (% of GDP) .................................................... 18 Figure 1-5: Share of Budget Allocation to SoEs Under ADP (%).................................................................. 19 Figure 1-6: Trends in Central Government Expenditure (% of GDP) .......................................................... 19 Figure 1-7: Per Capita GDP and Tax-To-GDP Ratio: Low-Income Countries ............................................... 28 Figure 1-8: Tax Revenues (2005) (% of GDP) .............................................................................................. 28 Figure 2-1: Functional Composition of Total Expenditure (% of GDP) ........................................................ 36 Figure 2-2: Functional Composition of Total Expenditure (% of Total Expenditure) .................................. 36 Figure 2-3: Expenditure Trends – Recurrent and Development Expenditure ............................................ 36 Figure 2-4: Interest Payments: Bangladesh and Countries in South Asia (% of GDP) ................................ 37 Figure 2-5: Interest Payments: Bangladesh and Countries in South Asia (% of Total Expenditure) ........... 37 Figure 2-6: Composition of Social Expenditure (FY00-FY07) ...................................................................... 38 Figure 2-7: Composition of Infrastructure Expenditure (FY00-FY07) ......................................................... 39 Figure 2-8: General Administration and Security Spending (FY00-FY07) ................................................... 40 Figure 2-9: Development Expenditure Shares (FY00-FY07) ........................................................................ 41 Figure 2-10: Composition of Development Expenditure (by Economic Classification) .............................. 42 Figure 3-1: Enrollment Has Shifted Away From Government-Recognized (and Funded) Schools ............ 50 Figure 3-2: Distribution of Total Health Expenditures (by Provider) .......................................................... 55 Figure 3-3: MoHFW Expenditures by Source of Funding (1998/9 to 2005/6) ............................................ 55 Figure 3-4: Transport ADP Allocation by Subsector (FY2003-2007) ........................................................... 58 Figure 3-5: RHD ADP Portfolio Average Duration (in Years up to June 2007) ............................................ 59 Figure 3-6: Access to electricity has improved…......................................................................................... 61 Figure 3-7: Energy SoEs: Profit/Loss ........................................................................................................... 62 Figure 3-8: Energy Subsidies Impose a High Opportunity Cost but are Largely Inequitable ...................... 63 Figure 3-9: Public Expenditures on Agriculture Have Increased, but Agricultural Subsidies Dwarf Expenditures on Public Goods .................................................................................................................... 65 Figure 4-1: Effective Aggregate Fiscal Discipline Index and Inflation ......................................................... 72 Figure 4-2: Effective Allocative Efficiency Index and Public Expenditure on Education ............................. 77 Figure 4-3: Effective Operational Efficiency Index and Public Expenditure on Education ......................... 80 Figure 4-4: Budget Planning, Execution and Reporting at the Line Ministry .............................................. 81 Figure 5-1: MTBF and Annual Budget Preparation ..................................................................................... 89 v Figure 5-2: MTBF Committees and Technical Working Groups .................................................................. 90 Figure 6-1: Subnational Government Structure ....................................................................................... 105 Figure 6-2: Bangladesh is One of the Most Fiscally Centralized Countries in the World ......................... 107 Figure 6-3: UP Revenue Sources ............................................................................................................... 108 Figure 6-4: Pourashava Revenue Sources ................................................................................................. 109 Figure 6-5: City Corporation Revenues ..................................................................................................... 110 Figure 6-6: UP Expenditures ..................................................................................................................... 112 Figure 6-7: Pourashava Expenditures ....................................................................................................... 113 Figure 6-8: Expenditure by Type of Local Government ............................................................................ 113 Figure 6-9: Per capita Development Spending Varies Across Districts and Districts With Higher Poverty Rates Spend Less ....................................................................................................................................... 115 LIST OF TABLES Table 1-1: Fiscal Trends: Bangladesh, Low-Income Countries and South Asia Region (% of GDP)............. 16 Table 1-2: Central Government Debt Stock, 2000-2008 (% of GDP) .......................................................... 18 Table 1-3: Public Sector Fiscal Performance (% of GDP)............................................................................. 20 Table 1-4: Sustainable Primary Balances (% of GDP) .................................................................................. 21 Table 1-5: Sustainable Levels of Deficit (% of GDP) .................................................................................... 22 Table 1-6: Central Government Debt Dynamics (FY99-07)(% of GDP) ....................................................... 23 Table 1-7: Debt Indicators........................................................................................................................... 24 Table 1-8: Medium-Term Budget Framework (% of GDP) .......................................................................... 27 Table 1-9: Revenue Elasticity in Bangladesh ............................................................................................... 28 Table 1-10: Revenue Productivity Ratios for Bangladesh and Comparators .............................................. 29 Table 1-11: Share of Trade and Nontrade Taxes for Selected Asian and African Countries....................... 29 Table 1-12: Bangladesh Tax Expenditure (FY2005/06)(% of GDP) .............................................................. 30 Table 2-1: Interest Payments: Bangladesh and Countries in South Asia .................................................... 37 Table 2-2: Composition of Development Expenditure ............................................................................... 42 Table 2-3: ADP: Resource Allocation vs. Cost of ADP Projects/Programs (FY07) ....................................... 44 Table 2-4: Bangladesh ADP Budget Deviation Index .................................................................................. 45 Table 3-1: Government Financing Modalities for Education ...................................................................... 48 Table 3-2: Internal Efficiency (%) of the Education System (2005)............................................................. 51 Table 3-3: Composition of Total Public Education Spending (Various Years) ............................................. 52 Table 3-4: Key Health Status Indicators (1990-2007) ................................................................................. 54 Table 3-5: Distribution of Curative Care Subsidies ..................................................................................... 56 Table 3-6: Transport Sector Expenditure by Sub-sector for FY 2003-2007 (Billions of Taka) ..................... 58 Table 4-1: Aggregate Fiscal Discipline Indexes of Budgetary Institutions .................................................. 70 Table 4-2: Indexes of Effectiveness Characteristics Affecting Budget Discipline........................................ 71 Table 4-3: Indexes of Institutional Arrangements for Allocative Efficiency................................................ 75 Table 4-4: Indexes of Institutional Arrangements for Operational Efficiency ............................................ 78 Table 4-5: Indexes of Effectiveness Characteristics Affecting Budget Discipline........................................ 79 Table 6-1: Revenue Composition of Individual City Corporations ............................................................ 110 Table 6-2: Fiscal Equality Measures for Deconcentrated Expenditures (2006/7) .................................... 114 vi ACKNOWLEDGMENTS The Public Expenditure and Institutional Review (PEIR) is a joint product of the World Bank (WB) and the United Kingdom’s Department for International Development (DFID). It has been prepared by a team led by Alma Kanani (author of the report) based on background papers prepared by Zahid Hussain (macro- fiscal analysis), Zaidi Sattar (Revenue Administration), Samer Al-Samarrai (Education), Tania Dmytraczenko, Tahmina Begum, Suraiya Zannath (Health Nutrition and Population), Binyam Reja and Reefat Sultana (Transport), Sebastian Eckardt and Raihan Elahi (Energy), Sebastian Eckardt (Agriculture), Mirafe Marcos (Review of the Annual Development Program), Shilpa Pradhan (Vulnerability to Mismanagement), Hadi Esfahani (Institutional Assessment and Political Economy Analysis), Andrew Bird (Review of the MTBF experience), and William Fox, Balakrishna Menon, Sebastian Eckardt (Subnational Finance), Zafrul Islam, Ahmad Ali (Integrating Procurement and Budgeting). Farria Naeem and Sanjana Zaman provided research assistance. A number of colleagues have also provided background information and reviewed early versions of the background papers, including Raisuddin Ahmed, Mohinder Mudahar, Johannes G.P. Jansen, Ceren Ozer, Alan Townsend, Jose Edgardo Campos. The report has also benefited from earlier technical assessments of PFM institutions and PEFA assessments done by P.K. Subramanian, Ismaila Cessay, Suraiya Zannath and DFID team on the impact assessment of the Financial Management Reform Program led by Ken Robson. Ernesto May (Sector Director), Joel Hellman (Sector Manager) and Vinaya Swaroop (Lead Economist) provided overall guidance. Shahnaz Sultana Ahmed, Jyoti Sriram, Mehar Akhter Khan, Midred Gonsalvez and Marinella Yadao, assisted with the formatting and processing of the report, as well as with the preparations for the various consultation meetings organized in Dhaka. DFID counterparts have provided comments on selected papers through Tom Crowards, DFID’s focal point on the task. The report has benefited from consultations and presentations of early findings at various stages. In addition to the government counterparts, the concept note was presented and discussed by the local group of development partners working on PFM issues. A number of discussions have been held at the technical level with GoB on selected issues from the background papers with the Bangladesh counterparts, including Arastoo Khan (Additional Secretary MoF), Ranjit Chakraborty (Joint Secretary Budget), Mohiuddin Khan (Deputy Secretary MFO), Muslim Chowdhury (Deputy Secretary, MoF) and several other officials from MoF, Planning Commission, Line Ministries and development partners who have attended the various presentations. Early findings of the background papers were presented for comments to the entire country team. Two presentations on early findings from the PEIR work have been held in workshops within the Bank: one at the “Public Expenditure Analysis and Management” workshop and the other at the “Frontiers in PFM” workshop. The report has been prepared in parallel with the Poverty Assessment for Bangladesh with which it shares the background work on the Benefit Incidence Analysis in Health and Education. It has also been prepared in parallel with the government’s efforts to put in place a follow-up five-year program of assistance on PFM. Peer reviewers are Bill Byrd (Lead Economist, SASPF), Marijin Verhoeven (Lead Economist, PRMPS), Parminder Brar (Lead Financial Management Specialist, AFTFM), Yatsuhiko Matsuda (Senior Public Sector Management Specialist, EASPR). The report was edited by Chris Stewart. vii LIST OF ACRONYMS AND ABBREVIATIONS ADB Asian Development Bank IMTP Integrated Multimodal Transport Policy ADP Annual Development Program LGD Local Government Division APR Annual Program Review LGED Local Government Engineering Department BANBEIS Bangladesh Bureau of Educational LTU Large Taxpayer Unit Information and Statistics BB Bangladesh Bank LM Line Ministry MBF Ministry Budget Framework BBS Bangladesh Bureau of Statistics MDG Millennium Development Goals BDHS Bangladesh Demographic and Health MIS Management Information System Survey BDT Bangladesh Taka MMR Maternal Mortality Rate BINP Bangladesh Integrated Nutrition MoF Ministry of Finance Program BMC Budget Management Committee MoHFW Ministry of Health and Family Welfare BMRC Budget Monitoring and Resource MoPME Ministry of Primary and Mass Committee Education BPC Bangladesh Petroleum Corporation MSA Management Support Agency BPDB Bangladesh Power Development Board MSR Medical and Surgical Requisites BR Bangladesh Railway MTBF Medium-Term Budgetary Framework CAAB Civil Aviation Authority of Bangladesh MTBO Medium-Term Budget Outlook CAMPE Campaign for Popular Education MTBPS Medium-Term Budget Policy Statement CFA Central Finance Agency MTMF Medium-Term Macroeconomic Framework CGA Controller General of Accounts NBR National Board of Revenue CPuS Consolidated Public Sector SCB State Commercial Bank DB Development Budget NDB Non Development Budget DC District Council NEC National Economic Council DDCC District Development Coordination NGO Nongovernmental Organization Committee DFIs Development Financial Institutions NHA National Health Accounts DGFP Directorate General of Family Planning NILG National Institute for Local Government DGHS Directorate General of Health Services NIPORT National Institute of Population Research and Training DP Development Partner NNP National Nutrition Program DPE Directorate of Primary Education NPV Net Present Value DPHE Department of Public Health Engineering NSAPR National Strategy for Poverty Reduction DSA Debt Sustainability Analysis OP Operational Plan viii ECNEC Executive Committee of the National PAD Project Appraisal Document Economic Council ERD Economic Relations Division PER Public Expenditure Review FD Finance Division PMA Performance Monitoring Agency FMAU Financial Management and Audit Unit SDP Strategic Development Plan FMRP Financial Management Reforms Project PRGF Poverty Reduction and Growth Facility GDP Gross Domestic Product SoE State-owned Enterprises GoB Government of Bangladesh SWAp Sector Wide Approach HDI Human Development Index TFR Total Fertility Rate HEU Health Economic Unit THE Total Health Expenditures HIES Household Income and Expenditure U5MR Under 5 Child Mortality Rate Survey HNP Health, Nutrition and Population UESD Utilization of Essential Services Delivery HNPSP Health, Nutrition and Population Sector UHC Upazila Health Complex Program HPI Human Poverty Index UDCC Upazila Development Coordination Committee HPSP Health and Population Sector Program UNO Upazila Nirbahi Officer IEDCR Institute of Epidemiology, Disease UP Union Parishad Control and Research IMED Implementation Monitoring Evaluation WASA Water Supply and Sewerage Department Authority IMR Infant Mortality Rate WHO World Health Organization ix EXECUTIVE SUMMARY A. Country Context and Public Financial Management (PFM) 1. Bangladesh has recorded many impressive achievements in economic and social development in the last decade and a half. It has achieved steady economic growth of over 5.5 percent annually (over 6 percent on average since 2004), maintained relatively low inflation and stable levels of domestic debt. More remarkable have been its achievements in reducing income poverty and improving social indicators, such as life expectancy, fertility control, child mortality, literacy and enrollment rates, which, in many cases, exceed other countries in the region and outside. 2. Sustained public spending in key priority areas has contributed to the impressive social outcomes observed over the years. The country has outperformed most low-income countries on a range of social indicators. Sustained expenditures on education and health in partnership with the NGO community have helped to increase gross primary school enrollment from 72 percent in 1980 to 92 percent in 2005. Bangladesh has already attained the Millennium Development Goal of eliminating gender disparity in primary and secondary school enrollment. With the sharp decline in infant mortality from 96 per 1,000 live births in 1991 to 52 in 2003, and in child mortality from 151 per 1,000 to 65, Bangladesh is set to achieve the targeted two-thirds reduction in these indicators from 1990 levels by 2015. Sustained investments in roads and rural infrastructure have established one of the most extensive road networks in a developing country with significant social and economic benefits. Food security has improved markedly, even for the very poor. At the same time, there has been a steady fall in income poverty with the share of poor people declining from 59 percent in 1990 to 40 percent in 2005. 3. Notwithstanding its economic and social achievements, the country’s performance has not matched the income surges of India and the East Asian economies. Although per capita growth accelerated over the past three decades, the incidence of poverty remains high. To achieve the ambitious poverty reduction objectives and related MDG targets, growth will need to accelerate to, and be sustained at, around 7 percent annually. This would require: (i) maintaining macroeconomic stability; and (ii) re-orienting the role of the state to facilitate private sector-led growth, as well as additional investments to the tune of 4-5 percent of GDP, especially in power and transport infrastructure. More importantly, a more equitable growth agenda means improving the ability of poor people to participate more fully in the growth process, through access to better-quality education, health, water, nutrition and local governance. The recent global financial crisis makes these objectives even more challenging. 4. Public Financial Management (PFM) policies and institutions are key to the successful achievement of these objectives. Effective fiscal policies and fiscal management institutions can ensure mobilization of necessary tax revenues without major distortionary effects on private activity. Prudent fiscal policies can ensure that fiscal deficits and its financing are consistent with fiscal sustainability objectives and good macroeconomic management. Furthermore, effective allocation of resources between and within sectors can ensure financing and implementation of priority programs and activities in the most efficient manner so that scarce resources are put to good use. 5. Bangladesh’s PFM policies and institutions have witnessed important changes over the last two decades. Substantial progress has been made through implementation of a series of first- generation reforms. The government has controlled the budget deficit and largely allocated resources in accordance with the country’s strategic priorities. It is able to collect better information on the use of budget funds and do it faster due to improvements in budget classification and computerization of some accounting transactions and reporting processes. The framework of accountability for PFM has improved with initial steps towards the separation of accounting and auditing functions, although this 1 remains an unfinished agenda. More recently, the introduction of a more strategic approach to budget planning, through the Medium-Term Budget Framework (MTBF), has marked a new era in budget preparation, bringing the recurrent and investment budget closer and introducing a strategic phase to budgeting (Appendix Five on PEFA indicators). 6. The upcoming agenda, however, is more challenging as Bangladesh aspires to modernize its economy and improve outcomes on a range of poverty and social indicators. A modernizing economy and improvements in the quality of public services would require deeper and more fundamental institutional changes in PFM practices with a significant upgrade of skills in both central and line agencies. B. Objectives of the Public Expenditure and Institutional Review 7. Drawing upon this agenda, the main objectives of the Public Expenditure and Institutional Review (PEIR) study are to: • influence the ongoing policy dialogue about public financial management policies and institutions and develop an understanding on the importance of the budget as a key instrument through which the government formulates and implements public policy. While a lot of progress has been made in the 20 MTBF ministries, budgeting is still viewed more as an accounting rather than a policy-making exercise; • build a more systematic dialogue with the government and facilitate the integration of strategic policy planning initiatives, in particular the Poverty Reduction Strategy (PRS) into the budget policy formulation by enabling central and line ministries to use the PEIR analysis as appropriate in the preparation of their MTBF over time; and • support the ongoing implementation of the next stage of the PFM reform program (2009- 2014) that will be supported by a joint group of development partners through a Multi- Donor Trust Fund. 8. The PEIR work provides an integrated analysis of policies and institutions.1 In addition to standard expenditure policy analysis typically associated with Public Expenditure Reviews (PERs), this study combines an assessment of key institutional elements that are likely to affect budget outcomes with political economy analysis of the budget institutions, in order to discover the underlying reasons behind the existing equilibrium. While the PEIR does not include a description of the technical aspects of PFM institutions, it makes use of the numerous studies that exist in this area2 to illustrate their effect on incentives. This approach is motivated by Bangladesh’s long journey of 14 years towards reforming PFM systems, which is yet to result in a fundamental break with past practices of input-oriented incremental budgeting and relatively weak expenditure controls, especially at the sector-specific level. This requires different thinking and an approach aimed at addressing the incentive structure of those who can influence more fundamental reforms in PFM. 1 Institutions, in the context of this study are defined as processes, procedures and organizational arrangements that affect the incentive structure of key agents that decide on the size, allocation and use of public resources. 2 Several studies and reports from the Bank and other donors (DFID and the Netherlands), have provided a comprehensive description/assessment of the technical aspects of PFM institutions. These include the PEIR 2003, the CFAA (2003), the CPAR (2003), the PEFA assessment 2005 and the update of the PEFA assessment from DFID (2007). 2 9. The report looks at three levels of outcomes which theory and experience consider as necessary for a PFM system to achieve full efficiency, namely: • intertemporal efficiency (also called aggregate efficiency or fiscal discipline): Effective PFM policies and institutions should take account of the trade-off of expenditure over time and minimize the differences in marginal value of spending or the marginal cost of taxation over time. For this purpose, policy makers should have the incentive to take a long-term view of public spending and coordinate their decisions; otherwise the "common pool" nature of the public funds will motivate them to opt for overspending. • allocative efficiency (across projects and programs): Given a level of spending at a point in time, the PFM machinery should be able to prioritize projects based on the marginal values and allocate the public funds accordingly. This requires an effective flow of information within the system and its interface with society. Allocative efficiency also calls for effective mechanisms to aggregate preferences and assess projects and to use the results as the basis for the allocation of funds. • operational efficiency (administrative or technical efficiency): Once public funds are allocated across a set of projects, implementation can be efficient if the administrators have adequate means and the correct incentives. This type of efficiency depends on the quality of public management at the agency level, rules for recruiting, training, performance evaluation, and promotion in the bureaucracy as well as the salaries paid and the resources available to the administrators. 10. The following analysis shows that Bangladesh does fairly well on intertemporal and allocative efficiency. This is evidenced by its strong record in maintaining fiscal discipline and sustained spending on broad sectoral priorities over the years. However, it does very poorly on operational efficiency at the more micro, sector/spending agency level. Following the budget policy outcomes and issues (Part I), the report focuses on an assessment of the institutional incentives that dominate the fiscal machinery in Bangladesh. To find the underlying reasons that preserve the incentive structure, Part II examines the political economy constraints that influence budget management institutions and outcomes in an effort to define the medium-term horizon and priorities for influencing these incentives on the way forward (Part III). C. Public Financial Management Outcomes 11. Bangladesh’s efforts to improve PFM, underway for several years, have been commendable. Although tremendous progress has been made on several fronts, PFM outcomes present a mixed picture. Fiscal discipline has been maintained despite pressures placed on public resources by various external shocks, recurring natural disasters, widespread poverty and social needs. Management of public resources within sectors and programs, however, needs to be improved significantly to achieve the level of efficiency and service delivery outcomes that the country aspires for. 12. Bangladesh’s share of public expenditures, which is often used as a measure of the size of the public sector, remains among the lowest in the world and reflects a very low revenue-to-GDP ratio of about 10 percent. Its public sector, however, extends far beyond the central government, comprising a large state-owned enterprise (SoE) sector and a dominant public financial sector. SoEs that operate in traditional utilities, infrastructure and the manufacturing sector account for more than 20 percent of public sector employment. In addition to being a drain on public resources through a range of subsidies 3 and contingent liabilities, they also exert considerable influence on the economy through the supply of vital inputs and services and their pricing policies. 13. In the context of constrained revenue capacity, Bangladesh has managed to maintain fiscal discipline and allocate spending to the country’s strategic priorities. Modest improvements in revenue performance and declining overall public expenditures have stabilized the central government’s fiscal deficit at 3.7 percent of GDP in FY08. After stagnating at around 10 percent of GDP over several years, revenue collection improved markedly in FY08 due to the caretaker government’s drive to reduce tax evasion. Positive growth performance and improving fiscal balances have reversed the increase in government debt which, as a share of national income, fell to 47 percent of GDP at the end of 2007. Spending in the social sector has been sustained, financing for rural development, especially in rural infrastructure, has increased and support for targeted poverty reduction programs, including social safety nets for the poor, has continued (Figures 1 and 2). Figure 1: Social Sectors Remain in the Lead Figure 2: Expenditure Trends (% of Total Expenditure) Development Expenditures Have Been Declining 16.0 4.0 14.8 Total Expenditure 14.0 13.8 13.9 13.7 13.6 3.5 13.2 12.0 3.0 10.0 2.5 Recurrent Expenditure Percent GDP 9.1 2.0 8.0 8.4 8.4 8.0 8.1 7.7 1.5 6.0 1.0 5.6 Annual Development Program 5.4 5.0 5.0 4.7 0.5 4.0 3.9 0.0 2.0 General Social Services Economic Inf r ast ruct ur e Int er est Administ rat ion Ser vices Ser vices 0.0 2002 2003 2004 2005 2006 2007 FY98-FY02 Average FY03-FY07 Average Source: MoF and WB staff estimates. 14. The quality of fiscal performance and expenditure management, however, has been weak, with particular deficiencies in the tax system, SoEs and structural rigidities. Despite recent progress, mobilization of public revenues remains one of the lowest in the world. Wrapped in old policies and outdated administrative practices, the tax system has not been able to bring new income-generating activities within the tax net. Revenue generation relies heavily on trade taxes, with nearly 45 percent of tax revenues coming from import-based taxes. Weaknesses in tax administration continue to result in leakages, evasion and inefficient revenue mobilization. The recent surge in global oil and food prices, coupled with the failure to address the issue of loss-making SoEs after 2003, resulted in a significant increase in transfers in the form of subsidies (15.8 percent of total expenditure in FY08) to cover SoE losses, especially in the petroleum, electricity and fertilizer industries. In this environment, fiscal adjustment has been driven by a decline in development spending (investments), which reached 3.4 percent of GDP in FY08 after peaking at 6.6 percent in FY01. Significant structural rigidities and implementation problems continue to hamper the implementation of the investment budget, which has been constantly underspent by a large margin (around 20-25 percent). Interest expenditures, which at 2 percent of GDP are not large by international standards, have been rising and are now almost equal to the education budget or double the amount of health budget. 4 15. These vulnerabilities are threatening Bangladesh’s prospects for maintaining its existing positive macro-fiscal outcomes, as well as prospects for maintaining fiscal sustainability. Given the significant size of the public sector beyond the central government’s civil service agencies, the sustainability of Bangladesh’s fiscal stance depends not only on the fiscal position of the central government, but also on the performance of the broader public sector, especially in light of high fiscal and economic costs associated with the situation of SoEs. While the recent fiscal stance of the central government is sustainable at the primary deficit levels of 1.5 percent of GDP, the fiscal stance of the Consolidated Public Sector (CPuS), (which includes transfers to SoEs), appears to be questionable. 16. While Bangladesh’s current fiscal stance appears to be sustainable, there is little room for complacency because of several vulnerabilities that are looming on Bangladesh’s medium-term economic horizon that could adversely affect its fiscal sustainability. These include: (i) lower economic growth; (ii) inadequate adjustment of the administered prices of diesel, power, natural gas, and fertilizer; (iii) public sector pensions and wage increases; (iv) higher interest rates and a large depreciation of the real exchange rate; and (v) insufficient reforms to improve revenue mobilization. If Bangladesh were to suffer a sustained rise in interest rates (higher than 9 percent) or a substantial decline in growth (lower than 6.5 percent), it would need to run a substantially smaller primary deficit or even a primary surplus to meet its debt target. Further, the CPuS sustainable level declines to 2.1 percent of GDP, if interest subsidies are removed, well below the 5.2 percent observed in FY07. D. Important Challenges Remain at the Operational Level 17. Bangladesh’s very good record of maintaining fiscal discipline and sustaining the allocation of public resources to broad strategic priorities is not matched at the operational, sector and agency levels, where its record is relatively poor.3 While Bangladesh is becoming well-known for having achieved good economic and social outcomes over the last two decades, the review of expenditure policies and programs in five key sectors--education, health, transport, energy and agriculture-- highlights a number of inefficiencies and structural issues, which need to be addressed urgently. Failure to do so, in many cases, could jeopardize the sustainability of the already achieved social and development outcomes. 18. A number of common features across all sectors demonstrate the need for significant improvement in the quality of public expenditure at the sector level. While sectoral policy frameworks are in place and appear to be appropriately formulated, intra-sectoral expenditure allocations and usage do not always reflect the sector’s policy, revealing a passive expenditure policy stance. Public expenditure policy in transport continued to favor the extension of the road network (FY03-FY07) despite the sector’s requirement for a more balanced and integrated multimodal transport system. Major transfers in the form of subsidies to energy and agriculture are crowding out much needed investments in operations and maintenance and core public goods such as research and technology. While overall enrollment rates in education have been sustained and even slightly increased between 2000 and 2005, the composition of enrollment has shifted away from government-recognized schools (with an 8 percent decline) where most public funding is committed. 19. In almost all cases, expenditure distribution needs to address the issue of equity as they remain skewed towards the non-poor. In education, the poor represent 40 percent of the total 3 Operational efficiency has two dimensions: (i) minimizing cost per unit of output or achieving more value for the money spent; and (ii) achieving the outcome for which the output is intended. This chapter deals more extensively with (ii), although makes references to (i) as appropriate. A full coverage of (i) is beyond the scope of this document. 5 population of school-aged children but receive only 32 percent of the total recurrent education expenditure. Even the primary stipend program, a conditional cash transfer program designed to target the poorest 40 percent of students, is only marginally pro-poor. In the health sector, the share of recurrent spending that goes to the 16 poorest districts is less than what is going to the 16 richest districts. Since energy consumption is progressive, with higher income groups consuming more, general public subsidies going to the sector are, in many ways, largely inequitable and benefit the more affluent disproportionately. 20. Poor planning and implementation problems continue to affect the quality of the investment portfolio in the sectors’ Annual Development Program (ADP), which has been a long-term concern of policy-makers. Time and cost overruns, deferral of project benefits long into the future and declining returns on investment have historically been associated with poor management of the development budget in Bangladesh. A review of the investment projects/programs in the ADP4 for FY03-07 reveals that on average, only about one-third (34 percent) of their cost was covered by the development budget. Consequently, the ADP each year has carried over a significant backlog of incomplete projects/programs. New projects/programs entering the portfolio grew by 2.7 percent per year, while the development budget grew by 3.5 percent. The continuing suboptimal allocation of resources suggests that this trend of implementation delays, despite small improvements, will persist and lower the overall effectiveness of the ADP. Project completion rates are going to be very low for the foreseeable future. 21. Addressing these challenges is not a straightforward matter of technical fixes as demonstrated by the country’s decade-long efforts to reform public financial management. While many improvements have taken place, the record shows that several fundamental weaknesses persist in the functioning of PFM practices. A new approach is needed that goes beyond treating PFM systems as isolated (self-standing) technical reforms, which will tackle the institutional incentives and political economy constraints to create an enabling environment for improving the outcomes. E. Institutional incentives and Political Economy Constraints for Better PFM Outcomes 22. The management of public finances in Bangladesh is concentrated in the hands of two key central finance agencies (CFAs), the Ministry of Finance (MoF) and the Planning Commission (PC). These institutions have the delegated responsibility within the executive to play the leading role in planning, allocating, managing and monitoring of public finances. This institutional structure has served the country well for the purposes of maintaining aggregate fiscal discipline and allocating resources to broad strategic priorities (Chapter Four). However, by limiting the flexibility and the autonomy of the line ministries and spending agencies in the face of weak administrative and personnel management practices, PFM institutions have inhibited operational efficiency and caused poor expenditure outcomes at the sector level. 23. The delegation of powers with decision-making authority over fiscal aggregates is given to a few key members of the Budget Monitoring and Resource Committee (BMRC),5 effectively insulating decision makers from diverse interests and views at the preparation stage. The cabinet only has authority to accept or reject fiscal aggregates, which are presented to parliament at a much later stage for approval. Parliament does not make changes to the budget aggregates. The dominant role of CFAs also helps in the strategic allocation of resources. Complemented by the development of a national 4 Excluding technical assistance projects/programs. 5 The resource committee consists of the following members, chairman of NBR, the Governor of the Bangladesh Bank, Secretary of Finance, and the Planning Secretary. 6 strategy and, to some extent, the introduction of a budget strategy (through the MTBF), these agencies have a major influence in maintaining strategic priorities. Given the limited capacity of the line ministries to present different alternatives as well as their limited flexibility during the implementation to make many changes, priorities are likely to stick. Further, the limited role played by the cabinet and especially the parliament in the review and scrutiny of the budget limits their opportunities to alter the priorities, as they enter the process only in the last month before approval. Moreover, as these priorities emphasize social programs and rural development, areas that are of interest, and visible, to voters, politicians have found it difficult not to support them. 24. These effective institutional arrangements at the preparation stage suffer from some weaknesses at the implementation stage, the main ones being related to weak ex-post accountability and lack of budget comprehensiveness. Weak accountability is primarily related to the role of the Comptroller and Auditor-General (C&AG). While a lot of progress has been made to improve and modernize the external audit function, more needs to be done to make it independent of the executive and a key instrument of accountability. Lack of budget comprehensiveness is manifested in the buildup of contingent liabilities. The use of off-budget expenditures through mechanisms, such as public bank/enterprise borrowings as a means of redistribution, can satisfy many political constituencies without having a tangible effect on voters for a significant period of time. Even when they are later recognized as a government liability, they would simply be added to the government's debt and their amortization would be spread over time and passed on to future generations. 25. While the dominance of central agencies in maintaining fiscal discipline and prioritizing and allocating resources among the sectors appears to be the strength of the current system, the limited participation in decision-making and flexibility of line ministries in the design and implementation of the budget within their agencies is a significant weakness. Specifically, until the introduction of the MTBF approach, the procedures did not give the ministry executives much leeway in designing and allocating their budgets to achieve their organizations’ objectives. These weaknesses were reflected in poor operational efficiency at the sector/agency level in particular (Chapter Three). Moreover, the effectiveness of whatever agency autonomy and merit-based recruitment/promotion exist in the Bangladesh bureaucracy is further undermined by a lack of the economic/institutional characteristics needed for the proper operation of those mechanisms. Perhaps the two most striking characteristics in this respect are low salaries of civil servants and the limited resources available to line agencies for their day-to-day operations. Bangladesh’s performance in this respect is much weaker than other developing countries. 26. The above findings demonstrate that operational efficiency is very closely linked to the proper functioning of the public administration, whose reform has long been under consideration, but needs to be prioritized for more effective management of public finances. There are many forces that militate against the tendency to reform. First, civil service reform is a risky undertaking with a time horizon that goes beyond the tenure of a single government. Second, much of the change has to be implemented by units and people who most likely will be affected by the change. Third, given the political fragmentation in Bangladesh, there has not been much incentive on the part of politicians to encourage bureaucratic autonomy and break the system of patronage. 27. The tendency to maintain the status quo in the civil service, with the perseverance of its ineffective cadre system in particular, has created a fragmented and inefficient PFM system at the level of the line ministries. The PFM administration at the line ministry level is a collection of deconcentrated units of the CFAs, MoF and PC. In addition to being inefficient, these vertical reporting/loyalty lines outside the line ministry structures divert responsibility and undermine the system of checks and balances that can be established to minimize waste and corruption in public funds. 7 These institutional arrangements, with dominant CFAs and limited flexibility for line ministries in budget management, might have been warranted in the face of macro-fiscal instability and economic and political challenges facing the country after independence. In today’s realities, however, they need to be reconsidered and gradually restructured to better serve the goal of effective management of public finances. F. Looking Ahead: The Medium-Term Horizon 28. If PFM reforms are to be successful in achieving the desired outcomes at all three levels, particularly at the operational level, they need to motivate the holders of public office and politicians to strengthen their development vision and programs and take a greater stake in implementing them effectively. If political capital can be built on successes with initiatives that improve the delivery of services financed by public resources, especially in strategic sectors, some leaders may start to place a greater emphasis on achieving policy outcomes and improving service delivery as a route to success. 29. Any future strategy to improve budget outcomes needs to focus on two main directions: (i) introduction of a performance focus in the budget process; and (ii) broadening the constituency for change both within and outside the executive through participation and transparency. Once the budget activities and programs become publicly observable with clear input and output indicators, the incentive structure tends to change direction. Political competition is likely to reduce its rent distribution aspects and steer it to a focus on delivering better results on public services. 30. The reforms initiated by the government, especially the introduction of the MTBF approach (introduced in 2004) and computerization of the budget accounting systems are important steps in this direction. Although these initiatives have got off to a very good start, they remain focused on serving the more narrow objectives of the financial management apparatus in controlling inputs, rather than the broader management objectives of monitoring sector and program performance. There is broad recognition that this focus needs to change, but that significant barriers exist for this to happen on several fronts. Given the challenges on fiscal policy and budget institutions, the medium-term agenda is ambitious but feasible, if macroeconomic stability is sustained and progress with policy and institutional reforms continues. 31. Strengthening domestic revenue mobilization and enhancing operational efficiency to make better use of available public resources should be at the core of the medium-term strategy to improve budget policy and institutional performance in the context of the MTBF approach. This is particularly important in light of a constrained growth environment, the anticipated reduction in trade taxes and growing development expenditure needs. Raising domestic revenue levels and improving the efficiency of available resources are a high priority, required to create fiscal space and improve public expenditure outcomes. Improving Revenue Performance: the Cornerstone of the Medium-Term Fiscal Strategy 32. Stagnating at around 10 percent of GDP for almost a decade, total revenues in Bangladesh have been largely unresponsive to changes in economic activity.6 This is remarkable for an economy that has been growing rapidly over the last decade and a half. It appears that the existing tax system has been unable to bring new income-generating activities within the tax net. Revenue collection increased significantly in FY08, due to several measures taken by the caretaker government to reduce tax evasion; 6 This section draws on the assessment that was done by the IMF in January 2007. 8 however, many of the fundamental weaknesses that have been part of the system for a long time remain unaddressed. 33. Introducing a simple, coherent and effective tax policy should be a priority, especially for domestic taxes such as VAT and income tax, which are likely to be the future mainstay of Bangladesh’s tax system, as the need to reduce dependence on international trade tax approaches. Levels of tax collection are low even though the statutory nominal tax rates are high. A low revenue-to-GDP ratio can be attributed to: (i) heavy reliance on trade taxes and low level of domestic taxes; (ii) significant tax expenditure, which erodes the tax base; and (iii) major organizational and institutional weaknesses in tax administration. Despite the adoption of numerous tax policy measures and tinkering on the margin with improvements in administrative practices over 10-15 years, Bangladesh has been broadly unsuccessful in lifting the revenue-to-GDP ratio to a level warranted by the development objectives. The current situation calls for a comprehensive reform agenda. 34. To improve its revenue performance and efficient functioning, Bangladesh’s VAT system should be urgently reformed with the principle objective of extending the VAT to the retail stage and applying it universally with an exemption that applies only to small businesses. This would require a thorough review of the VAT legislation, and the redrafting of the VAT Act, the tightening of the exemptions that are currently provided for by the regulations, a different application of the turnover tax for small businesses, and a review of the excise duties on the basis of efficiency and revenue mobilization objectives. Bangladesh’s VAT operates like an extended excise system of taxation and, after 15 years of application, it still resembles the system it sought to replace. The VAT method is applied mainly at the production/wholesale stage, at a rate of 15 percent. This partial application is only applicable to transactions involving manufacturers, importers and perhaps large raw material producers. 35. Despite being based on a sound original legislative framework dating from 1984, the income tax system in Bangladesh, consisting of corporate and personal income tax, operates inefficiently. The original legislation, which is still in force today, incorporates a number of modern income tax concepts. However, many amendments over the last two decades have significantly altered the once coherent structure. Redrafting the law to make it understandable, simple and coherent is a high priority. The increased use of the withholding method to collect taxes threatens to turn a major part of the income tax regime into simply a series of excise taxes on readily identifiable financial payments. In most cases, the application of the withholding tax does not have any relationship to the taxpayer’s income tax liability, which greatly distorts the horizontal equity concept of income taxation that individuals with the same income should pay the same amount of tax. A serious omission from the current legislative framework that is likely to lead to a great deal of revenue loss is the lack of detailed tax accounting rules, which describe how businesses must account for their profits for tax purposes. 36. Four key pillars should be part of any NBR effort to simplify processes, improve governance and modernize revenue administration: (i) organizational improvements; (ii) automation of tax collection; (iii) flexible human resource management practices, and: (iv) development of a culture of taxpayer services. Bangladesh’s revenue administration performance is way below its potential and is perceived by the business community as one of the most important constraints to investment and growth. To improve productivity and taxpayer compliance, the NBR firstly, needs to move towards a functional design within its tax-by-tax structure, sharing some central common functions across taxes. Secondly, with the exception of customs administration, there is minimal automation of tax collection or recording in the VAT and income tax wing of NBR. ICT can play an important role in modernizing revenue administration. However, this needs to be accompanied by simplified business processes and the functional organization of the agency. Thirdly, there are major HR issues with cadre rivalry among various tax departments, lack of integrity, low pay and other inefficiencies that result from a centralized 9 and inflexible human resource management system. HR practices need to be reformed to allow the agency to build a more flexible policy to motivate staff, improve productivity and strengthen professionalism and integrity. Finally, there is a need to reform taxpayer services by building a year- round channel of consultations with taxpayers and other stakeholders to keep them updated on significant changes in policies and procedures; creating a web-based taxpayer information and communication service; and enabling electronic submission of returns, introducing facilities for electronic payments, and direct trade inputs for importers. Introducing a Performance Focus in Budgeting: The Medium-Term Budget Framework (MTBF) 37. Recognizing the limitations of an annual, input-oriented and fragmented process, the government introduced a three-year MTBF with the FY05-06 budget in four pilot line ministries.7 The MTBF was subsequently rolled out to 20 key line ministries (in FY09-10) covering over 62 percent of total budget expenditure. The introduction of the MTBF has marked a new era in Bangladesh’s PFM reforms. However, it is still too early for the MTBF to have had a measurable impact on performance and budget outcomes. Nevertheless, significant progress has been made in developing an MTBF adapted to the particular circumstances of Bangladesh. In a difficult political and institutional environment, the MTBF is now seen by both central and line ministries as an integral element of the budget planning cycle. The initiative has positively influenced the incentive structure of key players, especially in line ministries. It has been supported by their senior managements, primarily because it is seen as having the potential to provide greater autonomy and empowerment to line ministries in the planning and management of their budgets. The challenge moving forward will be to build on these early successes to achieve a sustainable MTBF process that can contribute effectively to the realization of improved budget outcomes. The available “institutional space” within which to introduce further improvements in the budget management process, however, has now been filled. Making progress on the MTBF requires greater emphasis on issues of budget integration, institutional mandates, organizational reform and institutional capacity constraints. Impact on Incentives and Next Steps for Building a Sustainable MTBF 38. The introduction of the MTBF as a more strategic and performance-oriented approach to budgeting is having a positive impact in a number of areas, including improvements on the demand side for better budgeting within the executive. This has been a consequence of: (i) engaging higher- level officials in the budgeting exercise; (ii) motivating LMs to prepare better budget submissions; and (iii) highlighting the need for better budget execution and procurement systems. In order to keep them engaged, it will be important to ensure that MTBF reforms are perceived by stakeholders as bringing about improvements in budget planning and management, particularly in strengthening the management role of line ministries and gradually giving them more flexibility. The Finance Division (FD) needs to be responsive to the feedback from line ministries, particularly in providing predictable resources and streamlining budget execution procedures, as well as reducing the present duplication in budget planning and management procedures. Ultimately the sustainability of MTBF reforms will depend on the continued buy-in of key participants and stakeholders. This applies to the central ministries and agencies (principally FD and PC), the line ministries and at the political level. 39. The MTBF has brought a strategic phase to budget development (October-December), which focuses on the implications of strategic policy priorities for the allocation of resources under both the 7 The first pilot covered the Ministry of Agriculture, Ministry of Education, Ministry of Social Welfare and Ministry of Women Affairs. 10 Non-development Budget (NDB) and Development Budget (DB). This stage has created a platform for participation of high-level officials of line ministries in budget discussions through the Budget Management Committees (BMCs).8 While these committees have been able to attract high-level officials in the budget discussions, their effectiveness has been hampered by the lack of technical capacity and secretariat support. In the medium term, however, building a sustainable and inclusive MTBF process will require greater attention to the institutionalization of such structures and a stronger engagement at the political level. To strengthen political engagement in the budget process, the following initial steps should be considered: (i) approval of the MTBF resource ceilings in a high-level decision-making forum such as the Cabinet of Ministers, accompanied by a Medium-Term Budget Outlook (MTBO) paper setting out the preliminary macro-fiscal framework and identifying the priorities to be addressed; and (ii) political level consultation at the outset of the budget process on the MTBF in the form of a MTBF concept paper, as the process matures. 40. Strengthening the management role and flexibility of the line ministries in budgeting remains a priority to improve budget performance at the sector level (operational efficiency). The MTBF approach has been well received by the line ministries and is seen as having brought about a more strategic and forward-looking perspective. It is encouraging ministries to feel more empowered and plan their programs for the medium term, thus giving them a greater voice and decision-making power on how resources are allocated. At the same time, it has placed new demands on them with regard to budget planning and management. If the processes are to be sustainable, it will be important to ensure that the necessary institutional framework and incentives are in place to sustain the engagement of the line ministries in the MTBF process, including a stronger and more integrated role for the Ministry Budget Frameworks (MBFs) in the budget planning process and decision-making. 41. There is a need to reform budget execution and procurement practices as data limitations, the absence of a program classification, delays in the budget release process, and inconsistencies between the procurement cycle and the budget cycle, hamper the proper preparation and implementation of the MTBF. Addressing these complementary reforms to budget execution processes, in particular the fund release process, classification structure, computerization of the Treasury and accounting transactions, and integration of procurement and budgeting processes will be important. In the longer term, FD should consider the scope for delegating additional financial powers to the line ministries that meet benchmark PFM standards, with the aim of moving from ex-ante controls towards accountability backed up by strengthened ex-post inspection and audit. Addressing the Budget Integration Agenda: The Cornerstone of Improving Expenditure Performance 42. Measures to strengthen the autonomy of the line ministries should progress simultaneously with measures to better integrate the two sides of the budget, the DB and NDB. The FY06 MTBF document stated that a key objective of the GoB in introducing the MTBF was ‘removing the demarcation between the development and non-development budgets gradually by joint programming of development and non-development spending’. This segregation is at the heart of several deficiencies in the expenditure management process highlighted in Chapter Two of the report. It will become increasingly difficult to make progress on the MTBF unless the budget integration agenda is addressed. The immediate requirement will be to resolve the outstanding issues surrounding the way the budget is prepared, presented and implemented. 8 The Budget Monitoring and Review Committee (BMRC), chaired by the Minister for Finance, oversees the MTBF process. The Medium-Term Budget Coordination Technical Committee, chaired by the Secretary of Finance, exercises overall coordination over MTBF preparation. Two technical working groups-the MTMF Working Group and the Resource Ceilings and Estimates Evaluation Working Group-provide support to the FD with the preparation of the MTMF analysis and the development of the budget strategy and spending proposals. 11 43. Perhaps the most difficult and the most important reforms would be those that need to streamline the responsibilities among different institutions such as PC, MoF and the line ministries. The move towards a unified budget under the MTBF requires a significant change in the participation of the PC in the budget process. This should emphasize a higher-level role that reflects the primary responsibility of the PC in overseeing the management of the GoB’s policy and strategy agenda. At the same time, the PC should become less engaged in the details of budget planning and monitoring. The FD should take responsibility for overseeing the entire budget process to enable GoB to move progressively towards a single integrated budget. The distinction between capital and recurrent spending should replace the current one between development and non-development expenditure. Achieving this will require the FD to integrate economic and planning skills into its budget planning and management. The MTBF also requires a more proactive role for ministry secretariats in the planning and budgeting process, with a stronger emphasis on sector policy and strategy development and monitoring, as well as a better integration of the currently segregated PFM functions. Strengthening Budget Transparency: Involving the Parliament and Civil Society 44. Using MTBF as a vehicle for stronger consultation with parliament, civil society and the broader public on key budget issues can significantly increase the transparency of the budget process. To date, the focus in Bangladesh has been on building understanding of the MTBF reforms within government. Engagement of external stakeholders on the MTBF has, therefore, been limited and has not been directly sought by the FD. For example, there has been little coverage of the MTBF document in the press. Discussion between the FD and business leaders and civil society has been limited to the final stages of budget preparation and has, therefore, not significantly influenced planning of the budget. 45. There are, in particular, opportunities to refocus parliamentary discussion of the budget towards issues of strategy and performance and make a presentation of the MTBF to the parliament at the conclusion of the strategic phase of MTBF/budget development. In addition, the MTBF can be used as a basis for enhancing dialogue with civil society on budgetary issues and more generally for increasing the accountability of the budget process. Finally, focusing on the new reforms for creating an adequate system of local governance and subnational finance would gradually contribute to the demand for better budget outcomes and service delivery. Perhaps one of the most effective avenues to increase participation and consultations on the budget process with potentially significant impacts on performance and outcomes is through greater fiscal decentralization. F. Increasing the Demand for Better Budgeting Through Decentralization 46. In terms of bureaucratic and fiscal structure, Bangladesh is highly centralized with a complex system of subnational administration with various types of overlapping deconcentrated and devolved local administrative entities. It lags behind other countries of its size and per capita income on both aggregate expenditure and revenue measures of fiscal decentralization. Subnational expenditures as a percentage of total consolidated government expenditures are estimated to be in the range of 3-4 percent. On the revenue side, most major tax bases remain under the control of the central government and while own-source revenue potential varies vertically across the different levels as well as horizontally across different entities at each level, less than 2 percent of total government revenue is collected at subnational levels, again placing Bangladesh at the lowest end internationally. 47. Local revenues are collected from property taxes, various user fees and lease income on community properties; however, tax rates are uniformly set at the central government level. Besides the restricted revenue base, revenue collection is undermined by weak municipal and local government 12 tax administration. Consequently, transfers from the central government are the main source of financing for many local governments, especially the Union Parishads (UPs) and Pourashavas (municipalities) with low own-source revenue potential. Service delivery responsibilities are primarily performed by the deconcentrated offices at the Zila (district) and Upazila levels in particular. Offices of line agencies at the upazila level manage service providers, including schools and hospitals, and are used as conduits for several grant programs, including some grants provided by the Local Government Division. Nevertheless, line ministry staff rather than local government officials, deliver these services. 48. Deconcentrated government spending across Bangladesh’s 64 districts is fairly equalized, but the allocation of development spending to poor districts is lower. Coefficients of variations are below 1 for both per capita development and recurrent expenditures at the district level, which is low in comparison to other developing countries with more devolved fiscal structures. Of some concern is that per capita allocations under the development budget are inversely related to poverty levels in districts (if the three Chittagong hill districts are excluded). While aggregate spending levels respond to a variety of spending needs and capacity that might not necessarily correspond with poverty levels, spending needs on pro-poor social infrastructure, such as health and education as well as spending under social welfare programs, are presumably higher in districts with a higher incidence of poverty. The largely discretionary ADP allocation system--at least at an aggregate level--does not seem to accommodate pro- poor spending considerations. 49. Reforms of local government institutions and intergovernmental fiscal relations can be a critical element supporting more effective and efficient public spending in Bangladesh. The government has successfully maintained fiscal prudence and expanded service access in a number of sectors, including education and health and more recently rural roads. The challenge now is to improve quality and efficiency in service provision, which is more difficult to achieve in the overly centralized framework. Greater devolution of government would offer significant advantages to Bangladesh by enhancing the quality of local public service delivery, encouraging local resource mobilization, allowing for greater local voice in public service decisions, and generally increasing satisfaction with government. 50. Significant restructuring of government would be necessary to create a more devolved system that operates with incentives for efficient behavior at all levels. Greater local control would be necessary on expenditure choices and staffing in conjunction with greater ability to mobilize own-source revenues. The intergovernmental revenue structure would need to be re-examined and restructured in light of different expectations from local government, with the likelihood that more transfers are necessary. Furthermore, major reforms would be necessary in a number of national government institutions, such as the Department of Public Health Engineering (DPHE) and the Local Government Engineering Department (LGED). These organizations would need to serve as facilitators and monitors rather than being highly integrated into local government functions, as they are today. While this reform agenda has gained some momentum, actual change has lagged behind, and it would be important to scale up implementation based on experience with pilot programs in some local governments, for example the Local Governance Support Program at the UP level. 13 PART I PUBLIC FINANCIAL MANAGEMENT OUTCOMES 51. Effective delivery of three levels of outcomes is considered necessary by theory and experience for a PFM system to achieve full efficiency, namely: • intertemporal efficiency (also called aggregate efficiency or fiscal discipline): Effective PFM policies and institutions should take account of the trade-off of expenditure over time and minimize the differences in marginal value of spending or the marginal cost of taxation over time. For this purpose, policymakers should have the incentive to take a long-term view of public spending and coordinate their decisions accordingly; otherwise the "common pool" nature of the public funds will motivate them to opt for overspending. This will cause inefficiency because it eventually leads to constraints on borrowing and shortages of funds. • allocative efficiency (across projects and programs): Given a level of spending at a point in time, the PFM machinery should be able to prioritize projects based on the marginal values and allocate the public funds accordingly. This requires an effective flow of information within the system and its interface with society. Allocative efficiency also calls for effective mechanisms to aggregate preferences and assess projects, and to use the results as the basis for the allocation of funds. • operational efficiency (administrative or technical efficiency): Once public funds are allocated across a set of projects, implementation can be efficient if the administrators have adequate means and the correct incentives. This type of efficiency depends on the quality of public management at the agency level, rules for recruiting, training, performance evaluation, and promotion in the bureaucracy as well as the salaries paid and the resources available to the administrators. 52. The following analysis shows that Bangladesh does fairly well on intertemporal and allocative efficiency. This is evidenced by its strong record in maintaining fiscal discipline and sustained spending on broad sectoral priorities over the years, but does very poorly on operational efficiency at the more micro, sector/spending agency level. 14 CHAPTER ONE MAINTAINING AGGREGATE FISCAL DISCIPLINE Bangladesh has a commendable record in maintaining fiscal discipline despite poor initial conditions and constant pressures placed on public resources by recurring natural disasters and widespread poverty reduction and social needs. The country has managed to stabilize its fiscal deficit, modestly improve revenue mobilization, lower spending, reduce reliance on the central bank to finance public spending, stabilize the debt situation, and undertake significant reforms to improve its public financial management institutions. Despite all the positive gains, the quality of the fiscal performance remains far from satisfactory. Weaknesses in tax administration continue to result in tax leakages, evasion and inefficient revenue mobilization. The recent surge in global oil and food prices coupled with the failure to continue closing SoEs after 2003 resulted in a significant increase in transfers in the form of subsidies. In this environment, fiscal adjustment has been driven by a decline in investments despite major infrastructure needs. While the fiscal stance appears to be sustainable, several vulnerabilities are looming large in the medium-term economic horizon. Key among them are: (i) lower economic growth; (ii) inadequate management of contingent liabilities; and (iii) insufficient reforms to improve revenue mobilization. In light of growing development expenditure needs and a vulnerable global economic environment, strengthening domestic revenue mobilization through better tax policy and administration should be at the core of the medium-term budget strategy. Key Policy Priorities: • Revise tax legislation to introduce simple, coherent and effective tax policy provisions for domestic taxes such as VAT and income tax before modernizing revenue administration. • Strengthen revenue administration by: (i) moving towards a functional organizational structure; (ii) introducing automation and more flexible HR rule; and (iii) developing a culture of taxpayer services. • Establish a system to monitor and manage contingent liabilities, in particular subsidies rising as a result of changes in the global commodity and food prices. 1.1 Macro-fiscal Outcomes 53. Bangladesh has a small government budget if analyzed on the basis of its share of GDP. A government’s budget is often used as a measure of the size of the public sector. At 14-15 percent of GDP (FY03-FY07), Bangladesh’s share of public expenditures remains among the lowest in the world and is consistent with a very low revenue-to-GDP ratio of about 10 percent. Bangladesh’s public sector, however, extends far beyond the central government, comprising a large SoE sector and a dominant public financial sector. SoEs that operate in traditional utilities, infrastructure and the manufacturing sector account for more than 20 percent of public sector employment. In addition to being a drain on public resources (see paragraph 33), a range of explicit and implicit subsidies and contingent liabilities means they also exert considerable influence on the economy through the supply of vital inputs and services and their pricing policies. Although total public expenditures have increased consistently and almost doubled since the 1980s, as a share of the country’s GDP they remain among the lowest relative 15 to low- and middle-income countries9 (Table 1-1). This reflects the correlation between the level of economic development and government spending power observed over time in many different contexts (Fan & Rao 2003). Table 1-1: Fiscal Trends: Bangladesh, Low-Income Countries and South Asia Region (% of GDP) Low-Income South Asia Period Countries Region Indicator 2001-02 2002-03 2003-04 2004-05 2005-06 2006-07 2005 2005 Total Revenue 10.1 10.3 10.1 10.5 10.7 10.5 15 12.5 Total Expenditure 14.8 13.7 13.2 13.8 13.9 13.6 15.6 15.2 Overall Budget Deficit 4.7 3.4 3.1 3.3 3.2 3.1 -1.7 -3.2 Primary Deficit -2.9 -1.5 -1.4 -1.7 -1.4 -1.2 — — Net Foreign Financing 2.1 2.1 1.3 1.6 1.2 1.2 1.1 1.4 Net Domestic Financing 2.5 1.2 1.8 1.7 2.1 1.9 1.1 0.3 Total Central Government Debt 52.7 51.1 51.0 50.1 49.3 47.4 — 57.2 Source: Ministry of Finance, Bangladesh. Note: Non-ADP & Net Lending in FY06 includes bonds (10 billion taka) issued to a State Commercial Bank to assume BPC's liabilities. 54. Bangladesh’s record of maintaining fiscal discipline has been consistently positive: • The overall central government deficit has been declining. After peaking at 5.1 percent of GDP in 2000 and 2001, the central government deficit declined to 3.1 percent of GDP in 2007. The primary deficit also declined by over 60 percent, from 2.9 percent of GDP in 2002 to 1.2 percent in 2007. Total revenues have increased marginally but have stayed at around 10 percent of GDP10, while total expenditures have been declining from 14.9 percent of GDP in 2002 to about 13.5 percent in 2007 (Figure 1-1). Revenue performance has improved significantly in FY08 reflecting improvements in revenue collection stemming from the caretaker government’s drive against tax evasion, introduction of a universal self- assessment system and intense monitoring of activities of taxpayers and officials by NBR and other relevant government agencies. Total expenditures, however, increased in FY08 on account of increased subsidy and transfers to the state enterprise sector and unanticipated expenditures in response to floods and cyclones. Despite the recent increases, Bangladesh remains a country with one of the lowest levels of public revenue and spending among the world’s low- and middle-income countries. 9 Economic literature offers limited guidance on what is the “appropriate” size of government from a practical perspective. Certain criteria have been developed based on international comparisons of size and scope of governments. According to a typology proposed by Tanzi and Schuknecht (2000), for example, governments can be categorized in three groups based on their share of government expenditure to GDP: small (public expenditure below 40 percent of GDP), medium (public expenditure between 40 and 50 percent of GDP), and large (public expenditure above 50 percent of GDP). 10 Total revenues have been increasing in the first three quarters of FY08, reflecting improvements in tax administration and reduced tax evasion as a result of the major anti-corruption drive of the caretaker government that took power in early 2007. 16 Figure 1-1: The Fiscal Deficit Has Been Declining (% of GDP) 1 6 T o ta l E x p e n d it u r e 1 2 F is c a l D e fic it Tot al 8 R ev e nue 4 F Y 01 F Y 02 F Y0 3 F Y 04 F Y 05 F Y0 6 F Y 07 F Y 08 Source: Ministry of Finance. • A secular decline in the foreign-financed component of the deficit has changed the structure of deficit financing, with domestic financing now comprising the largest share. As Bangladesh’s growth performance remained strong, net foreign financing of the deficit declined significantly from 4.8 percent of GDP in the 1990s to around 1.2 percent in 2005- 2007 before rising to 2.2 percent in FY08. Net domestic financing increased slightly, while domestic bank financing declined from 1.6 percent of GDP in 2000 to 0.9 percent in 2007 (Figures 1-2 and 1-3). Figure 1-2: Foreign Financing Has Been Figure 1-3: Positive Growth Performance Declining (% of GDP) (% of GDP) 7 6 B udget 7.0 deficit 5 6.5 6.0 4 5.5 3 Net fo reign 5.0 financing 2 4.5 Net 4.0 1 do mestic financing 3.5 0 3.0 FY90 FY93 FY96 FY99 FY02 FY05 FY90 FY92 FY94 FY96 FY98 FY00 FY02 FY04 FY06 Source: Ministry of Finance. • Positive growth performance and improving fiscal balances in recent years reversed the increase in government debt. Strong GDP growth, averaging over 5 percent annually during the same period and the reduction in the value of domestic debt due to inflation lowered the debt-to-GDP ratio. As a result of these and other factors, total debt to GDP stood at around 44.4 percent of GDP at the end of 2008 (Table 1-2). Prior to that, debt had increased by nearly 10 percentage point of GDP, reaching 53 percent of GDP in 2002 from 44 percent in 1999. This was caused largely by an increase in the primary fiscal deficit. The share of domestic debt also increased (Figure 1-4). 17 Table 1-2: Central Government Debt Stock, 2000-2008 (% of GDP) Component 2000 2001 2002 2003 2004 2005 2006 2007 2008 Total 47.3 52.9 53.1 51.1 48.7 47.6 46.9 45.0 44.4 Domestic 14.8 19.7 18.7 18.3 18.3 18.1 18.2 16.8 18.5 External 32.5 33.2 34.4 32.8 30.4 29.5 28.7 28.2 25.9 Source: IMF and WB staff estimates. Note: Central Government Debt underestimates the actual size of total debt as debt of SoEs is not included. Figure 1-4: Increase in Central Government Debt Reversed (% of GDP) T o ta l 5 0 4 0 E xte r n a l 3 0 2 0 D o m e s ti c 1 0 F Y00 F Y01 F Y0 2 FY0 3 F Y 04 FY 05 FY 06 F Y07 F Y08 Source: Ministry of Finance. 55. The quality of the fiscal performance has been weak. • Mobilization of public revenues has been very low, although the last year recorded encouraging improvements. Total revenues have stagnated at around 10 percent of GDP in the last seven years, with tax revenues at about 8.5 percent of GDP, indicating a very low buoyancy to the country’s increasing growth performance. Wrapped in old policies and outdated administrative practices, the tax system has not been able to bring new income- generating activities within the tax net. Revenue generation relies heavily on trade taxes with nearly 45 percent of tax revenue coming from import-based taxes. Tax expenditures stemming from a proliferation of preferential treatments have cost the budget an estimated 2.5 percent of GDP in FY05-06. Weaknesses in the administration continue to cause tax leakages, evasion and inefficient revenue mobilization. • The recent surge in global oil and food prices increased significantly transfers in the form of subsidies to cover the losses of SoEs, especially on petroleum, electricity and fertilizer. The failure to continue closing loss-making manufacturing SoEs after 2003 and to appropriately adjust administered prices of energy and fertilizer has restored the rising trend in SoE deficit in FY06 and FY07. Overdue loans and mounting losses have already resulted in substantial transfers from the government to SoEs in the form of subsidies, debt assumptions, on-lending and equity participation. The share of budget allocations to SoEs increased from 26 percent of the investment program (the Annual Development Program: ADP) in 2001 to about 44 percent in 2006 before declining to 28 percent in FY08 (Figure 1- 5). These shares, however, might appear larger due to more transparent reporting on SoE losses from the CTG. 18 Figure 1-5: Share of Budget Allocation to SoEs Under ADP (%) 5 0 4 5 4 0 3 5 3 0 2 5 2 0 1 5 1 0 5 0 F Y 0 1 F Y 0 2 F Y 0 3 F Y 0 4 F Y 0 5 F Y 0 6 F Y 0 7 F Y 0 8 Source: Ministry of Finance. • In the context of low revenue mobilization, fiscal adjustment has been driven mainly by a decline in public development spending. Development expenditures have declined to 3.4 percent of GDP in FY08, after peaking at 6.6 percent of GDP in 2001, due to low absorption and weak implementation capacity (Figure 1-6). Significant structural rigidities continue to hamper the implementation of the development budget (investment program). At the same time, recurrent expenditure, which, at 60 percent, comprises the largest share of total central government expenditure, has risen by 20 percent in the period FY00-FY08, from 7.2 percent of GDP to 9.6 percent. Figure 1-6: Trends in Central Government Expenditure (% of GDP) 16 Total Expenditure 14 12 10 Recurrent Expenditure 8 6 4 ADP 2 0 FY 91 FY 94 FY 97 FY 00 FY 03 FY06 Source: Ministry of Finance. 56. These vulnerabilities, including low revenue mobilization, increased fiscal subsidies and declining public investments are threatening Bangladesh’s prospects for maintaining the positive macro-fiscal outcomes that have been achieved. They are also threatening prospects for maintaining fiscal sustainability that is financing core government activities without incurring additional debt. 19 1.2 Maintaining Fiscal Sustainability 57. Fiscal policy in Bangladesh extends well beyond the central government through a large public participation in nonfinancial and financial enterprises as well as through the Bangladesh Bank (BB). While not unique in having a dominant public enterprise sector involved in the provision of goods and services in key sectors of the economy, Bangladesh has been rather slow to reform the sector to better fit the changing business environment. The sustainability of its fiscal stance, therefore, depends not only on the fiscal position of the central government, as discussed in the preceding sections, but also on the performance of the broader public sector, especially in light of high fiscal and economic costs associated with the current condition of SoEs (Appendix 1 has a detailed description of the structure of the public sector). 1.2.1 The Fiscal Performance of the Public Sector 58. The Consolidated Public Sector Deficit (CPuS)11 experienced a dramatic improvement during the FY02-06 period, before slipping somewhat in FY07. The CPuS deficit, as measured by the saving- investment gap, declined by half from 6.6 percent of GDP in FY00 to 3.2 percent in FY06, driven by significant decreases in both the central government and SoE deficits. However, as a result of deteriorating fiscal performance, in particular for SoEs in the energy sector, the CPuS lost all its gains and increased by almost the same magnitude to reach 5.2 percent of GDP in FY07 (Table 1-3). Table 1-3: Public Sector Fiscal Performance (% of GDP) Element FY98 FY99 FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07 Central Government Saving 1.6 0.5 0.6 0.8 1.4 1.8 2.2 1.6 1.4 0.5 Investment 4.5 4.8 5.1 5.3 3.8 3.9 3.6 3.7 2.9 2.9 Deficit 2.9 4.3 4.5 4.5 2.4 2.1 1.4 2.1 1.5 2.4 State-Owned Enterprises (non-financial) Saving 0.6 0.6 -0.1 -0.2 0.1 0.5 0.4 0.3 -0.3 -0.6 Investment 1.5 1.6 1.9 1.9 2.1 1.8 1.6 1.3 1.2 2.0 Deficit 0.9 1.0 2.0 2.1 2.0 1.3 1.2 1.0 1.5 2.6 Public Financial Institutions Saving 0 -0.1 -0.2 0 0 0.2 -0.6 -0.2 0.3 -0.2 Provisioning shortfall 2.2 2.2 1.7 1.6 1.7 1.8 1.6 1,2 1.3 — Consolidated Public Sector Saving 2.1 1.0 0.3 0.6 1.7 2.2 2.1 1.8 1.0 -0.3 Investment 6.5 6.1 6.4 6.9 5.2 5.7 5.3 5.0 4.1 4.9 Deficit, excl. Provisioning. shortfall) 4.0 5.4 6.6 6.3 3.5 3.4 3.2 3.2 3.2 5.2 Deficit, incl. provisioning shortfall 6.2 7.6 8.3 7.9 5.2 5.2 4.8 4.4 4.5 – Source: World Bank staff estimates. 59. Despite the deterioration in FY07, the recent fiscal performance by the central government is sustainable. This is evidenced by the analysis using both the long-run sustainability method and the extended accounting framework method (see Appendix 3). The fiscal performance of the consolidated public sector, however, appears to be questionable. Analysis indicates that sustainability of the fiscal 11 CPuS is defined as the sum of the investment–saving gap of the consolidated public sector, consisting of central government, nonfinancial public enterprises and public financial institutions, including the Bangladesh Bank. For more details on the definition of the Consolidated Public Sector see Appendix 2. (attach attachment 2 20 stance is heavily dependent on concessional financing, which translates into low interest rates. It is also very sensitive to growth performance. 1.2.2 Central Government Long-run Sustainability 60. Current primary deficit levels (1.2 percent of GDP in FY07 and the 1.5 percent in FY08) are sustainable provided that GDP growth is 6.5 percent or higher and the nominal interest rate is 9 percent or lower. Assuming a debt-to-GDP ratio of 47 percent, Table 1-4 presents a range of primary balances corresponding to varying nominal interest rates and real GDP growth rates. It shows that the estimated primary balance is very sensitive to assumptions about growth and interest rates. Roughly speaking, at any given nominal interest rate, for each additional percentage point of growth the government has room to run a larger primary deficit by about 0.2-0.3 percentage points of GDP. For each additional percentage point on the interest rate, the required reduction in primary deficit is about 0.4 percentage point of GDP. This indicates that if Bangladesh were to suffer a sustained rise in interest rates or a sustained decline in its growth, it would need to run a substantially smaller primary deficit, or even a primary surplus, in order to meet its debt target. Table 1-4: Sustainable Primary Balances (% of GDP) Nominal GDP Growth Rates (%) Interest rates (%) 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 6 -1.1 -1.4 -1.6 -1.9 -2.1 -2.4 -2.6 -2.9 -3.1 7 -0.7 -0.9 -1.2 -1.5 -1.7 -2.0 -2.2 -2.5 -2.7 8 -0.3 -0.5 -0.8 -1.0 -1.3 -1.5 -1.8 -2.0 -2.3 9 0.2 -0.1 -0.4 -0.6 -0.9 -1.1 -1.4 -1.6 -1.9 10 0.6 0.3 0.0 -0.2 -0.5 -0.7 -1.0 -1.2 -1.5 11 1.0 0.7 0.5 0.2 -0.1 -0.3 -0.6 -0.8 -1.1 12 1.4 1.1 0.9 0.6 0.4 0.1 -0.2 -0.4 -0.7 13 1.8 1.6 1.3 1.0 0.8 0.5 0.2 0.0 -0.3 14 2.3 2.0 1.7 1.4 1.2 0.9 0.7 0.4 0.1 15 2.7 2.4 2.1 1.9 1.6 1.3 1.1 0.8 0.5 Note: Benchmark values for other parameters: Total debt-GDP ratio = 48.3%; Inflation rate (GDP Deflator) 4%; and base money-GDP=7.1%. 1.2.3 Central Government: The Extended Accounting Framework12 61. The sustainable level of the central government primary deficit is estimated at 2.8 percent of GDP, well above the 1.2 percent outturn in FY07 and the 1.5 percent for FY08. Adding domestic and foreign interest payments that are consistent with the financial and macro variables in FY07, the total central government deficit reaches 4.8 percent, which is well above the actual level of 3.1 percent in FY07 and 3.7 percent in FY08. However, the sustainable level of the primary deficit declines to 1.0 percent of GDP when domestic and foreign interest subsidies are excluded (Table 1-5). 12 This method disaggregates the debt into its two components–domestic and foreign–and tries to quantify the impact of real interest rate subsidy on domestic and foreign debt that underlies Bangladesh’s debt sustainability scenarios. 21 Table 1-5: Sustainable Levels of Deficit (% of GDP) Stationary Debt Stationary Debt Without Without Domestic Domestic and Foreign Debt Stationary Debt Debt Subsidy Subsidy Component Central Government Total Deficit 4.8 4.8 4.4 Domestic Interest Payments 1.7 2.7 2.7 Foreign Interest Payments 0.3 0.3 0.7 Primary Deficit 2.8 1.8 1.0 Net Growth Effect on Domestic Debt -0.6 -0.6 -0.6 Domestic Real Interest Subsidy 1.0 0 0 Net Growth Effect on Foreign Debt 1.6 1.6 1.6 Foreign Real Interest Subsidy 0.4 0.4 0 Foreign Grants 0.4 0.4 0 Consolidated Public Sector Primary Deficit 3.5 2.5 2.1 Net Growth Effect on Money 0.6 0.6 0.6 Inflation Tax 0.6 0.6 0.6 Source: World Bank staff estimates. 62. The government’s medium-term budget framework aims at reducing the primary deficit from 2.9 percent of GDP in FY09 to 1.8 percent by FY11. The magnitude of the reduction in the primary deficit is in line with what is needed to maintain sustainability with declining concessional foreign financing and the increasing cost of domestic debt. If foreign grants are reduced to zero, a plausible scenario in the not-too-distant future, the sustainable level of primary deficit would be lowered to 2.4 percent of GDP. In addition, if the actual interest rate paid by the government on its domestic debt increased (to the level of the real domestic shadow rate of 10 percent), the sustainable primary deficit level would drop to 1.8 percent of GDP, above the FY08 outturn but below the 2.9 percent central government primary deficit projected for FY09. While a rise in the real interest rate paid by the government is unlikely to materialize in the short term, it may affect future government financing when all domestic debt is serviced at market-based rates. Furthermore, if the foreign interest rate subsidy is reduced to zero, the sustainable primary deficit is reduced to 1 percent of GDP. This situation will only materialize when Bangladesh attains middle-income country status with no further access to concessional funding from abroad. For the time being, however, the mobilization of concessional foreign financing will remain critical in avoiding any major fiscal adjustment to maintain sustainability. 1.2.4 Consolidated Public Sector: Extended Accounting Framework 63. The sustainable level (in a narrow sense) for the CPuS primary deficit is estimated at 3.5 percent of GDP. This is close to the ratio (not including State Commercial Banks (SCB) provisioning shortfall) observed in FY06, but well below the 5.2 percent CPuS deficit observed in FY07. The sustainable level of the CPuS primary deficit declines to 2.1 percent of GDP when domestic and foreign interest subsidies are reduced to zero. This long-term sustainable primary deficit is determined by the combination of seigniorage financing, negative financing from the net growth effect on domestic debt, and the net growth effect on foreign debt. Compared to FY07 deficit levels, meeting this long-term fiscal sustainability condition would necessitate cutting the primary deficit by nearly 60 percent–from 5.2 percent of GDP to 2.1 percent. 22 1.3 Debt Dynamics and Sustainability 64. Bangladesh’s debt indicators improved significantly in the period FY02-07 due to fiscal adjustment, rising GDP growth, higher inflation tax and real exchange rate appreciation. Expenditure reduction and improvements in revenue mobilization reduced the primary deficit from an average of 3.4 percent of GDP during FY99-02 to 1.4 percent during FY03-07. Strong growth effects, averaging 2.7 percent of GDP, and inflation tax averaging 0.8 percent of GDP, reinforced the effect of primary deficit reduction on total debt. In addition, appreciation of the real exchange rate eroded the real value of foreign debt by about 0.5 percent of GDP per year. Reliance on seigniorage was also reduced somewhat, although it increased very significantly in the last two years (Table 1-6). Table 1-6: Central Government Debt Dynamics (FY99-07)(% of GDP) Component FY99 FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07 Change in Debt 0.7 3.5 5.7 0.2 -2.0 -2.4 -1.1 -0.7 -1.9 A. Sum of Standard Terms* 4.3 4.2 4.4 3.0 3.1 2.3 2.5 1.3 1.8 Interest Payments 1.3 1.5 1.6 1.7 1.9 1.6 1.7 1.8 1.9 Primary Balance -3.5 -3.6 -3.5 -3.0 -1.5 -1.2 -1.7 -1.5 -1.3 Seigniorage 0.5 1.0 0.7 1.7 0.3 0.6 0.9 2.0 1.4 B. Sum of Extra Items -2.4 -1.6 0.5 -3.3 -4.5 -4.0 -3.5 -2.7 -5.4 Growth Effect -1.9 -2.4 -2.3 -2.2 -2.5 -2.9 -2.6 -2.8 -2.6 Inflation Effect -0.5 -0.2 -0.2 -0.6 -0.8 -0.7 -0.9 -0.9 -1.0 Revaluation Effect 0.0 1.0 3.0 -0.5 -1.2 -0.4 0.03 1.02 -1.8 Memo Items Overall Budget Deficit 4.8 5.1 5.1 4.6 3.4 2.9 3.4 3.2 3.2 Errors and Discrepancy -1.2 1.0 0.9 0.5 -0.6 -0.7 -0.1 0.7 1.7 Inflation Rate 4.7 1.9 1.6 3.2 4.5 4.2 5.1 5.2 5.6 GDP Growth Rate 4.9 5.9 5.3 4.4 5.3 6.3 6.0 6.6 6.4 Source: World Bank staff estimates. Note: *Equals interest payments minus primary balance minus seigniorage. 65. Bangladesh’s risk associated with the level and composition of its overall debt is moderate and it is low for external debt as shown by the recent debt sustainability analysis.13 1.3.1 External Debt Sustainability 66. All external debt indicators are below the policy-dependent debt burden thresholds under the baseline scenario, but one threshold is breached under the most extreme standard stress tests.14. The Net Present Value (NPV) of debt-to-GDP ratio decreases from about 17 percent at end FY07 to 15 percent by end FY2011, compared to an indicative threshold of 40 percent; the NPV of the debt-to- export ratio decreases from 92 percent to 68 percent, compared to an indicative threshold of 150 percent; and the debt-service ratio decreases from over 5 percent of exports to about 4 percent, compared to an indicative threshold of 20 percent (Table 1-7). 13 This section highlights the results of the joint IMF-WB Debt Sustainability analysis that was completed in 2008. 14 These provide the upper ceiling of a confidence interval to the baseline projections. The first four tests assume that real GDP growth, export growth, inflation, and nondebt flows, including both FDI and current transfers, in each of the first two years, are one standard deviation below the historical average. Another test combines all four variables and assumes that, in each of the first two years, they are a one-half standard deviation below the historical average. A more severe test assumes a one-time 30 percent depreciation of the domestic currency in the first year of the projection period. 23 Table 1-7: Debt Indicators Indicator Threshold 2005 2011 PV of Debt as a Share of: GDP 40 17 15 Exports 150 92 68 Revenues 250 166 115 Debt Service as a Share of: Exports 20 5 4 Revenues 30 10 8 Source: Joint Bank-Fund DSA 2008. 67. The standard stress tests do not reveal any serious vulnerability; however, Bangladesh’s GDP will have to shrink at a rate of over 3 percent per annum for the next 10 years for debt ratios to exceed indicative thresholds. The baseline projections and the associated standard stress tests show little risk related to external debt; none of the indicators breaches or is close to the indicative debt burden thresholds. There is a temporary (for two years) and minor breach of one threshold in which a half standard deviation shock is given to GDP growth, exports, net current transfers, and FDI.15 This, however, does not change the conclusion. A failure to implement structural reforms, particularly those in power, ports and banking, could significantly depress GDP growth rates. 1.3.2 Public Debt Sustainability 68. With both external and domestic debt declining relative to GDP in the baseline scenario, both the Net Present Value (NPV) of public debt-to-GDP and NPV of public debt-to-revenue are projected to decline. However, the alternative scenarios and bound tests point to significant risks. Contingent liabilities from underpricing of energy products have been growing by 1 percent of GDP per annum, thus elevating the risks of large debt-creating flows. Even under a very conservative assumption of a 10 percent one-off debt creating flow, the debt service to revenue ratio reaches 36 percent in 2010, compared with the baseline ratio of 23 percent. Contingent liabilities combined with the risk of lower GDP growth or a failure to improve revenue performance pose material risks to public debt sustainability. 1.4 Key Risks and Priorities in the Medium Term 69. While Bangladesh’s fiscal stance appears to be sustainable, the level of comfort that can be derived from it is limited because of several vulnerabilities that are looming large on Bangladesh’s medium-term economic horizon: • insufficient reforms to improve revenue mobilization. The government’s program envisages a decrease in the primary deficit from 2.9 percent of GDP in FY09 to 1.8 percent in FY11. This is achievable provided that the projected level of concessional external financing and the increase in revenue mobilization materialize. The risk that the proposed strengthening of government revenues may not take place cannot be discounted, particularly considering recent experience. Whether or not the exceptional revenue growth performance this year will be sustained is still an open question. Significant efforts would be needed to ensure that the tax administration reforms translate into revenue gains. 15 The breach emerges largely because the relative magnitude of the shock is substantially higher compared to the previous DSA (2006). 24 Resistance to fundamental changes in the National Board of Revenue (NBR) remains strong and technical capacity is a significant constraint. While revenue collection in FY08 exceeded the original budget target by 7.6 percent, tax revenues collected by the NBR in FY07 were 0.9 percent of GDP lower than the original FY07 budget; reflecting continued weaknesses in tax administration (see Section 1.52 for a more detailed discussion). • inadequate adjustment (until recently) of the administered prices of diesel, power, natural gas, and fertilizer, the brunt of which were borne by the intermediating state- owned companies. Administered petroleum prices were last adjusted in October 2008 and power prices in March 2007, while fertilizer prices were adjusted in June 2008. The financial loss from underpricing of fertilizer, diesel, power and natural gas was around $1.8 billion, or 2.5 percent of GDP in FY08. The government financed a large part of this deficit through the central government budget.16 Recent falls in international prices–particularly for oil and fertilizer–have eased the pressure on SoE finances, but their sustainability remains to be seen. • public sector pension and wage increases. Some 828,000 civil servants, 140,000 defense staff and 560,000 teachers on the Monthly Payment Order (MPO) are currently on the central government payroll. Pension and related liabilities are expected to rise following the implementation of the revised pay scale in 2005, which offered 53-60 percent salary increase to be implemented over a three-year period. This has expanded the base on which a pension is determined. Cost of living adjustments for pensions received by existing pensioners further increase the pension liabilities, which have grown on average by 19 percent annually. This is a liability that is projected to increase by around 30 percent following the implementation of the new pay scale. This is a very sizable increase in budgetary liabilities, which will create pressure to reduce discretionary spending or to allow the deficit and debt to grow. • higher interest rates and large real exchange rate depreciation. Bangladesh is fortunate that the cost of interest payments places a relatively low burden on the budget. Interest payments are programmed to be slightly less than 2 percent of GDP in FY08-11. This is mainly because most of the public debt is multilateral or bilateral and is on highly concessional terms. The effective nominal interest rate facing the government was about 4.8 percent in FY07,17 which is less than two-thirds of the borrowing cost on international or domestic markets.18 If prospects for concessional lending diminish and Bangladesh is pushed towards market-based rates, the cost of financing could increase significantly, which would worsen the budget balances and the debt situation. Large real exchange rate depreciation can add additional pressure. • lower growth. The Bangladesh economy has recently been on a modest growth path by East Asian standards. The average growth rate of real GDP during FY04-08 was about 6-plus percent. Several factors constrain the growth performance, including exposure to external shocks, weak private investment, the high cost of doing business and export competitiveness problems.19 The benchmark scenario above assumes that Bangladesh will be able to address these profound challenges sufficiently to substantially lift the steady- 16 Starting from the FY08 revised budget, the definition of budget deficit has been broadened to explicitly include the financing of the operating losses of the four largest loss-making nonfinancial energy and fertilizer SoEs. The FY08 revised budget covers 76 percent of these losses. 17 This is the average of the interest rates on foreign and domestic debt weighted by the relative amounts of the two types of debt. 18 The average commercial bank nominal lending rate was 15 percent at the end of June 2004. 19 World Bank, Growth and Competitiveness report, 2005. 25 state growth performance. However, despite the pick-up in growth rates in recent years, a credible strategy to lift growth rates to a steady state of 6.5 to 7 percent is still lacking. As a consequence, there remains a plausible risk of steady-state growth remaining below the 7- 7.5 percent rate required to achieve middle-income status over the next decade and a half. The ongoing global economic crisis has materially increased the risk of a growth slowdown in Bangladesh, at least temporarily, due to the likely adverse impact on exports and remittances. Box 1-1: Macroeconomic Assumptions Underlying the Debt Sustainability Analysis (DSA)20 The macroeconomic assumptions are as follows: Real GDP growth, at about 6.4 percent, is above the recent historical average of 5.6 percent, picking up a bit in the outer years. This is close to (but slightly lower than) Bangladesh's PRSP projections, and assumes continued progress in broad-based structural reforms and increased openness of the economy that should allow Bangladesh to benefit from dynamic growth elsewhere in the Asian region. Inflation, as measured by the GDP deflator, decreases in the short term and stabilizes at around 4 percent. The growth of exports and imports is strong in the medium term (14 percent and 17 percent respectively) as the economy opens. Both will increase in terms of GDP with imports gaining the most as increasing investment and intermediate goods are imported. The current account (including grants) moves from a small surplus to a deficit, which peaks in the outer years at about 2 percent of GDP, as a result primarily of continued strong growth of capital and intermediate goods imports related to increasing investment projects. These effects are offset to some extent by strong growth of remittances, which are projected to grow at an annual average of about 13 percent over the medium and long term. Net aid inflows reach 2 percent of GDP and stabilize in that range (consistent with Bangladesh’s medium-term framework). The projections assume that the grant element of new borrowing decreases over the 20-year period from 48 percent to 39 percent in the outer years. The overall fiscal deficit (excluding grants) is assumed to remain close to the historical average (around 4 percent of GDP), while the primary deficit declines slightly over time. A modest rise is assumed in the revenue-to-GDP ratio (excluding grants) in the initial years (from 11.5 percent in FY2008 to 13 percent in FY13), supported by efforts to mobilize domestic revenues. Real interest rates on domestic currency debt are assumed to stay more or less constant at about 3.5 percent. 70. Considering these risks, the government’s medium-term fiscal policy is ambitious but not beyond the realm of feasibility, if macroeconomic stability is sustained and progress with policy and structural reforms continues.21 The medium-term fiscal policy is consistent with public debt sustainability. Overall, the central government budget deficit is projected to fall back to less than 4 percent of GDP by FY11. This decline is assumed to come from both a sustained increase in the revenue- to-GDP ratio in FY09-11 and about a 0.4 percentage point of GDP decline in the expenditure-to-GDP ratio after FY09. After rising to 2.8 percent of GDP in FY09, domestic financing is projected to decline to 2 percent of GDP in FY11. External financing is projected to remain at its current level of 2.2 percent of GDP in FY09-10 before declining to 2 percent of GDP in FY11. Nevertheless, achieving these parameters 20 Note that the DSA was completed before the meltdown in the global financial system. Since the crisis is still unfolding (at the time of writing this report), it is premature to factor in its impact on debt sustainability. 21 The fiscal path envisaged above does not take into account the potential capital costs of SoE and NCB reforms that are tentatively estimated at up to 10 percent of GDP to be absorbed during FY08-11, mainly below-the-line in the budget for eventual NCB recapitalization and resolution of SoE arrears. 26 will be a challenge that requires concerted efforts to implement several policy and institutional reforms. In light of growing development expenditure needs and the envisaged reduction in trade taxes, strengthening domestic revenue mobilization should be at the core of the medium-term budget strategy (Table 1-8). Table 1-8: Medium-Term Budget Framework (% of GDP) Actual Estimate Projections Component FY04 FY05 FY06 FY07 FY08 FY09 FY10 FY11 Total Revenue 10.2 10.5 10.8 10.4 11.4 11.0 11.0 11.9 Tax Revenue 8.3 8.5 8.6 8.3 9.1 8.9 8.9 9.8 NBR Taxes 7.9 8.1 8.2 7.9 8.7 8.5 8.5 9.4 Non-NBR Taxes 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 Non-tax Revenue 1.9 2.0 2.2 2.1 2.3 2.1 2.1 2.1 Total Expenditure 13.1 14.0 14.1 13.5 15.1 14.4 16.1 15.9 Current Expenditure 7.8 8.4 8.4 9.0 9.6 9.9 10.9 10.8 Pay & Allowances 2.3 2.3 2.4 2.5 2.5 2.3 2.8 2.8 Goods & Services 1.4 1.5 1.5 1.5 1.5 1.4 1.6 1.7 Interest 1.6 1.7 1.8 1.9 2.2 2.1 2.2 2.2 External 0.3 0.4 0.3 0.3 0.2 0.2 0.2 0.2 Domestic 1.3 1.3 1.7 1.6 2.0 1.8 2.0 2.0 Subsidies & Transfers 2.4 2.8 2.6 3.0 3.6 3.8 3.9 3.8 Block Allocations 0.1 0.2 0.1 0.1 0.1 0.6 0.4 0.3 Food Account Balance 0.1 0.0 0.0 0.1 0.1 0.1 0.1 0.1 Annual Development Program 5.0 5.0 4.7 3.8 3.4 3.4 4.1 4.0 Non-ADP Capital & Net Lending 0.5 0.7 1.0 0.6 1.6 0.8 0.8 0.8 Non-ADP Capital Spending 0.8 0.8 1.2 0.7 1.3 0.6 1.0 1.1 Net Lending -0.3 -0.1 -0.2 -0.1 0.3 0.2 -0.2 -0.3 Other Expenditures 0.2 0.1 0.0 0.0 0.3 0.2 0.2 0.2 Check float & discrepancy -0.5 -0.2 0.0 0.0 0.0 0.0 0.0 0.0 Overall Balance (excl. Grants) -2.9 -3.5 -3.4 -3.1 … … -4.5 -4.0 Overall Balance (excl. Grants and incl. budget support to SoEs) … … … … -3.7 -3.5 … … Primary Balance (excl. Grants) -1.3 -1.8 -1.6 -1.2 … … -2.3 -1.8 Primary Balance (excl. Grants and incl. budget support to SoEs) … … … … -1.5 -1.4 … … Net Financing 2.9 3.5 3.4 3.1 3.7 3.5 4.5 4.0 Net External Financing 1.1 1.8 1.2 1.2 2.2 1.7 2.2 2.0 Domestic Financing 1.8 1.7 2.2 1.9 1.5 1.8 2.3 2.0 Banks 0.4 0.9 1.6 1.2 1.1 1.3 1.9 1.7 Non-bank 1.4 0.8 0.7 0.7 0.4 0.6 0.4 0.3 Memorandum Item: Total Central Govt. Debt (% of GDP) 48.7 47.6 46.9 45.0 44.4 43.8 42.8 41.7 Source: Ministry of Finance. 1.5 Improving Revenue Performance: The Cornerstone of the Medium-Term Fiscal Strategy 71. Stagnating at around 10 percent of GDP for almost a decade, total revenues in Bangladesh have been largely irresponsive to changes in economic activity.22 Tax revenues have recorded a gradual but modest increase; however, their share to GDP has been virtually flat over the last 10 years. This implies that revenues have been largely unaffected by changes in economic activity as measured by the revenue elasticity,23 which is close to unity (Table 1-9). This is remarkable for an economy that has been growing rapidly over the last decade and a half. It appears that the existing tax system, based on old 22 This section draws on the assessment that has been done by the IMF in January 2007. 23 Revenue elasticity is calculated as the percentage change in revenues in respect of a percentage change in nominal GDP. 27 policies and procedures, is unable to bring new income-generating activities within the tax net. Cross- country comparisons within the region and outside confirm these results (Figures 1-7 and 1-8). Table 1-9: Revenue Elasticity in Bangladesh Component 1990s 2000-07 Total Tax Revenue (TR) 1.2 1.1 Income Tax Revenue (ITR) 1.1 1.6 VAT (VTR) 1.4 1.9 Customs Revenue (CTR) 1.2 0.8 Source: NBR, BBS, and WB staff estimates. Note: VAT = Domestic VAT; Customs Revenue = Rev from CD, SD, VAT on imports. Figure 1-7: Per Capita GDP and Tax-To-GDP Ratio: Figure 1-8: Tax Revenues (2005) (% of GDP) Low-Income Countries 50 38.2 20 40 T a x to G D P R a tio in p e rc e n t 16 30 18.7 16.4 20 11.0 8.5 12 10 8 0 Bangladesh OECD Upper Middle Lower Low Income Bangladesh 4 Countries Income Middle Countries (2003) Countries Income 0 0 200 400 600 800 1000 Countries Per Capita GDP in US dollars (2000) Sources: Government Finance Statistics 2004, IMF and World Development Indicators 72. In spite of the low level of tax collection, the statutory nominal tax rates in Bangladesh are generally high among the comparators. A VAT rate of 15 percent is higher than most countries in South and East Asia. The Corporate Income Tax (CIT) for publicly-listed companies (30 percent) is more or less compatible. However, the higher corporate income tax rates applied to non-listed companies (40 percent) and financial institutions (44 percent) are significantly above the CIT rates of comparable countries. On the other hand, the top marginal rate for personal income tax is lower than most comparator countries. 73. With comparatively high nominal tax rates and low revenue yields, Bangladesh’s revenue productivity24 is low (Table 1.10). Revenue productivity of both CIT and VAT, for instance, is well behind other countries in the region, clearly indicating the enormous potential for generating extra revenue from these sources. The data reveals that even customs and corporate taxes have sufficient scope for improvement. 24 Revenue productivity refers to the share of revenue (income tax, VAT, or customs) that is generated by each percentage point in the respective nominal tax rate (i.e. tax/GDP ratio divided by the applicable nominal tax rate). 28 Table 1-10: Revenue Productivity Ratios for Bangladesh and Comparators Revenue Source Bangladesh India Nepal Pakistan Sri Lanka Vietnam Thailand Malaysia Customs 8.2 10.6 21.1 10.5 18.5 22.9 16.2 17.3 VAT 18.8 N/A 29.3 25.8 40.4 58.1 56.2 N/A Corporate Income 3.4 7.9 4.6 6.7 2.5 27.7 14.3 30.7 Tax Source: NBR & BBS for Bangladesh, IFS and IMF (2007). Notes: FY07 data for Bangladesh, FY04 for rest of the countries. VAT = Domestic + Import Stage; Customs = revenue from CD and SD on imports. 74. The low revenue-to-GDP ratio can be attributed to: (i) heavy reliance on trade taxes and the low level of domestic taxes; (ii) significant tax expenditure, which erodes the tax base; and (iii) major organizational and institutional weaknesses in tax administration. With over 30 percent of its tax revenues coming from international trade, Bangladesh continues its long tradition of relying on more distortive import-based taxes. Relative to its comparators (Table 1-11), Bangladesh has a higher reliance on trade rather than domestic taxes. Although the NBR has been endorsing a strategic shift from its current policy for quite some time, the actual shift has only started over the past two years. With a highly restrictive trade regime25, the need for further trade liberalization precludes the scope for relying on import taxes to generate more revenues. There is thus a stronger case for increasing the domestic tax-to-GDP ratio through broadening the tax base and improving their management, in particular VAT and income tax. Table 1-11: Share of Trade and Nontrade Taxes for Selected Asian and African Countries As % of GDP As % of Tax Revenue Domestic International Domestic International Comparator Taxes Trade Taxes Taxes Trade Taxes South Asian and Pacific Countries Bangladesh 5.6 2.6 68.0 32.0 Nepal 6.6 3.1 67.7 32.3 Vietnam 18.5 3.0 83.1 16.9 India 8.5 1.7 82.9 17.1 Pakistan 8.8 1.5 89.6 10.4 Bhutan 7.1 0.5 93.7 6.3 Philippines 9.8 2.5 79.4 20.6 Sri Lanka 12.0 2.0 85.7 14.3 Indonesia 11.9 0.6 95.5 4.5 China PR 8.7 0.8 89.6 10.4 Thailand 15.8 1.6 90.5 9.5 Malaysia 16.3 1.3 92.5 7.5 Unweighted Average (excluding 11.3 1.7 86.4 13.6 Bangladesh) Sub-Saharan African Countries Uganda 7.6 3.4 69.3 30.7 Ghana 17.6 4.8 78.6 21.4 Kenya 16.6 0.6 96.7 3.3 Cote d'lvoire 6.7 7.9 45.9 54.1 Unweighted Average 12.1 4.2 72.6 27.4 Source: IMF (2007), pp 30-31, Data on China from Government Finance Statistics 2004 and WDI. 25 th In The World Bank’s World Trade Indicators 2007, Bangladesh is ranked 88 amongst 96 countries in respect of trade restrictiveness. 29 75. A recent assessment by the NBR estimated that revenue forgone due to various types of tax expenditure could be as much as 2.5 percent of GDP. This includes revenue foregone due to provision of tax holidays, exemptions and deductions, rate reduction, deferrals and tax credits (Table 1-12), a policy that Bangladesh has carried well beyond its optimal point. While their stated objective is to promote investments and growth, in all probability they are now counterproductive for at least two important reasons. First, they are strongly distortionary and adversely impact resource allocations in the economy, leading to lower than potential growth. Second, their adverse revenue impact severely constrains the financing of productive government expenditure. Anecdotal evidence also suggests that the primary beneficiaries of these concessions were the various lobbyists and vested interest groups. Table 1-12: Bangladesh Tax Expenditure (FY2005/06)(% of GDP) Component Amount Direct tax 0.28 Corporate Tax 0.22 Personal Income Tax 0.05 Indirect Tax 2.24 VAT 2.07 Imports 0.48 Domestic 1.59 Supplementary Duties 0.07 Customs Duties 0.10 Total 2.52 Source NBR, IMF (2007). 76. The administration of revenue collection is weak, with the revenue agency, NBR, organized on a tax-by-tax–income tax, VAT and customs–basis. Moreover, there are two distinct cadres of officials recruited to the customs and income tax administrations, with somewhat divergent career and promotion prospects. Revenue mobilization is negatively affected by the lack of intra-departmental exchange of information and minimal cross-tax monitoring. Intra-cadre rivalry and silo-type mentality prevails, resulting in the undermining of revenue mobilization efforts. Increasing voluntary compliance through modernization of revenue administration should be the ultimate objective of any revamped tax agency. 77. The current situation calls for a comprehensive reform agenda. Introducing simple, coherent and effective tax policy provisions must take priority, especially for domestic taxes, such as VAT and income tax. These are likely to be the future mainstay of the Bangladesh tax system, as the need to reduce dependence on international trade tax becomes imperative. 1.5.1 Modernizing the Value Added Tax System 78. Bangladesh’s VAT operates like an extended excise system of taxation. After 15 years of application, VAT still resembles the system it sought to replace. The method is applied mainly at the production/wholesale stage, at a rate of 15 percent. This partial application is only applicable to transactions involving manufacturers, importers and perhaps large raw material producers. Domestic wholesale and retail traders are generally subject to what is, in fact, a 1.5 percent turnover tax. The VAT in Bangladesh represents a pervasive excise-cum-turnover tax system with over 6,500 product definitions in the customs code. It distinguishes -- for tax purposes-between manufacturers, wholesalers and retailers, it has separate definition and treatment of goods and services, presumptive valuations methods and other features of an extended excise system. It is a far cry from a true VAT 30 system, which does not require product definition, the delineation of various stages of production and distribution, a distinction between goods and services, or approved prices. As the 2003 Report of the Revenue Administration Commission puts it “the excise psyche is still present in the VAT”. 79. The excise and turnover features of the VAT violate several good principles of taxation, including the neutrality criteria of a true VAT. These features make VAT in Bangladesh a highly differentiated tax whose incidence is indeterminate and capricious. More importantly, the workings of the VAT greatly increase uncertainty and the free functioning of business and trade. This is because every business has to engage in a bargaining process with the VAT administration to determine and settle its tax liability on the basis of the criteria that are probably not well understood by either party. Furthermore, the concern with legal, procedural and technical details instead of taxpayer education, self-assessment and accounting controls has made the administration of the VAT exceedingly complex. 80. The inefficiencies of VAT revenue performance are also clearly demonstrated by the low collection efficiency measured by VAT “productivity” on the basis of production26 (GDP). This indicates what percentage of GDP is collected for each percentage point of the standard VAT rate. In 2001, for example, Bangladesh collected 0.17 percent of GDP for each percentage point of the standard VAT rate. Since it may be misleading, however, to measure VAT’s revenue performance on the basis of production rather than consumption, VAT productivity is perhaps better expressed as the ratio of VAT revenues as a percentage of consumption (private) divided by the standard rate. On this basis, in 2005, Bangladesh’s VAT collected 0.21 percent of consumption for every 1 percentage point of the tax rate. These results are much lower than several neighboring countries. 81. To improve the revenue performance and efficient functioning of Bangladesh’s VAT system, a thorough reform of the design and workings of the VAT (and excise duties) is urgently needed. The principle objective would be to extend the VAT to the retail stage and apply it universally with an exemption that applies only to small businesses. This would require a thorough review of the VAT legislation, and the redrafting of the VAT Act, the tightening of the exemptions that are currently provided for by the regulations, a different application of the turnover tax for small businesses, and a review of the excise duties on the basis of efficiency and revenue mobilization objectives. 1.5.2 Income Tax: The Need to Start Again 82. Despite being based on the original model legislative framework of 1984, income tax in Bangladesh, consisting of corporate and personal income taxes, operates in a rather distorted, inconsistent and inefficient manner. The original legislation, which is still in force today, incorporates a number of modern income tax concepts. However, countless amendments, deletions, and additions over the last two decades, have changed the coherence and structure that the legislation might have had in 1984. Moreover, the amendments are scattered in different acts, regulations and special rules and orders, especially those that relate to various tax incentives. Redrafting the law therefore, to make it understandable, simple and coherent is a major priority. 83. Income tax collection seems to be heavily dependent on an extensive system of withholding on income and other payments. This withholding system partly explains why NBR has been able to collect as much income tax revenue as it does. Income taxes account for around 20 percent of total tax revenues. This share is expected to be higher in FY08, due to improved tax collection under the caretaker government. However, the increased use of withholding threatens to turn a major part of the income tax into simply a series of excise taxes on readily identifiable financial payments. In most cases, 26 VAT productivity on the basis of production is calculated as the ratio of the VAT revenues to GDP divided by the standard VAT rate. 31 the application of the withholding tax does not have any relationship to the taxpayer’s income tax liability. This practice greatly distorts the concept of income taxation, whose legitimacy is based on the principle that individuals who enjoy the same amount of income should pay the same amount of tax. 84. A serious omission from the current legislative framework that is likely to lead to a great deal of revenue loss is the lack of detailed tax accounting rules, which describe how businesses must account for their profits for tax purposes. The income tax legislation ought to contain detailed accounting rules that prevent assessors from reporting their profits using Generally Accepted Accounting Standards (GAAS) that defer profits to future years. Furthermore, there is a need for the legislation to define employment more accurately for tax liability purposes and include all incomes, whether in cash or in kind, for calculating tax liability. Finally, Bangladesh should reduce tax incentives delivered through the tax system. They are simply government spending programs that are legislated in the income tax ordinance, and are delivered by allowing assessors to offset the implicit subsidy the tax incentive provides against their tax liability. 1.5.3 Modernizing Revenue Administration 85. Bangladesh’s revenue administration performance suggests that the country is significantly underperforming its revenue administration potential. This has become apparent by looking at several objective indicators such as trade logistics costs, port clearance times, and VAT productivity indexes shown above (paragraph 71). According to the most recent business survey, revenue administration performance is perceived by the business community as one of the most important constraints to investment and growth in Bangladesh. Comprehensive tax administration reforms have become critical. The main objective of any improvement in revenue administration must focus on improving voluntary compliance by taxpayers. The NBR’s Strategic Development Plan (SDP) embraces this objective and incorporates many elements that will move NBR towards a modern tax administration system. 86. Four key pillars should be part of any NBR effort to strengthen governance, simplify procedures and modernize the revenue administration system: • Organizational improvements. To improve productivity and taxpayer compliance, NBR needs to move towards a functional design within its tax-by-tax structure, with the consequent sharing of some central common functions across taxes such as legal, human resource management and ICT. This means moving into an organizational structure that is based on various administrative functions that are carried out for each type of tax such as collection, enforcement, audit, taxpayer services and the like. This enables specialization, better sharing of information between customs/VAT and Income Tax, and economies of scale for a more efficient tax administration service. Core management functions also need to be strengthened by reinforcing the composition of the NBR Board structure and allowing certain advisory support services to be made available, including the creation of specialized functional units such as audit, human resources, legal and ICT. • Automation with business process changes. With the exception of customs administration, where ICT was introduced in 1990, there is minimal automation of tax collection or recordkeeping in the VAT and Income Tax wings of NBR. ICT can play an important role in modernizing revenue administration. However, the application of ICT in an environment where business processes are not simplified and transformed to address the main objectives of tax administration reform will not be able to produce tangible results in terms of improving revenue performance. The reengineering of business processes, however, will need to go hand-in-hand with the functional organization of the agency. 32 • Flexible human resource management. NBR practices of managing human resources suffer from the same issues as the overall civil service management. There are major issues with cadre rivalry among various tax divisions, a lack of integrity, low pay and other inefficiencies that result from a centralized and inflexible human resource management system. It is crucial that any tax administration reform is combined with serious transformation of human resources practices that allow the agency to build a more flexible policy to motivate staff, improve productivity and strengthen professionalism and integrity. • Developing a culture of taxpayer services. There is little or no history of a taxpayer service culture in Bangladesh’s tax administration. Although there is a tradition of NBR consulting with business leaders prior to budget formulation, primarily on tax policy issues (tax and tariff change), it is important to build a channel of consultations with taxpayers and other stakeholders throughout the year to keep them updated on significant changes in policies and procedures. The creation of web-based taxpayer information services is long overdue, as are steps to develop electronic means of communication between taxpayers and NBR, which could include enabling the electronic submission of returns, introducing a facility for electronic payments, and direct trade inputs for importers. 1.6 Key Conclusions and Recommendations 87. Bangladesh has shown a commendable record in maintaining fiscal discipline. Prudent fiscal policies have kept the fiscal deficit and debt indicators under control, despite numerous pressures placed on scarce public resources due to recurring natural disasters and rising global prices on food, fuel and fertilizer. The central government’s fiscal stance is sustainable at the current and projected levels of the fiscal deficit. The consolidated public sector deficit, however, which includes SoEs and state-owned commercial banks, appears to be questionable. Fiscal sustainability is highly dependent on the performance of economic growth and concessional foreign financing, which helps to keep interest rates low. Maintaining current rates of economic growth and mobilizing concessional foreign financing remain critical in avoiding any major fiscal adjustments to maintain sustainability. 88. The quality of fiscal performance has been weak and a number of vulnerabilities are threatening prospects for fiscal sustainability. • Revenue mobilization has remained poor, with a narrow tax base and heavy dependence on trade taxes, although FY08 has recorded encouraging improvements whose sustainability remains to be seen. Incremental measures to improve tax collection have been introduced over the last few years, including the creation of two Large Taxpayer Units (LTU) for VAT and income tax. The appropriate response to the situation, however, would require a major overhaul of tax policy and administration. Introducing simple, coherent and effective tax policy provisions must take priority, especially for domestic taxes such as VAT and income tax. These taxes are likely to be the future mainstay of Bangladesh’s tax system as the need to reduce dependence on international trade tax approaches. • Fiscal adjustment has been driven mainly by a decline in public investment spending; development expenditures, after peaking at 6.6 percent of GDP in 2001, have declined to 3.4 percent of GDP in FY08 due to low absorption and weak implementation capacity. Significant structural rigidities continue to hamper the implementation of the development budget (investment program). Measures to improve the implementation of the ADP are badly needed as the government embarks on a comprehensive program of PFM reforms. 33 • The recent surge in global food, fertilizer and oil prices has resulted in increased subsidies to cover financial losses of SoEs that are importing and selling these products at lower prices in the domestic market. Total losses from underpricing of fertilizer, diesel, power and natural gas were around $1.8 billion, or 2.5 percent of GDP in FY08. The recent decline in international prices–particularly of oil and fertilizer–has somewhat eased the pressure on SoE finances, but their sustainability remains to be seen as the government also lowered prices for these products in the domestic market. Maintaining a close watch on subsidies and establishing a system for accounting and managing them transparently–to contain their fiscal impact when global conditions are not favorable–will be important for maintaining fiscal sustainability. 89. Improving revenue mobilization, however, must be the cornerstone of the future fiscal policy of the government, as this will provide a buffer to adverse fiscal conditions. At 10-11 percent of GDP, Bangladesh has one of the lowest tax-to-GDP ratios in the world. The gradual reforms that have been attempted over the last 6-7 years have not produced the expected results and a comprehensive strategy that envisages an overhaul of the current system is needed. There is a strong case for increasing the domestic tax-to-GDP ratio through broadening the tax base and improving revenue management, in particular VAT and income tax. A thorough review of the VAT legislation is required as is a redrafting of the act to tighten exemptions and expand the application of the VAT to the retail stage. Likewise, redrafting of the income tax legislation to make it more simple and coherent is an urgent priority. Finally, improving revenue administration to increase voluntary compliance by taxpayers should accompany tax policy reforms. 34 CHAPTER TWO ALLOCATING RESOURCES TO BROAD STRATEGIC PRIORITIES: WHAT IS THE BUDGET FINANCING? Governments of different political persuasions have allocated public resources broadly in accordance with the country’s development priorities of improving social services and advancing rural development. Social sector spending has been maintained, rural infrastructure has been increased and support for targeted poverty reduction programs has been intensified. While broad inter-sectoral allocations of resources reflect strategic priorities, the country has been consuming more than it has been investing at a time of much-needed investments in infrastructure to support economic growth. Investments on national infrastructure services have declined, especially in power and transport, while transfers to loss-making state-owned energy enterprises have absorbed a significant share. Interest payments on domestic debt are crowding out much-needed expenditure on operations and maintenance. Structural rigidities and implementation problems continue to hamper effective implementation of the investment program. These developments do not point to the need for major re-alignments in inter-sectoral allocations, nor can they be resolved by simple re-allocations of resources. Key policy priorities: • Rationalize spending on subsidies and equity investments in loss-making SoEs, especially in the energy sector; • Increase spending for operations and maintenance, especially on transport, and for financing public goods on agriculture, especially for research and technology; and • Fundamentally improve the management of the ADP to increase the implementation rate of investment projects. 2.1 Government’s Strategic Priorities 90. In addition to achieving good fiscal outcomes and maintaining fiscal discipline, budget policies and institutions need to ensure that resources are allocated to the government’s strategic priorities. These should be delivered in accordance with the appropriate role of the public sector in delivering public services. Bangladesh’s development vision presented in the National Strategy for Poverty Reduction (NSAPR: the latest PRSP) puts strong emphasis on fostering an economic environment where everyone can benefit from the growth process, through employment generation activities, infrastructure development, and better health and education services, especially in the rural areas where 80 percent of the population live. 91. In the context of serious pressures put on public expenditures every year, the composition of spending has reflected the broad strategic priorities stated by the government. Spending in social sectors has been sustained, financing for rural development and especially, rural infrastructure has increased and support for targeted poverty reduction programs, including social safety nets for the poor, has continued (Figure 2-1 and 2-2). 35 Figure 2-1: Functional Composition of Total Figure 2-2: Functional Composition of Total Expenditure (% of GDP) Expenditure (% of Total Expenditure) 4.0 30.0 3.5 25.0 3.0 20.0 2.5 2.0 15.0 1.5 10.0 1.0 5.0 0.5 0.0 0.0 General Social Services Economic Infrastruct ure Interest General Social Services Economic Inf rast ruct ure Int erest Administration Services Services Administ rat ion Services Services FY98-FY02 Average FY03-FY07 Average FY98-FY02 Average FY03-FY07 Average Source: Ministry of Finance and WB staff estimates. 92. The country, however, has been consuming more than it has been investing at a time when investments in infrastructure, especially power and ports, are very much needed to boost economic growth. Recurrent expenditures as a share of GDP have increased significantly, while capital spending has declined (Figure 2-3). The revenue budget (recurrent expenditures), which has been rising in the last seven years, accounts for 60 percent of total spending, or 9.1 percent of GDP. Major increases have occurred in subsidies and transfers (30.5 percent in FY07) and some increase has taken place on interest payments (13.4 percent in FY07). This might have crowded out development expenditures, although they are constantly underspent. At the same time, the civil service wage bill has been stable over the years, accounting for about 28 percent of total expenditure in FY07. Figure 2-3: Expenditure Trends – Recurrent and Development Expenditure 16.0 14.8 Total Expenditure 14.0 13.8 13.9 13.7 13.6 13.2 12.0 10.0 Recurrent Expenditure D tGP 9.1 n 8.0 e 8.4 8.4 c 8.1 er 8.0 P 7.7 6.0 5.6 Annual Development Program 5.4 5.0 5.0 4.7 4.0 3.9 2.0 0.0 2002 2003 2004 2005 2006 2007 Source: Ministry of Finance, Bangladesh. 93. Interest expenditure remains a significant share of total expenditure, although as a share of national income, at 2 percent of GDP, it is well below other countries in the region. After their recent increase they account for an average of 13.4 percent of total expenditure, almost equal to education and double the resources going to the health sector. The change in the debt composition between domestic and foreign debt with a marked decline in foreign financing of the budget, has contributed to 36 the increase in interest expenditures. Their significant increase in FY03-07, by an average of 26 percent compared to FY98-02, has reduced fiscal space and likely crowded out expenditures on operations and maintenance in infrastructure and other priority sectors (Table 2-1, Figures 2-4 and 2.5). Table 2-1: Interest Payments: Bangladesh and Countries in South Asia (percent of GDP) Country 2000 2001 2002 2003 2004 2005 2006 Bangladesh 1.5 1.5 1.7 1.9 1.7 1.8 1.9 India 4.7 4.6 4.7 4.4 4.0 3.5 3.3 Pakistan 6.2 7.4 7.2 5.3 5.6 4.2 4.8 Sri Lanka 5.7 6.7 7.4 7.1 5.9 5.1 5.4 (percent of total expenditure) Bangladesh 11 10 12 14 13 13 14 India 30 29 29 28 26 23 22 Pakistan 36 46 42 31 39 29 31 Sri Lanka 25 28 32 34 28 24 24 Source: Bangladesh Bureau of Statistics, Bangladesh Bank, WDI. Figure 2-4: Interest Payments: Bangladesh and Figure 2-5: Interest Payments: Bangladesh and Countries in South Asia (% of GDP) Countries in South Asia (% of Total Expenditure) 8 50 40 6 30 4 6 20 2 10 0 0 B angladesh India P akistan Sri Lanka B angladesh India P akistan Sri Lanka 2000 2001 2002 2003 2004 2005 2006 2000 2001 2002 2003 2004 2005 2006 Source: Bangladesh Bureau of Statistics, Bangladesh Bank, WDI. 94. Development expenditure has been squeezed as a result of rising pressures on the recurrent budget, declining from 5.6 percent of GDP in 2002 to 3.9 percent in 2007. Consistent underspending has occurred in development expenditures in the range of 20 percent annually, leading to decreases in actual development spending. Slow disbursement of development funds is a result of weak downstream project management and procurement bottlenecks that have led to implementation delays, in particular in large-scale investment project expenditures. 37 2.1.1 Social Sectors and Rural Development Remain in the Lead 95. Governments of different political persuasions have maintained the high priority given to public social services in Bangladesh. During the past decade, the emphasis of public spending has shifted from agriculture and industries to the development of human resources and the rural economy. Recent budgetary allocations (FY03-07) have sustained these priorities, thus contributing to good social outcomes for which Bangladesh is becoming well-known in the developing world. By spending, on average, around one-quarter of its total government budget during FY03-FY07 on education, health, social security and welfare combined Bangladesh compares favorably to other countries with the same income level. Government spending in these sectors as a share of national income, however, remains small as a result of the low overall level of public spending.27 Within the social sector, education spending has taken the largest share at around 14 percent of total expenditure, while health accounted for 6.5 percent of the total budget. Although expenditures in these sectors have remained constant in relative terms, they have experienced a real increase of almost 50 percent since 2000 due to the increase in national income (GDP). (Figure 2-6). Figure 2-6: Composition of Social Expenditure (FY00-FY07) 30.0% Education Health Social Security & Welfare 24.5 25.1 25.0% 23.7 23.3 22.8 22.6 22.4 3.9% e s 6.7% 4.7% 3.4% 20.8 3.5% r 3.2% u 3.9% t 20.0% endi 4.2% p 6.7% lEx 5.9% 5.3% 4.7% 5.3% a 4.2% 5.7% ot 15.0% fT 4.9% c entageo 10.0% Per 13.7% 14.0% 14.0% 14.6% 14.5% 13.6% 12.9% 11.7% 5.0% 0.0% FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07 (R) Source: Ministry of Finance, Bangladesh. 96. The government is focusing its efforts on creating growth and income in the rural economy. Expenditures on economic services comprising agriculture, mining and rural development have increased to around 18 percent of total spending with the largest share going to rural development, which received, on average, 9.3 percent of total expenditure during FY03-07. Continued expansion of the rural roads’ network contributed to this larger share as investments by the government increased substantially.28 These increases have effectively exploited interlinkages within and between sectors. Support for rural infrastructure (rural roads and rural electricity) is reinforcing health and education results by contributing to an improvement in access to a school or health clinic while at the same time, increasing rural nonfarm incomes needed to pay for health and education services. 27 These sectors benefit from other sources of expenditure. The health sector,, and secondary education, for instance finance a large part of the expenditure that takes place in the sector from household incomes. 28 Studies have found that development of rural roads led to the improvement of service delivery in the project areas: access to schools increased by 7 percent, banks by 33 percent and health clinics by 60 percent. After the improvement of roads, passenger transport charges were reduced on average by about 50 percent, and freight passenger charges were reduced by 35 percent. 38 2.1.2 Infrastructure Spending Overall Has Declined 97. While spending in social sectors has been sustained and has increased for rural development, total spending on national infrastructure has declined, especially in the transport sector, which is crucial for accelerating economic growth.29 Infrastructure spending declined by 8 percent, from an average of 21 percent in FY98-02 to 19.3 percent of the total budget in FY03-07. The hardest-hit sector has been transport (excluding rural roads) while total expenditures on fuel and energy have increased, mainly reflecting increases in transfers to SoEs to cover their increased operational losses arising from increased global oil prices (Figure 2-7). At the same time, the capital-intensive infrastructure sector suffers more markedly than other sectors from implementation issues and low absorptive capacity. Figure 2-7: Composition of Infrastructure Expenditure (FY00-FY07) 40.0% Ministry of Science & Technology Fuel and Energy Transport and Communication Housing and Community Services 35.0% 33.4 32.7 32.3 31.3 30.0% 7.8% s 7.7% re 9.3% 26.7 itu 8.3% 25.8 d 1.1% n 25.0% 1.1% pe 22.6 lEx 1.0% ta 10.3% o 20.0% fT 11.9% 13.3% 11.3% eo 8.7% 17.0 12.4% 0.8% g 13.5% ta 15.0% rcen 8.0% Pe 7.6% 10.0% 6.6% 6.7% 6.4% 6.2% 5.3% 4.7% 4.1% 5.0% 6.6% 7.6% 6.7% 5.3% 6.2% 6.4% 4.7% 4.1% 0.0% FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07 (R) Source: Ministry of Finance and WB staff estimates. Note: Since FY04 , Local Government Division and Rural Development and Co-operatives have been merged. The Local Govt. Division was previously under Housing and Community Services 2.1.3 General Administration Spending Remains Significant 98. General administration (GA), including the civil service, defense and public order, continues to absorb significant budgetary resources. Bangladesh spends as much on general administration as it spends on the social sector. GA attracted, on average, a similar one-quarter of the total budget expenditures during FY03-07, although military expenditures have been at low levels compared to neighboring countries, thus creating fiscal space for other important priorities. While expenditures on the civil service have not been excessive by international standards (at an average of 1.3 percent of GDP in FY03-07), they have not produced the expected outcomes; showing major signs of inefficiency with over-employment at Class III and IV levels (the lower levels of the civil service) and outdated public management practices (Figure 2-8). 29 Total spending in Infrastructure does not include expenditure on rural roads. They are classified as rural development in the government’s budget classification structure. 39 Figure 2-8: General Administration and Security Spending (FY00-FY07) 25.0% 23.2 General Public Services Defence Public Order and Safety 22.1 22.5 22.0 21.4 21.3 21.6 21.1 5.0% 20.0% 4.4% 4.5% 4.2% 5.2% 4.7% 5.9% 5.1% pn edt iue r s 15.0% x lE 9.3% 8.1% a 8.2% 8.9% 7.4% o t 7.9% fT 7.8% 7.3% ae g o 10.0% e Prcn et 5.0% 9.8% 9.3% 9.0% 9.0% 8.8% 8.7% 8.4% 8.2% 0.0% FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07 (R) Source: Ministry of Finance, Bangladesh. 2.2 The Annual Development Program – Investment Policy and Performance 2.2.1 Priorities of the Investment Program 99. The government’s priorities are better observed in the more discretionary component of the budget, the ADP. The ADP accounts for about 40 percent of the total expenditure or 5 percent of national income. Large parts of the primary recurrent budget, including salaries and establishment costs, are nondiscretionary and do not lend themselves to reallocations, at least in the short term. The overall composition of these expenditures is thus fairly stable over time. Reallocations are driven at margin, mainly in the discretionary component of the budget, the ADP. 100. In line with the described focus on rural development, the allocation and use of ADP resources shows a considerable shift in the sectoral composition in favor of rural development and the energy sectors in FY01-07. The Local Government Division (LGD), which administers over 85 percent of investments in rural infrastructure, increased its share of development expenditure from 19.2 percent in 2001 to 27.3 percent in 2007. While part of the increase is due to an accounting change in FY04, part of the upward trend is due to refocused priorities that emphasize rural development. By contrast, the share of the social sectors remained relatively flat while the transport sector, which has consistently accounted for one of the top two shares of development expenditure, witnessed a steady decline. Development expenditures of the Ministry of Communications (which manages most of the investments in transport–except rural roads) have declined by an average of 1 percent annually. Similarly the agriculture sector, particularly the Ministry of Water Resources, has witnessed a steady decline in development expenditure (Figure 2-9). 40 Figure 2-9: Development Expenditure Shares (FY00-FY07) 100% 9.4% 10.4% 7.7% 8.0% 8.1% 9.5% 12.3% 90% 19.1% 15.9% 80% 22.6% 17.8% 22.2% 26.3% 26.4% 8.2% 70% 8.3% 7.8% 8.1% 60% 12.0% 11.2% 8.6% 19.5% 15.0% 22.6% 50% 21.0% 13.2% 12.5% 16.6% 8.3% 9.2% 40% 7.7% 7.2% 7.9% 5.9% 6.1% 13.9% 13.4% 30% 12.5% 10.8% 13.8% 13.8% 14.7% 20% 25.7% 27.7% 10% 22.0% 23.7% 19.2% 19.0% 17.2% 0% FY 01 FY 02 FY 03 FY 04 FY 05 FY 06 FY 07 Local Government & Rural Development Education & Technology Health Fuel & Energy Agriculture Transport & Communication Other Source: Ministry of Finance, Bangladesh. 2.2.2 Bangladesh’s Two Parallel Budgets 101. Despite its efforts to better integrate its budgeting practices, Bangladesh remains a country with two parallel budgets, as revenue and development budgets both contain elements of recurrent and investment spending. The main distinction between the two is that the development budget remains more discretionary as ADP is subject to less strict allocation and financing rules compared to the revenue budget. Further disaggregation of development expenditures (that is by economic classification) reveals that about 84 percent of the development budget financed capital expenditures (acquisition of assets, acquisition/purchase of land, construction and works, investments in shares and equities), while 16 percent financed recurrent expenditures (pay of officers, allowances, supplies and services) during FY01-07. (Figure 2-10). About 95 percent of capital expenditures were on construction and works and capital block allocation, and about 93 percent of the recurrent expenditures were on supplies and services, grants in aid, and block allocations. 41 Figure 2-10: Composition of Development Expenditure (by Economic Classification) 50 60 FY01-07 Average FY01-07 16% 45 FY01-07 50 40 35 40 30 25 30 20 15 20 10 10 84% 5 0 0 S&S Grants Revenue Expenditure Capital Expenditure Bloc All O Rev Exp C&W Cap Bloc O Cap Exp Key: S & S - Supplies and Services; Grants – Grants in Aid; Bloc All – Block Allocation; O Rev Exp – Other Revenue Expenditures (which is a grouping of all revenue expenditures minus S & S, Grants, and Bloc All); C & W – Construction and Works; Cap Bloc – Capital Bloc Allocation; and O Cap Exp – Other Capital Expenditures (which is a grouping of capital expenditures minus C & W and Cap Bloc). Source: Ministry of Finance and WB staff estimates. 102. There are significant differences across line ministries/divisions in the composition of development expenditures. Capital expenditures tend to account for a significant share in sectors engaged in infrastructure activities, while recurrent expenditures dominate in the social sectors. As Table 2-2 illustrates, the composition of expenditures of the four largest spenders identified earlier– Local Government Division, Energy Division, Ministry of Primary and Mass Education, and Ministry of Health and Family Welfare–underscore this trend. Table 2-2: Composition of Development Expenditure by Line Ministries/Divisions (by Economic Classification) Recurrent Expenditure Capital Expenditure 1. (FY01-07: Average) (FY01-07: Average) 2. 3. Local Government Division 3.93% 96.07% Supplies and Services: 21.69% Construction and works: 51.72% big spending items 4. Repairs, maintenance and Rehabilitation: 9.57% Capital Grants: 11.25% 5. Block Allocations: 40.85% Capital Block Allocation & Misc. Capital Expenditure: 33.41% 6. 7. Energy 0.17% 99.83% Pay of officers: 18.43% Construction and works: 19.56% 8. Supplies and services: 24.38% Investments in Shares and Equities: 24.81% 9. Repairs, maintenance and rehabilitation: 5.70% Loans: 20.93% 10. Capital Block Allocation & Misc. Capital Expenditure: 30.37% 11.Health and Family Welfare 54.61% 45.39% 12. Supplies and services: 54.74% Acquisition of assets: 17.64% 13. Grants in aid: 15.23% Construction and works: 48.37% Block Allocations: 13.73% 14. 15.Primary & Mass Education 55.61% 44.39% Supplies and services: 11.91% Construction and works: 50.25% Grants in aid: 68.15% Capital Block Allocation & Misc. Capital Expenditure: 38.69% Block Allocations: 17.64% Source: Ministry of Finance and WB staff calculations. 103. A noteworthy trend that emerges from such disaggregation is the significance of block allocations in the composition of development spending. Block allocations are general provisions that 42 are not tied to specific line items, and spending ministries/divisions have considerable discretionary power over their use. While these allocations undoubtedly increase the flexibility of line ministries/divisions to allocate and/or reallocate funds to areas they deem a priority, their extensive use without a clear strategy reduces the transparency and predictability of the budget process. Block allocations have been extensively used to fund a category of projects known as “unapproved projects”30 that proliferate in the ADP portfolio, reducing the much-needed discipline on the entry of new projects. 2.2.3 Performance of the ADP Portfolio 104. The quality and effectiveness of the ADP portfolio that averages more than one thousand projects/programs annually, has been of great and long-standing concern to policy makers. Incremental budgeting, particularly the practice of adding new activities while continuing with all existing projects/programs, has tended to favor stretching scarce resources to cover new activities at the expense of inadequate maintenance of existing ones. Time and cost overruns, deferral of project benefits long into the future, and declining returns on investment have historically been associated with the management of the development budget in Bangladesh. 105. A review of the ADP reveals that between FY03 and FY07, the development budget allocated for implementing projects/programs in the ADP portfolio averaged around 13 percent of their total cost. For instance, the FY07 ADP allocation of Tk. 260 billion was originally programmed to support implementation of 852 projects/programs (715 investment projects/programs and 137 technical assistance projects/programs) costing Tk. 1,968 billion. By mid-year, only 33.1 percent of investment projects/programs and 11 percent of technical assistance projects had been implemented. This implies that it takes an average of 7.1 and 6.6 years to fully implement the remaining investment and technical assistance projects/programs respectively, assuming the FY07 funding levels and the unlikely prospect of no new projects/programs entering the portfolio (Table 2-3). The FY07 ADP allocation was later revised downwards to Tk. 216 billion, while the number of projects/programs was revised upwards to 1,065 (885 investment and 180 technical assistance projects/programs), illustrating the tendency to favor the addition of new projects/programs without corresponding adjustment to resources. 106. The continuing suboptimal allocation of resources suggests that despite small improvements, the implementation delays will persist, resulting in a reduction in the overall effectiveness of the ADP. A review of the investment projects/programs in the ADP (excluding technical assistance projects/programs) for FY03-07 reveals that, on average, only about one-third (34 percent) of their cost was covered by the development budget. Consequently, the ADP each year has carried over a significant backlog of incomplete projects/programs. New projects/programs entering the portfolio grew by 2.7 percent per year, while the development budget grew by 3.5 percent. At this trend, project completion rates are going to be very low for the foreseeable future. 30 Unapproved projects are investment projects that have not gone through the government’s internal screening and evaluation process, but are wait-listed in the government’s ADP. They are usually approved and financed during the fiscal year. 43 Table 2-3: ADP: Resource Allocation vs. Cost of ADP Projects/Programs (FY07) Implementation Implementation Cumm. Expend. ADP Allocation Implied Project Upto Dec. 2005 Total Project Completion # of Invest. Number of Complete Portfolio for FY07 Achieve Years to Years to projects Current Period Sector Cost Taka in Millions % Taka in Millions Agriculture 96 56,404 10,748 19.05 10,303 4.43 1.04 5.47 Rural Development & Institutions 52 190,798 65,360 34.26 29,528 4.25 2.21 6.46 Water Resources 41 48,893 11,420 23.36 7,591 4.94 1.50 6.44 Industries 21 19,403 13,157 67.81 2,745 2.28 4.79 7.07 Power 44 203,578 74,877 36.78 29,324 4.39 2.55 6.94 Oil, Gas & Natural Resources 16 34,469 3,632 10.54 3,710 8.31 0.98 9.29 Transport 171 283,695 125,498 44.24 30,016 5.27 4.18 9.45 Communication 20 85,015 44,398 52.22 7,889 5.15 5.63 10.78 Physical Planning, Water Supply & Housing 94 98,815 41,060 41.55 12,016 4.81 3.42 8.22 Education & Religious Affairs 63 169,671 40,232 23.71 33,109 3.91 1.22 5.12 Sports & Culture 20 5,117 601 11.74 1,415 3.19 0.42 3.62 Health, Nutrition, Population & Family Welfare 15 179,262 23,251 12.97 24,439 6.38 0.95 7.34 Mass Media 5 1,644 99 6.01 750 2.06 0.13 2.19 Welfare, Women Affairs & Youth Development 25 5,746 2,761 48.05 1,228 2.43 2.25 4.68 Public Administration 8 1,115 125 11.25 481 2.06 0.26 2.32 ce, Information & Communication Technology 16 3,425 1,046 30.53 1,001 2.38 1.04 3.42 Labour & Employment 8 4,344 1,574 36.23 733 3.78 2.15 5.93 Total ADP Investment Portfolio 715 1,391,392 459,838 33.05 196,278 4.75 2.34 7.09 Source: Ministry of Finance and WB staff estimates. 2.2.4 Budget Deviation Indexes 107. The poor planning and resource allocation process has led to a persistent pattern of underspent ADP budget allocations. The unpredictability of resource flows has increased over the last few years as shown by the budget deviation indexes, which measure the deviation between the allocated budget and the actual spending. A budget deviation index can provide a measure of policy predictability and credibility for the budget. Expenditure deviations can occur as a result of reallocations due to shifts in fiscal aggregates or reallocations of sectoral resources. Reallocations that occur as a result of shifts in the fiscal aggregate are usually justified by the need to prioritize or maintain targeted levels of fiscal deficit. Bangladesh’s record in maintaining aggregate fiscal discipline has been good: the budget deficit was kept under 4 percent of GDP last fiscal year and its fiscal indicators have remained within the threshold of the macroeconomic framework agreed by the IMF’s Poverty Reduction and Growth Facility (PRGF). 108. The deviations seem to suggest lessened credibility of the budget and volatility in policy implementation, which could be a result of a highly unpredictable environment, poor quality of planning and/or implementation. Table 2-4 shows that the budget deviation in Bangladesh decreased on average between FY00 and FY03, but it has increased sharply since then. Policy variability appears to have been significant during FY05-06 and the unpredictability of the ADP budget more pronounced during those years. This trend continued in FY07 and FY08. 44 Table 2-4: Bangladesh ADP Budget Deviation Index31 FY00 FY01 FY02 FY03 FY04 FY05 FY06 ADP Budget Deviation Index (%) 26.22 16.54 23.02 18.93 18.85 21.60 30.16 Source: WB staff estimates. 109. A highly bunched pattern of reported ADP expenditure, particularly in the last quarter of the fiscal year, has also raised questions about the credibility of expenditure utilization and program/project implementation. ADP spending in the fourth quarter of the fiscal year has tended to be large in both absolute terms and as a share of the ADP. According to Bank staff estimates, around 60 percent of the annual ADP spending took place during the first nine months of the fiscal year, while a substantial 40 percent was spent in the last quarter of the fiscal year. This is significant even if one allows for seasonality and some normal bunching of project activities towards the second half of the fiscal year. In FY07, development expenditure stood at 35.3 percent as of the end of the third quarter (end April), increased to 39.1 percent by end May, and to 78 percent by end June.32 110. The spikes in spending in the fourth quarter can be linked to two factors; uncertainty and funding delays and the slow utilization of funds early in the fiscal year. The first factor relates to uncertainty and delays in third and fourth quarter fund releases. In many instances, funds have been released in the last week of June (or even retroactively in July), which appears to be out of line with financial norms. The second factor is the slow utilization of funds during the first two quarters of the fiscal year, which is linked to delays in the initiation of activities and the release of funds, and inadequate project planning and enforcement. A major cause for delays in initiating projects/programs has been the fact that line ministries/divisions initiate procurement activities only after the budget has been approved and released. New rules are now in place requiring line ministries/divisions to make submissions of their annual budgets along with their procurement plans that include specification of goods and works to be procured during the budget cycle. 111. The level of fourth quarter spending, however, raises questions about the credibility of official estimates of ADP expenditure as a true measure of realized implementation. This casts additional doubt on the efficiency of ADP spending when so much money is required to be spent in the last three months of the fiscal year. Anecdotal evidence suggests that much of the fourth quarter spending is reported on the basis of checks and vouchers issued, rather than actual implementation on the ground. The political incentive also encourages reporting large utilizations as ADP spending is seen in some quarters as a proxy for economic performance. Successive administrations have, therefore, implicitly or explicitly endorsed the practice and avoided instituting audits to investigate the reality of the reported ADP estimates. 2.3 Key Conclusions 112. Governments of different political persuasions in Bangladesh have maintained broad, intersectoral budgetary resource allocations in accordance with strategic policy priorities of strengthening social services and rural development. Spending in social sectors has been sustained, rural development and especially rural infrastructure has increased and support for targeted poverty reduction programs, including the social safety net for the poor has continued (Figure 2-2). At the same 31 The functional budget deviation index is calculated as the sum of absolute deviations by broad functional classifications as a percentage of total budgeted expenditure in respective years. 32 See Monthly Reports on Fiscal Position: April, May, June FY2006-07 Provisional (Finance Division, 2007); Overall Progress of ADP: 2006-07 (Implementation Monitoring and Evaluation Department, 2007). 45 time, total resources devoted to national infrastructure, in particular transport, have declined as have development expenditures for agriculture and water resources. In the meantime, equity investments from ADP in the loss-making energy SoEs have increased, while spending on general administration expenditures remains significant. 113. While broad intersectoral allocations are aligned with the government’s priorities, the government has been consuming more than it has been investing. The share of recurrent spending as a percentage of national income (9.1 percent in FY07) has been increasing. Subsidies and transfers to cover losses of SoEs hit by rising global food, fuel and fertilizer prices soared in FY07, although part of the increase came as a result of increased transparency in government’s reporting of subsidies. Prices then declined as global prices fell. Interest payments on domestic debt are crowding out much needed expenditure on operation and maintenance, especially on infrastructure. Structural rigidities and implementation problems that hamper the full implementation of the annual development program have caused investment spending to decline. 114. The performance of the ADP (investment program) portfolio remains a matter of great concern to policy makers and development partners as resources are spread thinly over thousands of projects. Time and cost overruns, deferral of project benefits long into the future, and declining returns on investment continue to plague the management of the development budget in Bangladesh. Budget deviation indexes, especially for the ADP, have been on the rise pointing to implementation and poor planning problems and, as a result, volatility in the policy implementation process. 115. These developments do not point to the need for major inter-sectoral reallocations; neither can they be resolved by simple resource allocations. However, there is a need to: (i) rationalize spending on subsidies and equity investments in loss-making SoEs, especially in the energy sector: (ii) increase spending on agriculture, especially for research and technology; and (iii) fundamentally improve the management of the ADP to improve the implementation rate of the investment program. Addressing these issues requires some structural reforms in these sectors and fundamental improvements in the management of resources and PFM systems at the sectoral/operational level. 116. Chapter Three discusses, in detail, sector expenditure policy and institutional issues in five key areas-- education, health, transport, energy and agriculture. It looks at how well equipped the sectors are to address the challenges they face in terms of the sector policy framework and their linkages with the expenditure management institutions. It also seeks to provide some recommendations for improving operational efficiency, an area where Bangladesh seems to have a poor record. 46 CHAPTER THREE ARE THE INTENDED OUTCOMES BEING ACHIEVED AT THE SECTOR LEVEL? OPERATIONAL EFFICIENCY Bangladesh’s very good record of maintaining fiscal discipline and sustaining the allocation of public resources to broad strategic priorities is not matched at the operational, sector/agency level, where its record is relatively poor.33 While the country is becoming well-known for having achieved good economic and social outcomes over the last two decades, this review of the expenditure policies and programs in five key sectors–education, health, transport, energy and agriculture–highlights a number of inefficiencies and structural issues which, if not addressed, will impede further progress on these outcomes. The public sector and public expenditure play an important role in sector outcomes through policy formulation, provision and/or financing. Sector policy frameworks are in place but public expenditure allocations do not always reflect sector priorities. Expenditure distribution within the sectors remains skewed towards the non- poor. Fragmented sector management structures, budget planning and management systems are hindering the effective and efficient implementation of sector policies. To address the issues, a number of sector-specific policy measures are outlined in this section. Improving the budget and financial management systems at the sector level, however, remains a key challenge for all. 3.1 Summary of Cross–cutting Challenges 117. The public sector and public expenditure play an important role in sector outcomes through policy formulation, provision and/or financing. 118. In the education sector, the government either directly provides or substantially supports– through full or partial financing–most education provisions. Education services are provided by a number of different providers. However, government schools at all levels are fully funded by public resources. At the secondary level, non-government schools and madrassas provide the largest share of costs with the government contributing teachers’ basic pay and tuition payments for female student participants in the stipends scheme (Table 3-1). 119. In health, the budget finances and maintains an extensive infrastructure of health facilities from primary to tertiary level, spread across the national and union levels. The government is the second largest financier of total health spending at 21 percent, after out-of–pocket expenditure (65 percent) that are spent mostly in private pharmacies.Transport services are regulated, provided and financed largely by the public sector through central government agencies and autonomous government departments. Private operators are limited to intercity road transport and the main routes of inland water transport. 120. Energy provision is dominated by a large number of state-owned enterprises, although there is some opening to private sector involvement. The budget finances most, if not all, capital investments in the energy sector and provides significant transfers in the form of explicit budgetary subsidies, while suffering major fiscal costs due to contingent liabilities from implicit transfers. Subsidy transfers for fertilizer, food, electricity and gas used in irrigation pumps are also present in the agriculture sector. 33 Operational efficiency has two dimensions: (i) minimizing cost per unit of output or achieving more value for the money spent, and: (ii) achieving the outcome for which the output is intended. This chapter deals more extensively with (ii), although makes references to (i) as appropriate. A full coverage of (i) is beyond the scope of this document. 47 While public resources contributing to agriculture GDP are not large at 12 percent, the government plays a significant role in the sector as it continues to intervene in the market for important agriculture inputs with a consequential impact on sector productivity and its fiscal position. Table 3-1: Government Financing Modalities for Education Enrollment % of Sector Level of Government Funding 2005 (000s) Total PRIMARY EDUCATION Government schools Fully government funded. 9,484 55 Registered non-government Government provides basic pay of teachers and limited allowances. 3,573 21 schools Provides free stipends to 40 percent of rural students and free textbooks to all students. Govt. Alia madrasaas Fully government funded. n/a Independent ebtadayee Teachers receive Tk. 750. Provides free stipends to 40 percent of rural 850 5 madrasahs students and free textbooks to all students. Recognized non-government Attached ebtadayee sections of higher madrasahs. Government 1,146 7 Alia madrasahs provides basic pay of teachers and limited allowances. Provides free textbooks. Unrecognized madrassas No government funding. n/a n/a Private schools No direct government funding. Free textbooks for schools following 404 2 national curriculum. NGO schools No direct government funding. Free textbooks for schools following 1,500 9 national curriculum. Community schools Teachers receive Tk. 750. Provides free stipends to 40 percent of rural 426 2 students and free textbooks to all students. SECONDARY AND HIGHER SECONDARY EDUCATION (CLASSES 6-12) Government schools Fully government funded. 387 4 Registered non-government Government provides basic pay of teachers and limited allowances. 7,812 79 schools Provides free tuition and a stipend to participating rural female students. Govt. Alia madrasaas Fully government funded. 1 0 Recognized non-government Government provides basic pay of teachers and limited allowances. 1,717 17 Alia madrasahs Provides free tuition and a stipend to participating rural female students. Unrecognized madrasaas No government funding. n/a n/a Private schools No government funding. n/a n/a NGO schools No government funding. n/a n/a TECHNICAL AND VOCATIONAL EDUCATION AND TRAINING Government institutes Fully government funded. 49 20 Non-government institutes Government provides basic pay of teachers and limited allowances 192 80 for most private institutes in this sector. TERTIARY EDUCATION DEGREE AND MASTERS EDUCATION Government colleges Fully government funded. 344 41 Government universities Government subvention. 116 14 Registered non-government Government provides basic pay of teachers and limited allowances. 223 27 colleges Govt. Alia madrasaas Fully government funded. 2 0 Recognized non-government Government provides basic pay of teachers and limited allowances. 55 7 Alia madraasas Private universities No government funding. 92 11 Source: NGO primary enrollment data is an estimate taken from World Bank (2006b). All other data for enrolment is for 2005 and taken from DPE (2006a) and BANBEIS (2006). Note: Private primary schools include non-registered non-government primary schools and kindergartens. 48 121. Sector policy frameworks are in place and appear to be appropriately designed, although public expenditure does not reflect the policies required to address sectoral issues and challenges. They present a reasonable set of policies to take the sector forward in order to address such challenges, and make an attempt to measure sector performance through various output or outcome indicators, in some cases as part of the overall strategy for accelerating economic growth and poverty reduction. However, public expenditure does not always reflect sector policies, so revealing a passive expenditure policy stance. Public expenditure policy in the transport sector continued to favor the extension of the road network, despite the sector’s policy requirement for a more balanced and integrated multimodal transport system. While education policy emphasizes improvements in the quality of education, overall spending in education remains low and insufficient to support the significant investments that are needed in the sector to improve quality. 122. In almost all cases, expenditure distribution needs to better address the issues of equity as they remain skewed towards the non-poor. In education, the poor represent 40 percent of the total population of school-aged children but receive only 32 percent of the total recurrent education expenditure. Even the primary stipend program, a conditional cash transfer program designed to target the poorest 40 percent of students, is only marginally pro-poor. In the health sector, the share of recurrent spending that goes to the 16 poorest districts is less than what goes to the 16 richest districts. Since energy consumption is progressive, with higher-income groups consuming more, general public subsidies going to the sector are, in many ways, largely inequitable and disproportionately benefit the more affluent. 123. Fragmented sector management structures and systems for budget planning and management are hindering timely and effective implementation of sector programs. The division of education sector policy and the implementation of educational programs between two different ministries–the Ministry of Primary and Mass Education, which is responsible for primary level education and the Ministry of Education, which is responsible for overall coordination between the two ministries- -will likely improve policy planning and implementation in the sector. 3.2 Education Sector: Policy, Financing and Expenditure Outcomes34 124. Increasing enrollment and achieving gender equity have been the hallmark of education policy in Bangladesh for the last two decades. More recently, prominence has been given to improving the quality of education. While overall enrollment rates have been sustained and even rose slightly during the 2000-2005 period, the composition of enrollment has shifted away from government- recognized (and funded) schools, which have suffered an 8 percent decline. While primary school enrollment has stagnated and tertiary enrollment declined, secondary school enrollment has exhibited the fastest growth, with a 10 percent increase (Figure 3-1). In contrast, government recognized madrassas appear to have registered substantial increases in enrolment.35 34 For a detailed analysis of sector expenditure analysis see Volume II of this report. 35 This increase is likely to be due both to better reporting as well as a shift in enrollment towards this type of school. 49 Figure 3-1: Enrollment Has Shifted Away From Government-Recognized (and Funded) Schools total other Ebtadayee madrasahs RNGPS GPS (1,500,000) (1,000,000) (500,000) - 500,000 1,000,000 1,500,000 male female total Source: (DPE 2002a; DPE 2006a). Note: GPS stands for Government Primary Schools, and RNGPS are Registered Non-Government Primary Schools 125. The aggressive pursuit of gender equity policies has shifted the gender gap in enrollment in favor of girls. The gap is greatest at the secondary level where female gross enrollment rates exceed those of their male counterparts by 8 percentage points. On the other hand, gender differences in enrollment rates in favor of boys at the higher secondary level have been contributing to the large gender gaps seen at the tertiary levels. While these gaps have been narrowing since 2000, there were approximately two male students enrolled for every female student in 2005. 126. While results in education access present a mixed picture, the education system is characterized by persistent low quality and inefficiency. Only half of all children beginning primary and secondary education survive up to the final grade, and at secondary level only one in five actually passed the SSC examination (Table 3.2). Failure to complete secondary education is a growing phenomenon. Between 2000 and 2005, the proportion of the 16-25 age group with an incomplete secondary education as their highest educational attainment increased from 23 to 33 percent (Al-Samarrai 2007a). These low completion rates are masked by an improvement in examination results at this level. In 2006, 60 percent and 76 percent of students, sitting for the SSC examination in general schools (government and nongovernment) and madrassas respectively, passed (BANBEIS 2007). Completion rates among the poor are much lower than for the non-poor, although the gap in primary completion rates has narrowed. Furthermore, as some earlier studies report, approximately one-third of individuals who had completed primary education were illiterate (CAMPE 2003). Learning outcomes for the poor are also lower than for the non-poor. These trends indicate that to improve sector outcomes and achieve the sector policy goals stated in the PRSP, significant investments will be needed to improve the quality of education, especially at the primary and junior secondary level. 50 Table 3-2: Internal Efficiency (%) of the Education System (2005) Repetition Rate Survival Rate Completion Rate Sector M F T M F T M F T Primary 12 11 11 49 57 53 - - - Secondary (general) 8 8 8 43 40 41 23 17 20 Secondary (madrassa) - - - - 57 61 - 14 22 Higher Secondary - - - - - - - 59 57 Source: (DPE 2006a; BANBEIS 2007). Notes: Primary statistics cover government, registered non-government and experimental schools. Statistics for madrassa education are for 2003. Figures for survival and completion rates are calculated using the reconstructed cohort method. 127. Education spending has retained its importance in the national budget and there have been substantial increases in real resources directed to the sector. Since the beginning of the decade approximately 15 percent of government resources have been devoted to education and this is comparable to developing and regional country averages.36 However, at 2.3 percent, the proportion of national income devoted to education remains low and this is largely due to the small share of government spending in national income. While the share of national income devoted to education was similar in FY 99-00 and in FY 06-07, real spending has increased by 50 percent owing to the high rates of economic growth. 128. There have been no significant shifts in the composition of spending, with the subsector shares of education remaining relatively stable across the period 2000-2007. Primary education has received relative importance with development spending shifting slightly away from the secondary sector, while non-development (recurrent) spending has maintained the relatively equal split between the primary and secondary sectors. The low proportion of national income spent on education implies, however, that Bangladesh spends much less on primary and secondary education compared to other countries in the region and outside. 129. Personnel costs take the largest share of recurrent spending, with 98 percent devoted to salaries over the last decade. This leaves little room for nonsalary related spending, although some of it is compensated by the development (investment) side of the budget financed by foreign aid (textbooks and stipends). Teacher salaries are the largest driver of differences in unit expenditure between different types of schools, showing that government school teachers are paid better and at comparable levels with regional counterparts, at 3-4 times their per capita GDP, while nongovernment teachers are paid substantially less (Table 3-3). 36 The average percentage of total public spending devoted to education in 2002 was on average 16 percent for developing countries as a whole and 14 percent for countries in South and West Asia (UNESCO 2006a). 51 Table 3-3: Composition of Total Public Education Spending (Various Years) (Constant 2006/7 Prices in millions of Taka) 2001-02 2003-04 2004-0537 Component Total % Total % Total % Primary Education Salary 18,007 57 18,456 64 19,357 67 Nonsalary 7,275 23 6,078 21 6,425 22 Capital 6,189 20 3,952 14 2,831 10 Other 43 0 255 1 265 1 Total 31,514 100 28,741 100 28,878 100 Secondary Education Salary 14,624 66 15,965 57 - - Nonsalary 4,412 20 4,245 15 - - Capital 3,193 14 7,049 25 - - Other 8 0 604 3 - - Total 22,237 100 27,863 100 - - Source: Ministry of Finance, Ministry of Primary and Mass Education, Ministry of Education. 130. Government education spending has remained heavily skewed in favor of the non-poor. The poor represent 40 percent of the total population of school-age children, but receive only 32 percent of the total recurrent education expenditure. Even the primary stipend program, a conditional cash transfer program designed to target the poorest 40 percent of students, is only marginally pro-poor. Primary education spending is more focused on the poor than the other levels of education. However, even primary education expenditure is not pro-poor. Fifty percent of the primary school-age population is classified as poor but the poor receive only 47 percent of public primary recurrent expenditure. This finding is similar to the incidence analysis conducted on the 2000 HIES data, which shows that since 2000, there has been no significant shift in education spending towards the poor.38 The gap becomes larger as students progress from primary to tertiary education. 131. The increase in real total spending, combined with a decline in enrollment, has increased spending per student in real terms, creating possibilities for improving quality. Spending per student, however, is still low by regional comparison. For example, in India, spending per student was approximately three times as high at the primary level in 2002. In addition, subsidies per student are different for different education providers, both at primary and secondary levels. Of the largest providers at the primary level, government and nongovernment registered madrassas tend to receive the highest per student subsidies from the government. However, registered nongovernment schools that enroll over one-fifth of all primary school students, receive less than half of the subsidy going to the other large providers. At the secondary level, both nongovernment schools and madrassas are similarly poorly funded. 132. If the NSAPR targets are to be achieved, substantial improvements and investments in the primary and junior sub-sector will be needed, as well as a stronger focus on balancing public spending between the poor and pro-poor more effectively. This is even more challenging at a time when the current emphasis in the sector has been on primary government and nongovernment schools where enrollment has been declining. 37 For 2004/05, actual primary education spending is low compared to budget figures because only 51 percent of the development budget was spent. This was due largely to the slow implementation of PEDP II at the start; budget execution for this project was only 34 percent in 2004/05. 38 In the previous study, public primary education expenditure is classified as being pro-poor because a higher percentage of this spending went to the poor compared to the percentage of poor individuals in the overall population. However, the paper also reports the percentage of primary school age children who are poor. This shows that 59 percent of the primary school age population is poor while only 56 percent of primary education expenditure accrues to this group (Glinskaya 2005). 52 133. Future sector expenditure strategies will require that attention be given to the following key issues: • improving management and accountability in the sector. Primary education requires a more holistic approach that focuses not just on government recognized schools. Bangladesh is an arena of co-existence of different education provision models, thus enabling a potential exchange and learning across the different types of schools. Greater coordination between the two education ministries would likely improve policy planning and performance of the sector. Furthermore, decentralizing financing directly to schools with stronger and more accountable school management committees could potentially lead to a more effective use of resources. • focusing more strongly on serving the needs of the poor. Better targeting and enforcement of stipend programs–which remains poorly targeted–is likely to have a larger impact on the poor’s access to schooling. In addition, more attention needs to be given to improving the learning outcomes of the poor, which largely rest on improving quality and ensuring that schools serving the disadvantaged areas have adequate resources and support. • providing more equitable levels of financing. If access to and quality of education is to be improved, increases in overall education spending are required. Investment in increased access, particularly at primary and secondary levels, needs to include demand-side interventions. This should be complemented with a review of the funding norms to ensure that they serve the government’s equity goals as well as in-service training for teachers. • providing improved incentives for nongovernment schools. Regulations providing incentives for better school performance to nongovernment schools and madrassas need to be improved and enforced in order to improve their incentive to provide good quality education. There is a need to review indicators used to monitor the performance of schools to ensure that they address the goals of education policy. 3.3 Health Sector: Policy, Financing and Expenditure Outcomes 134. Bangladesh’s health sector policy emphasizes its commitment to pro-poor health and the need to reduce inequalities in health service provision. The journey towards achieving these outcomes is on the right track with positive trends over the last decade. Infant and child mortality declined an impressive 46 and 57 percent respectively, outstripping progress in other countries in the region and setting Bangladesh on track to meet the MDG target of no more than 50 deaths per 1,000 by 2015. Fertility is once again on the decline, after being stabilized in the 1990s. Some progress has undoubtedly been achieved on maternal mortality, although the MDG target will be hard to reach at the current rate of change (Table 3-4). 135. Despite these good outcomes, there is much scope for improvement as inequities in service utilization remain a major concern and a substantial equity divide exists among service recipients. Among the wealthy, almost nine out of 10 children under the age of two are fully vaccinated compared with only six in 10 children in the lowest income quintile. Pregnant women in the top quintile are almost four times more likely to have received antenatal care from a medically trained person than those in the bottom quintile. One in every five women in the richest quintile delivered at a facility (whether public or private) as opposed to one in every 100 women in the poorest. 53 Table 3-4: Key Health Status Indicators (1990-2007) Base Indicator Year 1993 1996 1999 2003 2007 MDG Target Infant Mortality Rate (Number of infant deaths 96 87 82 66 65 52 – 1 per 1,000 live births) Under 5 Child Mortality Rate (Number of deaths 1513 133 116 94 88 65 505 per 1,000 live births of children aged under 51 Maternal Mortality Ratio (Number of maternal 478 452 444 320 3154 – 1435 2 deaths per 100,000 live births) 1 Total Fertility Rate 4.3 3.4 3.3 3.3 3 2.7 2.26 1 Prevalence of Underweight Children Under 5 67.0%3 - 56.3% 47.7% 47.5% 46.0% – Source: Ministry of Finance, Ministry of Health and Family Welfare. Note: Base year varies depending on data availability: MMR (1990), TFR, IMR, U5MR (1991). 1 BDHS, reported at end year. 2 BBS date from the Vital Registration System. 3 MDG Progress Report, 2005: Bangladesh. 4 Last year for which MMR is available is 2001. 5 By 2015. 6 By 2010. 136. The HNP strategic investment plan (2003-2010) lays out four broad policy directions that can have an impact on reducing health inequalities. They are: (i) shifting resources to poorer districts (or districts with poor health outcomes); (ii) targeting and demand-side subsidies (subsidies to consumers of health services) as an alternative way of reaching the poor; (iii) diversification of service provision through public-private partnerships to improve quality and coverage of HNP services; and (iv) intersectoral collaboration to link with other ministries that have a direct impact on the health status of the poor. These policy directions require a concerted effort of both public and private providers. 137. Health services in Bangladesh are provided to the population by an extensive infrastructure of health facilities and a vast network of public and private service providers. The public health infrastructure extends from national to union level and includes all levels of health care provision from primary to tertiary care. Although there is no census of private health care providers, the network is believed to be large, comprising pharmacies and dispensaries, physicians, nonmedical practitioners and NGO clinics. 138. A significant portion of the population seeks care from pharmacies and other private providers. The largest share of total health expenditure (46 percent) is spent in drug retail outlets–a likely indicator that people might be seeking health care from unqualified personnel.39 The Ministry of Health and Family Welfare (MoHFW), which is the principal actor in the public health sector, spends about 26 percent of total health spending (Figure 3-2). Households are the largest source of financing for health spending, at 64.6 percent, almost entirely spent through out-of-pocket (OOP) payments, followed by the government with 25.6 percent (including donor contributions). NGOs account for about 5-6 percent. 39 Conclusive evidence on that is difficult in the absence of information about drugs sold with or without a doctor’s prescription. 54 Figure 3-2: Distribution of Total Health Expenditures (by Provider) Total health expenditure by providers, 1999-2000 Insurance and f oreign providers 2% MOHFW providers 26% Other public Drug and medical providers goods providers 2% 45% Priv ate providers Non-prof it 18% institutions and NGO providers 7% Source: Health Economics Unit, MoHFW (2003). 139. Expenditure on health is low by international standards. Whether expenditure is measured as a share of GDP or translated into per capita terms and adjusted for purchasing power parity, Pakistan was the only country in the region to spend less than Bangladesh on health. 140. Although not high by international standards, the overall budget envelope for the MoHFW is not a constraint as the ministry consistently underspends its budget. Total public expenditure on health increased in real terms by 44 percent over the 1998/9 to 2005/6 period, mainly as a result of the increase in national income. Development expenditure grew at an average annual pace of 6 percent, with a lot of annual variations, while nondevelopment (recurrent budget) expenditure experienced an actual decline, although allocations continued to increase (Figure 3-3). Capacity to plan and execute the budget remains a major challenge for MoHFW, just as it is for many other line ministries in the government. Figure 3-3: MoHFW Expenditures by Source of Funding (1998/9 to 2005/6) (Constant Prices 2006 in millions of Taka) 3 0 0 ,0 0 0 2 5 0 ,0 0 0 2 0 0 ,0 0 0 15 0 ,0 0 0 10 0 ,0 0 0 5 0 ,0 0 0 0 19 9 8 / 9 9 19 9 9 / 0 0 2 0 0 0 / 0 1 2 0 0 1/ 0 2 2 0 0 2 / 0 3 2 0 0 3 / 0 4 2 0 0 4 / 0 5 2 0 0 5 / 0 6 r e v e n u e e x p e n d it u r e d e v e lo p m e n t e x p e n d it u r e ( G O B c o n t r ib u t io n ) d e v e lo p m e n t e x p e n d it u r e ( D P c o n t r ib u t io n ) Source: CGA, MOF, HEU PERs. 141. While the overall health public resource envelope is not a constraint, public health facilities are inadequately funded. Expenditures on central level administrative functions of planning, management and regulation of the sector by the Ministry of Health are on the rise, whereas allocations to facilities that attend to the curative care needs of the population declined from 72 to 63 percent of 55 MoHFW spending. The funding share of facilities at upazila level and below, consisting of upazila health complexes (UHC) and union level facilities, which deliver hospital and primary care to the majority rural population, dropped from 51 to 42 percent. Declining funds to district hospitals and UHC are particularly worrisome in light of statistics showing that these facilities are used beyond their optimal capacity. 142. Health public resources are also not allocated equitably, either from a geographic or income perspective. The share of recurrent spending that goes to the 16 poorest districts is less than what goes to the 16 wealthiest districts. The traceable share of development spending accruing to the wealthiest districts has declined over time, but gains were not shifted to the poorest districts. Instead, the second poorest districts benefited. Furthermore, public subsidies (for curative care) in health also remain skewed to the non-poor. In almost all cases, the rich received a greater share of subsidies than the poor, with inequities being starker at the tertiary-care level for in-patient care (Table 3-5). These disparities have resulted in widely divergent health outcomes, although other (cultural) factors may have also played a role. Reaching the poor and improving their health status remains a major challenge for the health sector. Table 3-5: Distribution of Curative Care Subsidies by Equivalized Consumption Expenditure Quintiles (%) Tertiary Secondary Primary Other GOB Total Out- Out- Out- Out- Out- Quintile Inpatient patient Inpatient patient Inpatient patient Inpatient patient Inpatient patient Poorest 8 11 21 18 11 16 7 8 12 14 2 12 22 10 24 9 18 7 15 10 20 3 8 15 14 21 15 21 0 10 13 18 4 23 20 15 20 25 27 20 33 24 25 Richest 49 32 40 17 40 18 66 34 41 23 Total 100 100 100 100 100 100 100 100 100 100 Source: Authors’ calculation based on data from CGA, HIES 2005, HDS 2000. 143. In conclusion, addressing weaknesses in the current system requires action on three fronts: • improving service provision to the poor through better planning, allocation and targeting; • strengthening administration and management systems for better staffing of vacant posts, better budget planning and execution and monitoring and evaluation functions; and • improving leadership and sector governance by the MoHFW, so that they can play a more active role in managing and regulating the health system in the country. 3.4 Transport Sector: Policy, Financing and Outcomes 144. Bangladesh is one of the few countries in the world that has all modes of transport serving the needs of its economy and people with very little integration among them. It has a vast network of highways and rural roads, a railway system, inland waterways, two seaports, maritime shipping and civil aviation and a national airliner. The economic expansion and social development witnessed in Bangladesh since independence was accompanied by a rapid growth in transport demand, which grew at 9 percent per year. Much of this growth was met by road transport, which has emerged as the dominant mode of transportation over the years. 56 145. Transport sector policy in Bangladesh aims to provide safe and dependable transport services to the population. The policy is designed to play an important role in helping to reduce the transport costs of goods for export and keeping them competitive in the world market. A multimodal transport policy has been prepared that introduces the notion of an integrated system, combining rail, road and water transport. 146. Continued public investments in the road sector have established one of the most extensive road networks in a developing country. The road network has expanded by an additional 50,000 kms since 2001 and now covers some 271,000kms. Bangladesh has developed major road corridors connecting Dhaka with key economic centers and towns, and a network of village roads connecting communities to market centers and the main road. New bridges connect communities with road transport, as well as integrate whole regions, including the multipurpose Jamuna Bridge that provides uninterrupted east-west road and rail connection in Bangladesh. This extensive network surpasses other countries in South Asia in total road density. 147. Notwithstanding the massive expansion of transport infrastructure since independence, the services provided to users have not kept up with the demand in terms of quality and safety. The quality of the road network is poor, as roads are often too narrow for the traffic they carry. Only about 40 percent of main roads (the national and regional highways and the Zila roads) are in good condition, and the situation of rural roads is not very different. Congestion, overloading, air pollution, and safety are major problems. The role of the Bangladesh Railway in the economy has been limited and is declining due to sector governance challenges and inadequate resource allocation. The inland water transport system has not been used to its full potential, partly due to inadequate dredging of the waterways and lack of berthing facilities. Despite recent improvements (for example in the vessel turnaround times), the poor quality of Chittagong port services (the main port in the country), the associated land transport system (road and rail connection to the port), and the inland connections (mostly dominated by the trucking industry) have been a major barrier to trade. 148. Public finances have played and continue to play a major role in the transport sector and its outcomes. Accounting for about one-tenth of total government expenditures and the largest component of the ADP, transport spending declined slightly during 2000-2007. As a share of national income, however, it remained constant at around 2 percent of GDP. Nevertheless, national incomes increased by almost 50 percent during this period and transport expenditures also increased in real terms (Table 3-6). The largest share of transport spending was used on investments (about 69 percent) in new roads at the expense of essential rehabilitation and maintenance to improve existing road conditions and safety, which accounted for only 31 percent. While the transport investment (ADP) allocation is large compared to other sectors, it appears to be below what is required to achieve sustainable economic growth. The Tk.62 billion investment in 2006/07 represented 1.40 percent of GDP. Normally, developing countries with a satisfactory GDP growth rate invest 2.5-3.0 percent of GDP in transport. In South Asia, in order to sustain a GDP growth rate of 6.5 percent, investment in sealing roads alone is estimated to require 2.0 percent of GDP.40 40 GDP in 2006 was US$62 billion, and the average exchange rate in FY06-07 was Tk.71.3 per US dollar. The estimate for South Asia is from Chatterton and Puerto (2006). 57 Table 3-6: Transport Sector Expenditure by Sub-sector for FY 2003-2007 (billions of Taka) Sector 2002/3 2003/4 2004/5 2005/6 2006/7 Total (FY03-07) Rural Roads 15.3 20.3 24.5 29.7 30.5 120.3 Highways 23.5 23.6 28.6 26.2 21.8 123.7 IWT 6.4 7.0 8.2 11.6 9.9 43.1 Railways 14.8 14.7 13.1 16.5 15.6 74.7 CAAB 1.6 1.7 1.9 3.3 2.8 11.3 Total Transport 61.6 67.3 76.3 87.3 80.6 373.1 Total Expenditure (GoB) 445.2 492.2 581.0 594.5 739.9 2852.8 Transport as % of Total Spending 13.8 13.7 13.1 14.7 10.9 13.1 GDP Growth Rate 5.30% 6.30% 6.00% 6.60% 6.40% 6.1 Source: Annual Development Program, Ministry of Transport. 149. Despite recognizing the need for a more balanced approach to development of the transport sub-sectors, the government’s expenditure policy continued to favor the construction of new roads during FY2003-07. This was especially the case in rural areas with rural roads accounting for about 45 percent of total transport sector investments (ADP). This was followed by roads and highways managed by the Roads and Highways Department (RHD) with 38 percent, which means that the road sector accounted for 83 percent of the total transport investment (ADP) allocation during the period. Railways received about 12 percent, while inland water transport and civil aviation each received only 2 percent each (Figure 3-4). Figure 3-4: Transport ADP Allocation by Subsector (FY2003-2007) Railways, 12% CAAB, 2% IWT, 2% Rural Roads, 45% Highways, 38% Source: Ministry of Finance, Ministry of Communication 150. This public expenditure policy that emphasizes the road sector is not consistent with the government’s recent policy pronouncements to provide a more balanced development of the transport system. Despite the 2004 National Land Transport Policy that recommended increased investment in rail, the road sector continues to receive the lion’s share. The allocation of transport ADP funds to rural roads also increased faster during the FY2003-07 period, while allocation for rail remained constant and increased only marginally in 2007. At the same time, funding for road maintenance, especially for highways, has remained substantially inadequate, resulting in major deterioration of the 58 highway network. One-third of the national highway network is in poor or very poor condition, and the situation is much worse for regional and Zila roads, with 54 percent and 79 percent respectively in poor or very poor condition. 151. The relatively weak quality of the investment portfolio in transport further highlights the inefficiency of public expenditure policy. Limited resources combined with an inadequate forward planning process lead to significant delays in the implementation of the sector budget. The pressure to include too many projects results in resources being spread thin in small annual allocations per project, and in long delays in project completion. On average, for instance, RHD projects take six years and the range is from one year to 21 years (Figure 3-5). In the highway sector, the RHD has 131 projects on its books as of FY07, with an average completion time of seven years compared to an average of two years for international practice. While the degree to which different subsectors are able to spend these resources varies, the overall trend is one of underspending, which points to serious institutional capacity weaknesses in implementing projects. Furthermore, political influences also play a role in resource allocation, distorting priorities and contributing to inefficiencies. Figure 3-5: RHD ADP Portfolio Average Duration (in Years up to June 2007) 9 8 7 6 5 4 3 2 1 0 Fo reig n A id ed Pro ject s GOB Pro ject s-M ajo r GOB Pro ject s-M ino r Source: Ministry of Finance, Ministry of Communication. 152. While the need for investments in transport remains large, foreign financing to the sector has fallen. Weak implementation capacity and related governance issues are perceived to play a major role in the decline of foreign aid. In 1998, almost half the ADP came from foreign aid, while by 2007 the share had declined to 25 percent. In absolute terms, the total ADP allocation more than doubled between 1998 and 2007, while foreign aid fluctuated during the period. Foreign aid increased during FY00-04, but declined after FY05 as a result of donors reducing or ending their assistance and is currently at the level it was in 1998. 153. Cost recovery systems, an important financing instrument for the sector, remain weak. Users of public transport infrastructure and services cover only a minor part of the cost of operating and maintaining the transport system through fares and fees. Road user payments cover less than one-fifth (less than 20 percent) of the cost of maintaining the road system. In addition, the cost recovery system for road use is wrongly structured. All fees and charges are fixed annual values and do not depend on the actual use of the roads. This creates a perverse incentive for users to use the roads more than they would if there were a variable charge. A proposal for establishing a fuel levy is being considered by the government. The situation in the Bangladesh Railways is similar with users covering about 42 percent of the operating costs. 59 154. Going forward the government needs to balance its investment allocations in the transport sector and increase funding to meet maintenance requirements. Achievement of these objectives from a technical perspective will require: • a more effective budget management system, which is informed by performance and expenditure analysis; • strengthening cost-recovery systems in all modes of transport; • refocusing the role of the public sector to support maintenance for better quality and safety of the transport to improve the regulatory framework, avoid fragmentation and strengthen enforcement standards; and • re-orienting and modernizing the transport sector agencies to promote a well-coordinated and integrated transport system. 3.5 The Energy Sector: Policy, Financing and Outcomes 155. The energy sector provides a range of products and services to households and businesses, including electricity, gas and petroleum. Traditionally, energy provision has been dominated by a number of SoEs. More recently, private sector energy provision has gained some ground, but it remains low by international standards, and below its potential in Bangladesh. The government has established the Bangladesh Energy Regulatory Commission (BERC) to strengthen independent regulation and oversight of the energy market, with the authority to set wholesale and retail electricity and gas tariffs as well as to issue and revoke licenses in the energy sector. However, the new agency is just becoming operational. The Ministry of Power, Energy and Mineral Resources is now responsible only for setting the overall policy framework, overseeing state-owned energy suppliers and executing capital projects financed by the state budget. 156. Despite some improvements, the challenges in the energy sector remain formidable. Access to electricity has increased from only 32 percent in 2002 to 42 percent in 2005, which is commensurate with Bangladesh’s income level. Improvements in operational efficiency have brought Bangladesh on par with countries in the region, although system losses remain large. Bangladesh’s energy sector continues to suffer from low coverage, unreliable supply, and high technical and commercial distribution losses. Per capita consumption of electricity (163 kw/h) remains low by regional comparison and by comparison with other countries of similar per capita incomes. Moreover, access to, and consumption of, energy services remains hugely inequitable. Distribution networks are concentrated in urban areas with insufficient coverage in rural areas and to the poor. The poorest 20 percent of households are more than three times less likely to have an electricity connection than the richest 20 percent, and while 68 percent of urban households have electricity connections, only 29 percent do so in rural areas. As a consequence, poorer rural households have to rely on other energy sources, such as kerosene for lighting and other household uses (Figure 3-6). 60 Figure 3-6: Access to electricity has improved… …but per capita consumption is low 100 450 2 R = 0.5558 90 400 N = 67 Low Income Average 80 350 Electrification Rate/Percentage Population 70 Sri Lanka 300 60 India South Asia Average kwh/per capita 250 50 Pakistan Bangladesh 2005 200 40 Nepal 150 30 Bangladesh 2000 Bangladesh 20 100 10 50 0 0 100.00 1,000.00 10,000.00 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Per Capita GDP/2005/Current US$/Log Scale 100% 89% 90% 86% 82% 79% 80% Kerosene Percentage Household Using Energy Source 72% 70% 62% 60% Electricty 50% 46% 40% 37% 30% 27% 19.32% 20% 17% Natural Gas 9.54% 10% 5.50% 4.98% 1.46% 0% Poorest 2 3 4 Richest Consumption Quintile Source: World Bank World Development Indicators, 2005. HIES, 2005. 157. Most of these problems have their roots in a weak institutional framework and restricted competition with large SoEs dominating the sector. In addition, inadequate pricing for many energy products has undermined the financial viability of energy suppliers, depriving them of the financing needed to carry out routine maintenance activities and to invest in the expansion of access. Prices for many energy products and services, including electricity, gas, diesel and kerosene, are regulated at levels below cost. Although politically costly to adjust, low energy prices drain scarce public resources and are poorly targeted, with more affluent social groups benefiting disproportionately. In addition, low energy prices also distort price signals to households and businesses, leading to inefficient consumption patterns and production technologies (for example captive diesel generators rather than grid). 158. The provision of energy services below market prices is a quasi-fiscal activity that imposes significant fiscal costs on the government and creates substantial contingent liabilities. In 2007, the energy sector incurred cumulative losses of more than Tk. 42 billion (Figure 3-7). Since utilities lack long- term financial sustainability, it is likely that large portions of these losses will be shifted to government. 61 Figure 3-7: Energy SoEs: Profit/Loss Source: Ministry of Finance, Bangladesh. 159. Explicit budgetary subsidies and discreet capital projects financed through the development budget cover only a small fraction of the total public financing provided to the sector. The lion’s share is hidden in implicit transfers not shown in the budget, such as loans provided to energy suppliers at below-market rates (in many cases on lending from international development partners), guarantees on loans by SCBs, ad hoc liquidity support, and cross-subsidization through administered input prices for gas and fuel to power plants. In addition, depreciation of state assets, mainly the depletion of the country’s fossil energy reserves and forgone returns in the form of SoE dividends, add to the cost. This results in huge direct and indirect costs for the budget. 160. The total fiscal cost of energy subsidies is in the vicinity of Tk. 106 billion, equal to 2.2 percent of GDP annually, about the same as government spending in education and more than health and social welfare expenditure (Figure 3-8). This estimate recognizes full coverage of SoE operational losses and the opportunity costs of low gas prices (but does not include capital injections from the development budget). Only about one-third of these transfers are allocated and accounted for in the budget. Given the predominantly indirect nature of energy subsidies, they are largely outside the resource prioritization and control processes of the budget framework. As such, they remain shielded from competition with other expenditure programs in the budget and some, such as loan guarantees, represent significant fiscal risks that threaten to undermine long-term fiscal sustainability and may cause adverse macroeconomic effects. 62 Figure 3-8: Energy Subsidies Impose a High Opportunity Cost but are Largely Inequitable 250 Annual Per Household Benefit from Kerosene Subsidy 244 223 216 200 192 Taka/Per Year/Per Household 169 150 100 50 - Poorest 2 3 4 Richest Consumption Quintile Source: ADP, MoF, BPC, Petrobangla, HIES 2005. 161. Since energy consumption is progressive, with higher income groups consuming more, these subsidies are largely inequitable and disproportionately benefit the more affluent. Subsidies are primarily captured by the urban middle- and upper-class that have both better access to and higher per capita consumption of energy. At the same time, they impose high opportunity costs on the government, constraining investments in social sectors and–with the exception of equity injections that account for roughly one-third of the total–they are also largely consumptive in nature. As such they do not create the long-term (human or physical) capital assets needed to generate future growth and income. 162. A number of policy options emerge from the analysis: • tariff reforms should be considered to improve the commercial viability of service provision and to reduce the fiscal burden resulting from underpricing. Reducing tariffs could substantially improve the targeting of subsidies in the power sector where price discrimination based on consumption levels is possible. While the current structure is already progressive, lower tariffs could ensure affordable electricity supply to poor households that are connected to the grid while, at the same time avoid mistargeting of higher-income households; • where direct price controls are deemed desirable, financing should be made more transparent and should be explicitly provisioned for in the budget as direct subsidies of the end-user selling price. This would reduce fiscal risks and expose resources devoted to lowering energy prices to the same political expenditure prioritization process as other competing expenditure programs; • it would be critical to establish predictable and transparent adjustment mechanisms to ensure timely responses to changes in the business environment; • a more balanced growth strategy in the power sector is needed. Generation capacity needs to be increased to keep pace with increasing demand; and • governance reforms in the sector are essential to improve operational efficiency. 63 3.6 The Agricultural Sector: Policy, Financing, and Outcomes 163. The agricultural sector in Bangladesh continues to play a critical role in ensuring food security and in generating employment and income, particularly for Bangladesh’s predominantly rural population. Agriculture is no longer the major source of economic growth, comprising 18.2 percent of GDP and contributing an average of 3.7 percent to its growth. Nevertheless, the sector continues to be the major source of income for Bangladeshi workers, with about 46 percent of the total labor force being employed in labor-intensive agriculture. 164. Bangladesh has achieved self-sufficiency in rice production as a result of impressive productivity-led growth in agriculture over the past three decades. There has been impressive progress in raising productivity and diversifying the agricultural sector, mainly as a result of investments in rural infrastructure, including disaster management and mitigation systems, rural roads, electricity, and liberalization of markets for important agricultural inputs. The enhanced use of agricultural inputs, mainly high-yielding seed varieties, fertilizer, and irrigation equipment, has played an important role in increasing cropping intensity and yield. Notwithstanding this achievement, gains in food security at both the national and household levels are frequently threatened by supply and price shocks caused by natural disasters. Large segments of the population remain highly vulnerable and continuously face the distress of food insecurity. This is compounded by the widespread prevalence of undernutrition, with 40.4 percent of the population consuming less than the 2,122 minimum caloric intakes per day set for Bangladesh. 165. To ensure food security over the medium term, significant increases in agricultural productivity are required for food supply to keep pace with rising demand. Population growth has slowed down to 1.5 percent at present but, at this rate, about 2 million people are added to the total population each year, generating a requirement for an additional 350,000 tonnes of food grain every year. Being one of the world's most densely populated countries, there is little scope for additional land to be brought under agricultural production and future growth will have to continue relying on raising yield on major food grains. Despite productivity growth, average yield rates in Bangladesh are low in comparison with other countries in the region and far below potential yields of modern rice production (5-7 tonnes/ha). In addition, there are substantial domestic yield gaps between modern irrigated production at 3.45 t/ha and rain-fed agriculture using local rice cultivars at 1.16 t/ha. 166. Public expenditures on agriculture have grown to 12 percent of agricultural GDP, which is commensurate with spending levels in countries of similar income. In addition, significant off-budget costs accrue in the form of cross-subsidies from the gas sector to the fertilizer sector and operating losses of Bangladesh Petroleum Corporation (BPC) on subsidized diesel sales, bringing total implicit and explicit liabilities to roughly 16.2 percent of agricultural GDP. While the level of spending is adequate, resource use is heavily skewed towards subsidies of agricultural inputs. Input subsidies, most notably on fertilizer and diesel, now account for roughly 40 percent of the total budgetary expenditures on agriculture. The rapid expansion of over the past five years as a result of rising international commodity prices has crowded out investments in core public goods such as research and development, extension services and water and irrigation infrastructure, which have fallen from 63.4 percent in FY99 to 22.4 percent of agricultural expenditures in FY07. In particular, spending on research and development is low at only 0.3 percent of agricultural GDP (Figure 3-9), compared to 0.62 percent for other developing countries. 64 Figure 3-9: Public Expenditures on Agriculture Have Increased, but Agricultural Subsidies Dwarf Expenditures on Public Goods 13.0% 5.0% Agricultural Development Expenditure as Share of Total Expenditures FY 2007 12.2% Agricultural Revenue Expenditure as Share of Total Expenditures 0.8% 4.5% 11.0% Non-Urea 2.7% Fertilizer 2.6% 4.0% 4.0% BPC Urea Subsidies Operatiing 9.0% 3.5% Loss from Percentage of Total Expenditures Diesel Sales to Percent of A gricultural G D P 8.0% 2.4% Agricultural 3.0% Sector 7.0% 2.5% 1.3% 5.8% 5.2% 5.4% Water and 2.5% Irrigation 5.0% 5.4% 9.3% 2.0% 4.6% 4.5% Infrstructure 4.2% 3.9% 4.1% 4.2% 3.9% 3.5% 2.5% 6.7% 1.5% 1.6% 3.0% Total Expenditures as Share of Agricultural GDP BOMG 1.0% Cross-Subsidy Extension 3.9% 1.0% on Gas Sales Services 2.9% to Fertilizer Public Food 2.3% 2.2% 2.3% 2.2% 2.2% Distribution System 1.0% Sector 0.5% 0.3% 0.3% 1999 2000 2001 2002 2003 2004 2005 2006 2007 0.0% R&D -1.0% On Budget 1 Off Budget 2 On Budget 3 4 Fiscal Cost of Subsidies on Production Side Subsidies on Consumption Side Public Goods Source: Ministry of Finance, Bangladesh. 167. There are substantial benefits to be gained from changes in the resources allocation as part of a comprehensive strategy to improve agricultural productivity. First and foremost, this would entail a reform of the markets for agricultural inputs. Input subsidies are not only costly but they are not targeted and price controls have undermined the development of commercial distribution systems, routinely causing the undersupply of key inputs. They also distort price signals on input markets, encouraging farmers to adopt inefficient production technologies, and have led to wastage and imbalanced fertilizer application. Where input subsidies are deemed necessary, their full cost, including off-budget liabilities, should be captured in the state budget and compete with alternative expenditure programs. In addition, voucher or coupon systems could improve targeting of subsidies to marginal producers while at the same time fostering the development of commercially viable wholesale and retail markets for agricultural inputs. 168. In step with reforms of input markets, policies to achieve food price stability and safeguard food security, especially for vulnerable groups, could be gradually shifted to interventions on the consumption side, such as food aid or cash transfers. As Bangladesh’s commercial agriculture grows and food markets develop, shifting support to the consumption side is likely to achieve efficiency gains. In addition to up-scaling existing food aid programs, this could also pertain to cash transfer programs to poor consumers, allowing households to make dietary choices based on relative costs and nutritional preferences, while at the same time providing incentives for producers to adjust their outputs to consumer choices. 169. Finally, investments in public goods aimed at increasing agricultural productivity could be stepped up commensurate with the implementation capacity needed to deliver quality expenditure programs: • Investments in extension services could help diffuse high productivity technologies. As there are significant yield gaps between modernized and traditional farming, an expansion would potentially raise productivity and production levels. • Research and development is an area that is likely to yield significant returns. Current spending in Bangladesh is far below that of other countries, and Bangladesh has a strong existing research infrastructure under the Bangladesh Agricultural Research Council (BARC). 65 • Increasing the rural electrification rate. This would allow for the greater use of electric irrigation systems, which are more cost effective than diesel powered ones. • Linking farmers to consumer markets through better rural roads. This reduces the cost of agricultural inputs as well as marketing costs of outputs, thereby improving farm profitability and reducing consumer prices. The recent expansion of the rural road network has produced some returns in this regard, but it is vital to provide sufficient financing to road maintenance to sustain these achievements. • Expanding the strategic grain reserve is an appropriate response to provide insurance against vulnerability to frequent supply shocks. The optimal reserve level could be based on a risk profile derived from historic trends in production volatility. 3.7 Conclusions and Recommendations 170. Bangladesh’s very good record of maintaining fiscal discipline and allocating resources to broad strategic priorities is not matched at the sector level. The record at this level is relatively poor as shown by the review of operational efficiency in five key sectors–education, health, transport, energy and agriculture. A number of common features across all sectors demonstrate the need to significantly improve the quality of public expenditure at the sector level. A failure to do so risks undermining the sustainability of the good social outcomes that have been achieved. 171. Sector policy frameworks are in place and appear to be appropriate for all sectors under review. They recognize the issues and challenges faced by the sectors and present a reasonable set of policies to take the sector forward in order to address such challenges. They also attempt to measure sector performance through various output or outcome indicators, in some cases as part of the overall strategy for accelerating economic growth and poverty reduction. 172. The state/public sector continues to play a substantial role in expenditure allocations within sectors. However, there is a need to realign expenditure allocations with sector policies in almost all sectors under review. Expenditure allocations do not always support the sector policy objectives, and expenditure trends reveal a passive expenditure policy. If access to, and quality of, education is to be improved, there is a need to not only maintain the level of spending but to further increase it, especially to support primary and junior secondary education. Investment in increasing access to education needs to go beyond providing school places and include demand-side interventions to attract and retain students in school, in addition to providing more training and better pay to teachers. Expenditure on the administration of health services are on the rise, whereas allocations to facilities that attend to the curative care needs of the population have declined, at a time when some of these facilities are used beyond their optimal capacity. The objective of the transport strategy to develop a balanced and integrated transport sector needs to be matched by an expenditure policy that places adequate weight on road maintenance instead of focusing mainly on road expansion. The energy sector continues to put pressure on public expenditure to cover operational losses and inject fresh capital into a sector that suffers from many problems, contributing to a buildup of contingent liabilities. 173. Expenditure realignments within sectors need to address the distribution of spending, which is not pro-poor. While success has been achieved in narrowing (and sometimes reversing) the gender gap in education, spending in the sector has remained heavily skewed towards the non-poor. Even the stipend program designed to target the poorest 40 percent is only partially pro-poor. Despite the health sector policy emphasis on achieving equity in service provision, health expenditures are not allocated 66 equitably, either from the perspective of geographical distribution or household income. Energy and fertilizer subsidies need to be better targeted in order to benefit the poor and less profitable enterprises. 174. The quality of investment portfolios in almost all sectors, especially those with a large capital investment share, requires particular attention. Investment spending suffers from delays and slow implementation. The pressure to spread resources thinly across many projects through an inefficient planning process results in long delays in project completion and underspending of investment programs year after year, with very few exceptions. Institutions for planning and management of public finances at the level of the line ministries, which currently suffer from segregation, lack of coordination, and outdated information and data sharing systems, will be key to improving the effectiveness and efficiency of spending at the operational level. 175. Addressing these challenges is not a straight forward matter of technical fixes. As the following chapter explains, a number of institutional incentives and political economy constraints will need to be surpassed to improve the outcomes. 67 PART II INSTITUTIONAL INCENTIVES AND POLITICAL ECONOMY CONSTRAINTS TO BETTER PFM OUTCOMES 176. The ability of, and incentive for, politicians to effectively address the issues and implement the key policy priorities highlighted in the previous chapters depends on the nature of the government institutions under their control. If the fiscal institutions and the budget system work effectively, then the policies must be the best choices from the politicians' point of view, subject to resource constraints. If the electoral system also works well, then those choices should further reflect some aggregate version of public preferences, thus leading to efficient outcomes for all. However, those systems always suffer from imperfections due to limited information. Therefore, to understand the budget policies and outcomes presented in Part I and to seek remedies for mitigating inefficiency, it is important to know to what extent the budget and the electoral institutions in Bangladesh are deficient. 177. In the following chapter, we focus on the analysis of institutions that dominate the budget machinery in Bangladesh across the three dimensions–maintaining fiscal discipline, ensuring allocative efficiency and reaching operational efficiency. We also discuss some of the political economy elements that may have influenced the creation and survival of the current institutional structure. The assessment follows the methodology developed by von Hagen and Harden (1994 and 1995), Campos and Pradhan (1996), and Alesina et al. (1999). In particular, we employ the extended version of the methodology attributed to Esfahani (2000). The approach consists of ranking institutional arrangements at each step of the budget process and then combining those rank-orders to come up with aggregate indicators of institutional capabilities. Esfahani (2000) extends the methodology of earlier studies in this area by examining the steps of budget-making and implementation in greater detail and taking into account the economic and institutional characteristics that affect the effectiveness of each arrangement41 (Appendix 4). Asiedu and Esfahani (2001) apply the methodology to a set of countries from the Middle East and North Africa in addition to Bangladesh. The methodology is also applied to a set of countries in Latin America, East Asia, and OECD. 178. The results drawn by applying this methodology in the case of Bangladesh are very consistent with several other studies that have assessed Bangladesh’s institutional arrangements in recent years. This framework, however, groups institutional characteristics according to their relationship with the three levels of public expenditure outcomes, fiscal discipline, allocative efficiency and operational efficiency, and give a sense of the relative priorities that need to be addressed in order to influence the incentive structure for improving budget institutions and outcomes. 41 The information required for the rankings in the case of Bangladesh has been collected on the basis of a survey questionnaire designed to extract information from the Bangladeshi PFM officials and experts, World Bank experts working in this area and previous papers and official reports in early 2008. 68 CHAPTER FOUR INSTITUTIONAL CHANNELS FOR INFLUENCING BUDGET OUTCOMES The management of public finances in Bangladesh is concentrated in the hands of two key central agencies, the Ministry of Finance and the Planning Commission. These institutions play a key role in planning, allocation, management and monitoring of public expenditures and revenues. This institutional structure has served the country well in maintaining aggregate fiscal discipline and broadly allocating resources to key strategic priorities, where Bangladesh does better than several other developing countries. The delegation of control over fiscal aggregates to MoF and the use of ex-ante agreements on the size of fiscal aggregates, coupled with capable staff with reasonably long-term horizons working in this area have contributed to these results. The limited autonomy and flexibility of the line agencies accompanied by weak administrative and personnel management practices associated with this framework have, however, proved to be problematic for operational efficiency as demonstrated by rather weak expenditure outcomes at the sector level. This institutional structure for budget management and the role it plays in the effective management of public resources has its roots in the country’s historical and political developments and the nature of the system that prevailed after independence. Changing these structures, therefore, has not proven to be an easy task. A different approach that addresses the political economy constraints and the underlying incentive structure for maintaining the status quo is required. 4.1 Institutions for Aggregate Fiscal Discipline 179. Institutional arrangements for aggregate fiscal discipline provide the machinery of rules, procedures and coordinating mechanisms that influence the incentive structure and decisions of the main players involved in budget management. This helps to avoid the possibility that too many claimants with differing self-interests will have control over the same pool of public resources. The key dimensions considered to influence the performance of the fiscal institutions in these areas are: (i) delegation of control over fiscal aggregates; (ii) use of macroeconomic programming; (iii) legal limits on spending and borrowing; and (iv) ex-ante agreements over the fiscal aggregates. Bangladesh does well in all of these dimensions.42 The overall Aggregate Fiscal Discipline Index is much better than several other countries for which this information is available (Table 4-1). 42 The only exception is the legal limits on spending and borrowing for which there is no provision. 69 Table 4-1: Aggregate Fiscal Discipline Indexes of Budgetary Institutions Aggregate Dominance of Discipline Central Limits on Country Arrangement Budget Ex-Ante Macroeconomic Borrowing/ (Period)1 Index Agencies Agreements Program Spending Bangladesh (2008) 0.52 0.64 0.67 0.44 0.00 Algeria 0.47 0.65 0.63 0.27 0.00 Egypt 0.54 0.78 0.40 0.63 0.00 Indonesia 0.61 0.93 1.00 0.08 0.00 Iran 0.37 0.28 0.33 0.63 0.00 Jordan 0.48 0.87 0.40 0.27 0.17 Korea 0.33 0.33 0.33 0.44 0.00 Kuwait 0.35 0.45 0.43 0.27 0.00 Lebanon 0.24 0.33 0.40 0.00 0.17 Morocco 0.54 0.72 0.80 0.27 0.00 Tunisia 0.81 0.77 1.00 0.81 0.33 UAE 0.36 0.92 0.23 0.00 0.17 Yemen 0.56 0.49 0.90 0.44 0.17 Brazil (1995/9) 0.47 0.69 0.43 0.44 0.00 Turkey 0.28 0.48 0.20 0.27 0.00 New Zealand 0.62 0.32 0.77 1.00 0.00 USA 0.51 0.40 0.50 0.81 0.00 Source: Staff estimates. 1 Note: The period is the 1990s unless otherwise indicated. - Indicator values are from 0 to 1 with 0 being the lowest. Higher indicators show better performance. - Data are based on a survey of budget experts (local and international) in Bangladesh conducted in early 2008 (see Annex v for the analytical and methodological framework. 180. The dominant role that central finance agencies (CFAs), in particular MoF, play in budget management and ex-ante agreements on the overall budget envelope and deficit limits, have contributed to Bangladesh’s strong record of fiscal discipline. At the start of the budget process, MoF prepares the macro-fiscal framework for the following two years and, in cooperation with the Planning Commission and the Bangladesh Bank, defines the budget envelope for the recurrent (revenue) and investment (ADP) budgets. After being prepared by MoF, the macro-fiscal framework and budget aggregates are reviewed and approved by the Resource Committee headed by the Finance Minister43 before presenting them for acceptance or rejection to the cabinet. 181. While MoF is responsible for the recurrent budget and the Planning Commission (PC) for the ADP, MoF has the delegated responsibility of guiding and coordinating the whole preparation and implementation process, especially with regard to fiscal aggregates. With the introduction of the MTBF approach in 2004, MoF introduced the practice of assigning indicative budget ceilings to selected line ministries that were part of the MTBF exercise, adding an additional element of good fiscal discipline that did not exist prior to MTBF. This delegation of powers and decision-making with regard to fiscal aggregates to a few individuals prevents decision makers with differentiated interests from expanding the budget envelope as the cabinet can only accept or reject the fiscal aggregates. 43 The resource committee consists of the following members, chairman of NBR, the Governor of the Bangladesh Bank, Secretary of Finance, and the Planning Secretary. 70 182. While these coordination mechanisms are necessary for disciplined budget making, they are not sufficient. The effectiveness of these arrangements depends on other factors. In particular, better outcomes are more likely when the politicians and administrators implementing the budget are more capable and accountable, with reasonably long-term horizons in their positions.44 Effectiveness should also depend on the extent to which the budget is comprehensive and free from segments with autonomous growth. Finally, maintaining discipline is crucially dependent on the presence of hard budget constraints on agencies and ex-post restraints on the executive more generally. Table 4-2 shows the results of the extended framework of indexes. Table 4.2: Indexes of Effectiveness Characteristics Affecting Budget Discipline Average Tenure of Auto- Central Education nomous Comprehen- Hard Budget Level of Country Budget siveness of Ex-post Budget Agency Decision- Account- (Period)1 Segments Budget Restraints Constraints Heads makers ability Bangladesh 0.80 0.40 0.71 0.88 1.00 0.75 0.40 Algeria 0.83 0.55 0.50 0.73 0.75 0.75 0.36 Egypt 0.77 0.76 0.27 0.86 1.00 0.75 0.43 Indonesia 0.78 0.48 0.70 0.50 1.00 1.00 0.28 Iran 0.67 0.32 0.47 0.55 1.00 0.75 0.59 Jordan 0.67 0.64 0.23 0.80 0.50 0.25 0.43 Korea 0.97 0.45 0.30 0.78 0.50 0.75 0.84 Kuwait 0.71 0.73 1.00 0.72 1.00 0.75 0.44 Lebanon 0.61 0.42 0.32 0.83 0.50 0.75 0.49 Morocco 0.93 0.33 0.16 1.00 1.00 0.50 0.54 Tunisia 0.73 0.74 0.54 0.72 0.75 1.00 0.40 UAE 0.73 0.55 0.42 0.66 1.00 0.75 0.43 Yemen 0.62 0.67 0.30 0.80 1.00 1.00 0.25 Brazil (1995-9) 0.40 0.40 0.27 0.68 1.00 1.00 0.81 Turkey 0.83 0.26 0.39 0.34 0.75 0.50 0.58 New Zealand 0.94 0.94 0.65 0.98 1.00 1.00 0.95 USA 0.60 0.92 0.86 0.93 1.00 1.00 0.96 Source: Staff estimates. 1 Note: The period is the 1990s unless otherwise indicated. - Indicator values are from 0 to 1 with 0 being the lowest. Higher indicators show better performance. - Data are based on a survey of budget experts (local and international) in Bangladesh conducted in early 2008 (see Annex v for the analytical and methodological framework 183. Combining the effectiveness indicators presented in Table 4-2 with the scores for coordination mechanisms can yield an overall Effective Aggregate Discipline Index. This measures (on a scale from 0 to 1) the degree to which budget institutions allow for the buildup of fiscal liabilities at the implementation stage. Figure 4-1 maps this index against the log of average inflation rates for the sample of countries where comparable data is available. The graph highlights the inverse relationship between the two variables and suggests that lack of fiscal discipline tends to lead to inflationary pressures.45 Based on the methodology used here and other analysis, it appears that the Effective 44 Capability is measured by the educational attainment and accountability is a combination of seriousness of audit, ex-post reconciliation, explicit sanctions, information availability, freedom of the press, extent of democracy, educational attainment of the public, and openness of financial markets. 45 Of course, inflation is driven by other factors as well and, in that sense, is an imperfect indicator of the outcome of discipline failure. However, it is still a better measure for the purpose at hand than the other measures commonly used for this purpose—budget deficit and 71 Aggregate Discipline Index in Bangladesh is higher than that in most other developing countries, and its inflation rate has been accordingly lower. The dominant role of the CFAs and ex-ante agreements reached on the fiscal aggregates without providing an opportunity for later changes in the process have played an important role in this respect. Figure 4.1: Effective Aggregate Fiscal Discipline Index and Inflation 8 7 Brazil 85-94 6 hown) n ss 5 io d Brazil 64-84 lat io I Inf r Turkey 90s rpe P 4 fC dove go Lebanon 90s Iran 90s Brazil 95-99 e o Algeria 90s g L 3 a Yemen 96-00 Indonesia 90-97 er Egypt 80s Av Egypt 90s ( 2 Jordan 90s Bangladesh 01-06 Tunisia 90s Korea 90s Morocco 90s 1 UAE 90s Kuwait 90s USA 90s NZ 94-98 0 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Effective Aggregate Fiscal Discipline Index Source: Staff estimates. 184. The comparison with other countries shows that the main areas of weakness for Bangladesh among the effectiveness characteristics are weak accountability and lack of budget comprehensiveness. Aspects of the accountability problem that have some bearing on aggregate fiscal discipline are: (i) inadequate and delayed external audit and the absence of ex-post reconciliation of the budget; (ii) lack of explicit sanctions in case of deviations; and (iii) limited information about the budget performance made available to interested parties, including the public. While a lot of progress has been made to improve and modernize the external audit function, there is still a long way to go to make it a key instrument of accountability (Box 4-1). More recently, the government has made available fiscal information on the MoF website but, given the limited access and low levels of education among the Bangladeshi people, the use and effectiveness of this information is somewhat limited. A major source of comprehensiveness shortfall is the large off-budget expenditures that take place through public enterprises and their borrowings (see Chapter One and Appendix One for more details).46 public debt—because countries with good discipline do run high deficits and debt when necessary and may sometimes do so more aggressively than the countries with low discipline that are less creditworthy. 46 A full assessment of budget comprehensiveness and contingent liabilities of the public sector is beyond the scope of this report. There will be a follow-up policy note in the coming months that will focus specifically on these issues. 72 Box 4-1: External Audit in Bangladesh The role and independence of the Comptroller and Auditor-General (C&AG) in Bangladesh is still evolving. In most countries with an effective accountability structure, the C&AG or its equivalent is independent of the executive and reports directly to the parliament. In Bangladesh, the C&AG is not yet fully independent as the institution is, to a large extent, still part of the executive. The C&AG is appointed by the president on the advice of the parliament. However, the C&AG cannot hire qualified auditors, restructure or promote staff without the approval of the government. Audit reports are first sent to the government, which then sends them to parliament. While reports are not usually altered and audit findings are not changed, this practice is likely to deter reporting on major violations that may implicate the ruling party. Audit and accounting cadre are not separate. Members of the audit and accounts cadre are transferred randomly. Accountants may be assigned to the C&AG and auditors may be transferred to the Controller General of Accounts office. This gives rise to several issues, not least of which is conflict of interest, as accountants assigned to the C&AG may end up auditing their own agencies and their peers or even their own transactions. Capacity constraints of the C&AG result in significant backlogs. The C&AG has 22,000 auditable units (including SoEs) and 3,500 staff, most of whom are junior level accountants. Public audits cover about 16-25 percent of C&AG’s mandate every year. The quality of the report does not draw the attention of line ministry management as audit reports include too many trivial issues at the expense of major violations. The follow-up process is often not timely and not coordinated by the respective Sectoral Committee on individual Ministries (SCM). Transparency of audit findings, responses and follow-up are generally limited, reducing demand-side pressures that provide added incentives for accountability. Once an audit report is submitted to the parliament, it should become a public document. Although this is legally mandated, audit reports, Public Accounts Committee (PAC) findings, PAC hearings and follow-up reports are not published by either the C&AG or by the PAC. 185. The delegation of fiscal management to the CFAs, such as MoF and PC, has had a positive impact in maintaining fiscal discipline. This has its roots in the country’s historical and political developments and has been highly influenced by the significant degree of political fragmentation that emerged after independence. Political fragmentation and the absence of strong political and administrative institutions led to a chaotic economic situation, which was brought under control through a military regime. This era was followed by a period of political consolidation and the emergence of two main political parties, Awami League (AL) and Bangladesh National Party (BNP), and bureaucratic centralization overall. The bureaucracy delegated substantial powers to the central finance agencies to bring the economic and fiscal situation under control (Box 4-2). 186. In addition, the fear of inflationary pressures associated with a lack of fiscal discipline is another factor that has worked in favor of the politicians’ incentive to maintain fiscal discipline. Politicians understand that inflationary deficits are quite noticeable by the electorate and can quickly turn into lost votes for the incumbent party, especially given the centralized nature of budget procedures in Bangladesh that makes the top few politicians responsible for the outcome. On the other hand, the use of off-budget expenditures through mechanisms such as public bank/enterprise borrowings as a means for redistribution–which seem to be a source of fiscal risk in Bangladesh–can satisfy many constituencies without having tangible effects on voters for a long time. Even when they are recognized as government liability years later, they would simply be added to government debt and their amortization would be spread over time and passed on to future generations. Indeed, the limitations of the regular budget and the internal and external pressure to keep deficits down seem to have added to the incentive to expand off-budget spending; hence the substantial use of public banks and enterprises for this kind of redistribution (Nizamuddin 2003) and low indexes on budget comprehensiveness. 73 Box 4-2: Political Developments and the Way to Bureaucratic Centralization Bangladesh emerged as an independent country in 1971 with a very fragmented political structure–a coalition consisting of different small groups around, and inside, one political party. Although the state that took shape under these circumstances had a democratic premise, it was highly fragmented and, consequently, ineffective in addressing the enormous economic problems facing the country. The entire movement at that time was mobilized around the issue of independence–with few concrete programs for dealing with post-independence problems and long-term development. At the same time, the civil service cadres and administrative procedures that had been in place deteriorated, owing to disruptions and exits from the civil service related to Independence. Matters were made worse from an administrative point of view by the new government's decision to nationalize many firms, running them as state-owned enterprises (SoEs) and hence adding to the burdens of the civil service. The high degree of political fragmentation at the time and the absence of strong political and administrative institutions to provide a coordinating role led to a serious common pool problem in the budgetary process, where many different groups demanded benefits from the government without proper procedures and rules in place to prioritize expenditures and maintain aggregate budget discipline. Attempts to bring the situation under control following the famine of 1974 and a state of emergency resulted in a constitutional change and the transfer of powers from the legislative and judicial branches to the newly established institution of the president. This attempt to bring order through executive centralization was eventually followed by a series of military coups and countercoups, which ended in a change of power and the emergence of another political party, the BNP. This period was followed by an era of liberalization and, although the administration reverted to a presidential- parliamentary system from one with an executive prime minister, bureaucratic centralization remained strong. 187. The dominant role played by central budget agencies has also helped Bangladeshi governments to achieve good results beyond fiscal discipline. Their role is also associated with achieving a reasonably good level of allocative efficiency and has allowed for scarce public resources to be directed to the broad strategic priorities of financing education, health and supporting rural development. (This was discussed in Chapter Two and will be covered in more detail in the following section). 4.2 Institutions for Allocative Efficiency 188. In addition to serving the objective of maintaining fiscal discipline, institutional arrangements must also ensure that resources are allocated to strategic priorities. Chapter Two showed that governments of different political persuasions have broadly maintained the critical priorities in the allocation of resources consistent with the country’s strategic development objectives. How has Bangladesh been able to maintain these broad inter-sectoral priorities and sustain the financing of key public services? Which institutional dimensions affect more directly the ability of the government to direct resources to key strategic priorities and achieve good outcomes on allocative efficiency? 189. The literature and experience shows that the institutional arrangements that matter most for achieving allocative efficiency are those that affect the prioritization of resources and the flow of information about costs and benefits of alternative programs. Three dimensions of the budget process are important for efficient allocation: (i) prioritization; (ii) line ministry participation and flexibility; and (iii) breadth of consultation over the budget. Bangladesh does well in some of these dimensions as shown by the Allocative Efficiency Indexes, which measure the degree to which objective criteria are used in the allocation of resources among sectors and whether these priorities can be altered during the review, approval and the implementation of the budget (Table 4-3). 74 Table 4-3: Indexes of Institutional Arrangements for Allocative Efficiency Allocative Efficiency Arrangement Participation of Flexibility of Breadth of Country – Period1 Scores Prioritization Line Agencies Line Ministries Consultations Bangladesh 0.39 0.61 0.24 0.00 0.71 Algeria 0.29 0.33 0.30 0.39 0.14 Egypt 0.27 0.50 0.11 0.17 0.29 Indonesia 0.28 0.67 0.02 0.00 0.43 Iran 0.47 0.67 0.40 0.25 0.57 Jordan 0.39 0.44 0.24 0.58 0.29 Kuwait 0.39 0.61 0.67 0.00 0.29 Lebanon 0.34 0.50 0.33 0.22 0.29 Morocco 0.41 0.83 0.00 0.36 0.43 Tunisia 0.41 0.89 0.19 0.11 0.43 UAE 0.22 0.17 0.44 0.11 0.14 Yemen 0.13 0.50 0.00 0.00 0.00 Brazil 1995/9 0.27 0.50 0.04 0.25 0.29 Turkey 0.08 0.17 0.00 0.00 0.14 New Zealand 0.81 1.00 1.00 0.67 0.57 Korea 0.50 0.67 0.38 0.39 0.57 USA 0.68 1.00 0.67 0.50 0.57 Source: Staff estimates. 1 Note: The period is the 1990s unless otherwise indicated. - Indicator values are from 0 to 1 with 0 being the lowest. Higher indicators show better performance. - Data are based on a survey of budget experts (local and international) in Bangladesh conducted in early 2008. (see Annex v for the analytical and methodological framework). 190. Good performance in prioritization and line ministry consultations seem to be the main factors affecting the ability of the government to maintain broad strategic priorities in inter-sectoral resource allocations. However, these factors mask some fundamental weaknesses of the existing accountability framework. The limited role played by the cabinet, and especially the parliament, in the review and scrutiny of the budget negatively affects the results that can be achieved at the operational/agency level. 191. Two central agencies, MoF and the PC, play the dominant role in inter-sectoral allocations. The process of prioritizing resource allocations among sectors begins with the preparation of the indicative budget ceilings from the MoF in cooperation with the PC.47 Budget ceilings are then vetted by the BMRC. While the quality of the analysis that goes into this process needs to be upgraded substantially, there has been some progress with the introduction of the MTBF approach (see Chapter Five), especially in the FY08 budget submission. Following these steps, the PC continues to play a further role in ADP inter-sectoral resource allocation. The PC issues separate budget guidelines with instructions for line ministries to allocate resources among various development programs/projects, which are then scrutinized by the PC which has the final say in ADP allocations. 192. The dominant role of the CFAs and the more technical nature of inputs at this stage of the budget process limit the space for political bargaining and shield the resource allocation process from the injection of individual political preferences. Inter-sectoral allocations are discussed with the entire cabinet much later in the budget preparation process, when the draft annual budget is presented in April-May, thus limiting the potential for bargaining in the early stages. Furthermore, there is no 47 Budget ceilings are prepared for all 19 LM that are part of the MTBF process, thus for about 65 percent of the budget resources. The rest of the ministries still follow the old process of budget preparation. However, they are a smaller share and will very soon be part of the same integrated process. 75 parliamentary role at this initial stage as the budget is usually presented to the parliament at the very end of the process (first week of June of the current budget year for approval by the end of June) in a format that does not yet clearly highlight policies and programs that are being financed and thus does not facilitate policy analysis. In addition to not having any qualified staff to review and scrutinize the budget, the parliament is also allowed very little time (about three weeks) to review, amend and approve the budget. As a result, the parliament usually does not alter the sectoral allocation of resources presented by the executive. 193. Line ministry participation in the budget preparation process needs to be analyzed at two different phases, before and after the introduction of the MTBF. Before the MTBF approach (prior to 2004), participation by line ministries in priority setting was limited mainly to providing data and information to MoF and the PC. The extent of consultations, however, between CFAs and the line ministries/agencies appears to have been broader than in most other developing countries. This reflects, to some extent, the lack of flexibility and autonomy of the line agencies to make any changes to the allocations during budget implementation, thus forcing more consultations upfront. 194. With the introduction of the MTBF, the participation of pilot line ministries in the budget preparation process and priority-setting has been extended significantly and reasonable steps are taken to prioritize spending categories within the budget ceilings set by the MoF. A series of consultations take place between MoF and PC and the line ministries between September and February of the current fiscal year. These consultations occur in two stages. The first consultation takes place in October, when indicative budget ceilings are prepared and distributed to line ministries, while the second consultation takes place around March when the ceilings are firmed and final LM submissions are sent to MoF. In addition, LMs are consulted by the PC in the allocation of the ADP resources. 195. In addition to the core dimensions described in paragraph 187, institutional arrangements for reaching allocative efficiency are more effective when the budget is less autonomous and more integrated, with the senior level officers being capable and having long-term horizons. In these respects, as we have seen in the previous section (paragraph 180), Bangladesh compares well with other developing countries. The effectiveness of allocative efficiency institutions also depends, however, on: (i) the comprehensiveness of the budget; (ii) accountability; and (iii) the extent of program evaluation, in which Bangladesh has a weak record.48 196. To provide a sense of how Bangladesh fares compared to other countries in the sample in terms of Effective Allocative Efficiency, the scores in Table 4-3 have been combined with the effectiveness measures. These are then charted against public expenditure on education as a share of total government expenditure for the observations in the sample (Figure 4-2).49 From the perspective of current analysis, the graph shows that the effective allocative efficiency index is relatively high in Bangladesh compared to most other developing countries, and that it is associated with a reasonable allocation of public funds to a public service such as education. A similar result, with somewhat lower relative public expenditure, can be obtained in the case of health expenditure. 48 In this case, accountability is a composite measure involving ex-post reconciliation, explicit sanctions, information availability, freedom of the press, extent of democracy, and the educational attainment of the public. 49 Given the importance of education for economic growth and the need for government expenditure to ensure broad access to education, spending a larger share of public resources on education—at least up to a point—seems to be a sign of more efficient allocation. 76 Figure 4-2: Effective Allocative Efficiency Index and Public Expenditure on Education 30 nditure Korea 90s pe 25 licEx Yemen 96-00 n) b Jordan 90s w u o lP Iran 90s USA 90s h 20 ss ota UAE 90s Tunisia 90s d fT io Algeria 90s NZ 94-98 o r Morocco 90s Bangladesh 01-06 e s% ep Brazil 95-99 r th 15 rea e itu Egypt 90s dov Turkey 90s nd e e Egypt 80s Kuwait 90s g p a 10 nEx ver Indonesia 90-97 (A atio Lebanon 90s duc 5 licE Brazil 85-94 Pub 0 0.01 0.10 1.00 Overall Allocative Efficiency Index (Log Scale) Source: Staff estimates. 197. The above analysis suggests that while resource allocation mechanisms in the budget process in Bangladesh have room for improvement, they are not the weakest aspects of the system when compared to other developing countries. It also shows that the existing weaknesses are likely to diminish as the MTBF and other improvements envisaged in the PFM program take root in the bureaucracy. Indeed, full implementation of such programs can raise the effective allocative efficiency index in Bangladesh from 0.21 to 0.38, making it quite comparable with the better systems in the developing world. 198. Moreover, from a political economy perspective, Bangladeshi politicians have usually been keen to maintain the prioritization of budget allocations at the broad program level because budget shares allocated to key social programs such as health, education and rural infrastructure are visible to voters. In the transport sector, for example, the largest share of the money has been directed to road expansion despite the continuous need for maintenance of the existing road network (Chapter Three, Volume II). However, the selection process that screens projects to the implementation stage is far less observable and can be used more effectively for rent distribution. As a result, political leaders might not have much incentive to strengthen the use of full cost calculation and a sophisticated cost-benefit analysis in the project selection or program evaluation processes, an issue that has been raised repeatedly by the reviews of the ADP, and will be discussed in the following section. 4.3 Institutions for Operational Efficiency 199. While the dominance of central agencies in prioritizing and allocating sectoral resources appears to be the strength of the current system, the limited participation and flexibility of line ministries in the design of the budget within their agencies is a significant weakness. Specifically, the procedures do not give the ministry officials much leeway in designing and allocating their budgets to achieve their organization's objectives based on their own observations. These weaknesses manifest themselves in poor operational efficiency at the sector/agency level in particular, as was observed in the sectoral analysis in Chapter Three. 77 200. The key institutional arrangements needed for operational efficiency are agency autonomy and merit-based recruitment and promotion. Personnel policy is important because it determines the ability of the bureaucracy to attract and motivate competent staff capable of implementing policies efficiently. Agency autonomy further enhances efficiency because it allows executive units to respond to the specific conditions under which they carry out their tasks. To be effective, of course, autonomy must be combined with accountability and reasonable assurance of resource availability and time horizons for agency heads. Accountability itself requires performance measurement, competitive procurement procedures, audit, and explicit sanctions, with opportunities for the public to access such information and provide feedback. For each one of these categories, an index is developed, which are then combined to generate indexes for the composite categories, as in the cases discussed in the previous sections. Table 4-4: Indexes of Institutional Arrangements for Operational Efficiency Merit-Based Country – Period1 Arrangement Scores Recruitment/Promotion Autonomy of Line Agencies Bangladesh 0.35 0.43 0.27 Algeria 0.49 0.44 0.54 Egypt 0.34 0.24 0.43 Indonesia 0.70 0.62 0.78 Iran 0.37 0.49 0.25 Jordan 0.51 0.48 0.54 Kuwait 0.51 0.33 0.69 Lebanon 0.48 0.44 0.52 Morocco 0.52 0.70 0.33 Yemen 0.37 0.48 0.25 UAE 0.43 0.63 0.23 Tunisia 0.67 0.63 0.70 Brazil 1995/9 0.63 0.75 0.51 Turkey 0.38 0.38 0.38 New Zealand 0.83 0.82 0.84 Korea 0.63 0.85 0.42 USA 0.66 0.69 0.63 Source; Staff estimates 1 Note: The period is the 1990s unless otherwise indicated. - Indicator values are from 0 to 1 with 0 being the lowest. Higher indicators show better performance. - Data are based on a survey of budget experts (local and international) in Bangladesh conducted in early 2008 (see Annex v for the analytical and methodological framework). 201. Table 4-4 provides indexes that are intended to show how well institutional arrangements create the potential for operational efficiency in Bangladesh. The summary results for autonomy of line agencies indicate a major weakness in that respect, reflecting the high degree of centralization and the limited room for maneuver to reach goals efficiently at the agency level based on the local information observed during implementation. This is an old problem that has remained unresolved despite its severity (Zafarullah 1998). The merit-based characteristics of recruiting and promotion in the civil service also seem to be seriously compromised by the use of political and other criteria (World Bank 2006). 202. The effectiveness of whatever agency autonomy and merit-based recruitment/promotion exist in the Bangladesh bureaucracy is further undermined by a lack of economic/institutional characteristics that are needed for the proper operation of those mechanisms. Table 4-5 shows the indexes produced for such characteristics and compares them across countries in the sample. Perhaps 78 the two most striking weaknesses in this respect are the low salaries of civil servants and the limited resources available to line agencies for their day-to-day operations. Accountability of agency managers is also quite low, with a lack of independent and effective agencies to carry out financial and performance audits undermining the accountability of administrators. The relatively short tenure of agency managers also adds to the intensity of the problem. Finally, the fact that the general level of education in the country is relatively low is likely to make it difficult to recruit good managers and to ensure that the agencies work with a relatively well-informed clientele. Table 4-5: Indexes of Effectiveness Characteristics Affecting Budget Discipline (Intertemporal Efficiency) Educational Attainment Average Tenure of Competitive- Resource Availability of the Public Predictability of Line Agency ness of for Day-to- Day Account- 1 Country – Period Officials Resource Flow Managers Salaries Operations ability Bangladesh 0.30 0.59 0.50 0.10 0.33 0.41 Algeria 0.39 0.66 1.00 0.92 0.69 0.54 Egypt 0.41 0.71 1.00 0.33 0.77 0.24 Indonesia 0.52 0.51 1.00 0.10 0.42 0.27 Iran 0.45 0.62 0.75 0.20 0.92 0.32 Jordan 0.63 0.51 1.00 0.50 0.77 0.32 Kuwait 0.66 0.79 1.00 0.67 1.00 0.29 Lebanon 0.96 0.66 1.00 0.58 0.77 0.41 Morocco 0.42 0.79 1.00 0.33 0.83 0.36 Tunisia 0.42 0.77 0.75 0.75 0.60 0.35 UAE 0.58 0.77 1.00 1.00 1.00 0.35 Yemen 0.35 0.53 1.00 0.10 0.08 0.18 Brazil 1995/9 0.48 0.88 1.00 1.00 0.77 0.66 Turkey 0.46 0.51 0.75 0.15 0.60 0.45 New Zealand 0.91 0.99 1.00 1.00 1.00 0.93 Korea 0.84 0.90 0.50 0.50 1.00 0.57 USA 0.94 0.95 1.00 0.67 1.00 0.92 Source: Staff estimates 1 Note: The period is the 1990s unless otherwise indicated. - Indicator values are from 0 to 1 with 0 being the lowest. Higher indicators show better performance. - Data are based on a survey of budget experts (local and international) in Bangladesh conducted in early 2008 (see Annex v for the analytical and methodological framework). 203. Combining the above effectiveness characteristics with the scores for agency autonomy and merit-based recruiting and promotion yields an overall effective index of operational efficiency. This is a measure of the degree to which the administrative and institutional characteristics of public administration, including educational achievement of personnel, average tenure of line agency managers, competitive salaries, predictability of resource flow and agency accountability, are adequate to support effective budget management at the agency level. Bangladesh scores very low on this index when compared to other developing countries, which may help explain the low rank that the country gets in survey measures of bureaucratic quality and corruption control. The high degree of operational inefficiency may also help explain the low levels of expenditure on public services, despite the fact that the shares of such expenditure in total government spending are reasonably high (compare Figure 4-2 and Figure 4-3). 204. The assessment demonstrates that operational efficiency is closely linked to the functioning of the public administration whose reform has long been under consideration but has never materialized. The political economy analysis shows that there have been many forces that are at odds with the tendency to reform. Firstly, given the high degree of political fragmentation, the ruling politicians have not had much incentive to encourage bureaucratic autonomy and fight the system of 79 patronage and rent distribution. Secondly, civil service reform is a risky undertaking that could take a long time and its rewards are not likely to materialize and be internalized by ruling politicians within one electoral cycle. Thirdly, reforming an inefficient system with weak personnel and low pay is very costly because salaries need to be raised substantially and some employees have to be retired or laid off with adequate compensation. This is by no means an easy task, especially when the government has been under pressure to keep its borrowing low, partly due to foreign aid and partly due to creditworthiness limitations. Finally, managing civil service reform requires a great deal of care because much of the change has to be implemented by units that are themselves affected by reform and have a natural interest in blocking change and preserving the status quo.50 The slow pace of reforming the management of the ADP is a clear example. 205. The tendency to maintain the status quo in the civil service, with its very centralized structure and the perseverance of the cadre system has created a very fragmented and inefficient PFM system at the level of line ministries/agencies with negative consequences for operational efficiency (Figure 4-4). The PFM administration at the line ministry level is a collection of deconcentrated units of the CFAs, MoF and PC. It is essentially made up of divisions and personnel that report directly to their main departments in the MoF and PC. As such, there is little autonomy and little accountability of these units to the line ministry itself. Until the introduction of the MTBF, no senior official at the line ministry had a full picture of the performance of the budget resources, and there was little participation by high-level ministry officials in budget management. In addition to being inefficient, these vertical reporting lines outside the line ministry divert responsibility and undermine the system of checks and balances that can be established to minimize waste and corruption for the better management of public funds. Figure 4-3: Effective Operational Efficiency Index and Public Expenditure on Education 8 7 fGDP NZ 94-98 o Tunis ia 90s s% 6 Jordan 90s ea Yem en 90s ur Brazil 95-99 it 5 Algeria 90s nd Morocco 90s pe Egypt 80s USA 90s nEx 4 Kuwait 90s Egypt 90s io Iran 90s t Korea 90s ca Brazil 85-94 u Brazil 64-84 d 3 Turkey 90s tPublicE 2 n Lebanon 90s e Banglades h 01-06 r UAE 90s Cur 1 Indones ia 90-97 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Ove ra ll Ope ra tiona l Efficie ncy Inde x Source: Staff estimates. 50 The calculus delineated above has not prevented the incumbent politicians from privatizing many SoEs. This could partly be because some firms nationalized in the aftermath of independence were run so inefficiently and were such huge burdens on public finances that any patronage benefit would simply not be worth the cost. For another group of SoEs, the potential private buyers were part of the political networks that could internalize the efficiency gains from privatization. This logic was clearly not applicable to bureaucratic units, except perhaps for a few marginal ones. It was also not applicable to some of the largest SoEs and public banks because the beneficiaries of redistribution through those firms were diverse and compensating them for the loss of their rents would have been difficult. 80 Figure 4-4: Budget Planning, Execution and Reporting at the Line Ministry MoF Execution MoF Planning PC Planning Budget CGA Administration Planning Spending Division Division Unit Report Budget Audit Report MoF Bill CAO Line LD/PD LD/PD LD/PD Accounting Ministry and Audit Report Internal Audit Reporting Spending Spending Unit Unit Bill DAO DDO Report UAO Line Ministry Report Accounts & Audit Information Reports 4.4 Key Conclusions 206. The dominant role that the CFAs play in the management of public resources has served the country well for the purposes of achieving macro-fiscal discipline and allocating resources effectively to broad strategic priorities. However, the lack of autonomy/flexibility of the implementing agencies, accompanied by weak administrative and personnel management practices, operating in the context of a weak accountability structure, impede the quality of services they have provided to the population. More importantly, it hampers their ability to improve the quality of expenditures and services and their social benefits, essential for moving into the next development stage. 207. The nature of the budget management institutions and the role they play in the effective management of public resources have their roots in the country’s historical and political development and the nature of the electoral system that prevailed after independence. Changing these institutional structures, therefore, has not proven to be an easy task. Several rounds of expenditure reviews, sector reviews and recommendations have been prepared over the years, highlighting what needs to be done from a technical perspective to address the issues. While many improvements have taken place, the record shows that several fundamental weaknesses persist in the functioning of public financial management practices despite almost a decade and a half of reform. This indicates that a different approach is needed, one that addresses the political economy constraints highlighted above and the incentive structure for maintaining the status quo. 81 PART III LOOKING AHEAD—THE MEDIUM-TERM HORIZON 208. If PFM reforms are to be successful in achieving the desired outcomes at all three levels– particularly at the operational level–they need to motivate public office holders and politicians to strengthen their development vision and programs and take a greater stake in implementing them effectively. If political capital can be built on successes with initiatives that improve the delivery of services financed by public resources–especially in strategic sectors–some leaders may start to place a greater emphasis on achieving policy outcomes and improving service delivery as a route to success. 209. Any future strategy, therefore, to improve budget outcomes needs to focus on two main priorities: (i) introduction of a performance-oriented budget process; and (ii) broadening the constituency for change both within and outside the executive through participation and transparency. Once the budget activities and programs become publicly observable with clear input and output indicators, the incentive structure tends to change direction. Political competition is likely to reduce its rent distribution aspects and steer it to focus on delivering better results on public services. 210. The recent reforms initiated by the government, especially the introduction of performance- based management of public resources through the MTBF approach and computerization of the budget accounting systems, are important steps in this direction. Although they have made a good start, the government remains focused on serving the more narrow objectives of financial management apparatus for controlling inputs, rather than the broader management objectives of monitoring sector and program performance. There is broad recognition that this focus needs to change, but also that significant barriers remain for this to happen on several fronts. The following chapter discusses these issues in more detail. 82 CHAPTER FIVE INTRODUCING A PERFORMANCE ORIENTATION IN BUDGETING The introduction of the medium-term budget framework (MTBF) has marked a new era in Bangladesh’s public financial management reforms. Recognizing the limitations of an annual, input-oriented and fragmented exercise, the government introduced a three-year MTBF with the FY05-06 budget in four pilot line ministries.51 The MTBF was subsequently rolled out to 19 key line ministries (in FY08-09), covering over 60 percent of total budget expenditure. It is still too early for the MTBF to have had a measurable impact on performance and budget outcomes. Nevertheless, significant progress has been made in developing an MTBF that can be adapted to the particular circumstances of Bangladesh. In a difficult political and institutional environment, the MTBF is now seen by both central and line ministries as an integral element of the budget planning cycle. The initiative has had a positive influence on the incentive structure of key players, especially in line ministries. It has been supported by their senior managements, primarily because it is seen as having the potential to provide greater autonomy and empowerment to line ministries in the planning and management of their budgets. The challenge moving forward will be to build on these early successes to achieve a sustainable MTBF process that can contribute effectively to the realization of improved budget outcomes. The available “institutional space” within which to introduce further improvements in the budget management process, however, has now been filled. Taking forward the MTBF requires greater emphasis on issues of budget integration, institutional mandates, organizational reform and institutional capacity constraints. Key Policy Priorities: • Institutionalize and deepen the MTBF approach through: (i) strengthening the engagement of politicians in the process; (ii) strengthening the management role and flexibility of the line ministries; and (iii) modernizing budget execution and procurement systems. • Gradually move towards a better integrated budgeting process between revenue and development budgets with revised responsibilities for the MoF, PC and Line Ministries. • Strengthen the demand side for better budgeting through: (i) involving parliament and civil society in preparation and monitoring; and (ii) strengthening the independence and effectiveness of the role of the C&AG as key institution of accountability. 5.1 Introduction of the MTBF 5.1.1 Objectives 211. The government’s objectives in introducing the MTBF initiative were summarized in the FY06 MTBF document. These objectives fall into four groups: (i) promoting macroeconomic and fiscal sustainability; (ii) improving the linkage between policy priorities and resource allocations; (iii) establishing a system for monitoring the performance of public expenditure programs against the 51 The first pilot covered the Ministry of Agriculture, Ministry of Education, Ministry of Social Welfare and Ministry of Women Affairs. 83 attainment of planned output targets; and (iv) removing gradually the demarcation between DB and NDB spending. 212. Promote macroeconomic and fiscal sustainability. GoB’s expenditure plans should be matched to a realistic and sustainable macroeconomic framework and revenue forecast. At the same time, public resources should be allocated in a way that promotes pro-poor economic growth. Since the early 1990s, Bangladesh has achieved strong economic growth within a stable macroeconomic environment. However, the government’s macroeconomic and fiscal analysis and forecasting capacity to support the management of key budget aggregates had remained limited. Raising the level of government revenues remains a key fiscal management concern for a country where the level of revenues, at around 10.5 percent of GDP, remains among the lowest in the world. 213. Improve the link between policy priorities and resource allocations. This could be achieved by: (i) allocating resources between sectors and ministries in a way that reflects strategic priorities; (ii) encouraging more strategic budget planning by providing ministries with a medium-term forecast of resource availability so that they can plan ahead; and (iii) requiring ministries to link spending proposals to their strategic objectives and associated program priorities. With a single-year time horizon, the planning of the NDB was primarily concerned with securing the financing for continuing existing activities. While the DB had a stronger strategic focus, new activities once established had proved difficult to transfer to the NDB with the result that the DB included a substantial element of recurrent spending. The ADP suffers from inefficiencies, such as the length of time taken to complete projects as a result of inadequate resources, cost overruns and underutilization of resources. The justification provided by ministries for their budget requests was often very weak (Chapter Two). 214. Monitor the performance of public expenditure programs. A system should be established to monitor the performance of public expenditure programs against the attainment of planned output targets. Budget monitoring outside of basic financial monitoring is not well developed. The outputs of budget activities are not adequately specified, either in what is to be achieved or in the setting of attainment targets. While procedures for the ADP and DB were more developed, these were focused at the level of the individual project rather than at the level of department or agency program. 215. Improve budget integration. Budget integration could be achieved by gradually removing the demarcation between DB and NDB spending. Over the years, the distinction between DB and NDB expenditure has become blurred with significant levels of current financing for operational programs being financed from the NDB. This had significantly undermined budget transparency and contributed to a lack of focus on the planning and appraisal of NDB spending despite its central role in funding key public service operations (paragraph 99). 216. A rough assessment of Governments’ progress on its objectives of introducing the MTBF is set out below (Box-5.1). It should be emphasized that experience from elsewhere indicates that the successful introduction of medium-term budgeting reforms can normally be expected to take several years. 84 Box 5-1: GoB Objectives in Introducing the MTBF Objective Progress 1. Improving macroeconomic stability by matching expenditure and resources as projected in the MTMF.  2. Ensuring stability in budgetary management in the short term, keeping public expenditure within a sustainable limit in the medium and long term and accelerating pro-poor economic  growth at the same time. 3. Enhancing the participation and role of the line ministries in the budget preparation process.  4. Improving predictability of policy as well as funding by providing a medium-term view of policy objectives and resource availability, so that line ministries can prepare and plan for  forward expenditure. 5. Allocating resources between and within sectors/ministries according to strategic priorities.  6. Ensuring proper, efficient and effective use of resources by establishing a more explicit linkage between the NSAPR and other policy documents and line ministries’ objectives,  policies and resource allocation. 7. Gradually removing the demarcation between the development and non-development budgets by joint programming of development and non-development spending.  8. Establishing a system for measurement of performance of ministries/departments and subordinate offices, so that a clear indication of the desired output from the budget  allocations can be shown.  = initial progress made;  = objective substantially achieved. Source: Medium Term Budgetary Framework 2005/6-2007/8, page 3. 5.1.2 The Key Elements of the MTBF 217. To achieve these objectives, the MTBF process has been built around five main elements or building blocks. They include: (i) a macroeconomic framework that provides the basis for forecasting government revenues and expenditure over the coming three-year period; (ii) an analysis of key budget strategies and choices that can inform the allocation of budgetary resources between competing priorities; (iii) ministry-level budget frameworks and operational spending plans that link strategic objectives to budget priorities and output targets; (iv) resource ceilings that allow ministries to prioritize their budget plans against realistic medium-term spending limits; and (v) streamlined procedures for timely and efficient execution of ministry budgets (Box 5-2). 85 Box 5-2: Coverage of the MTBF Aspect FY06-08 MTBF FY07-09 MTBF FY08-10 MTBF Macroeconomic  Macroeconomic  Macroeconomic framework Framework framework and and projection of aggregates. projection of fiscal aggregates. Budget Strategy  Strategic priorities  Strategic priorities deriving Analysis deriving from NSAPR. from NSAPR.  Limited analysis of intersectoral and sectoral expenditure priorities. Ministry Budget  Four ministries  Ten ministries  Fourteen ministries included Frameworks included (Agriculture; included (Primary and (Rural Development and Education; Social Mass Education; Cooperatives; Science and Welfare; Women Health and Family Information & Affairs). Welfare; Local Communications Technology; Government Food and Disaster Department; Fisheries Management; Environment and Livestock; Water and Forests). Resources; Communication). Resource Ceilings  Single resource ceiling  Single resource ceiling  Single resource ceiling covering covering both DB and covering both DB & both NDB & DB. NDB. NDB.  Resource ceilings reflected  Ceilings do not reflect some prioritization between any prioritization sectors and MTBF ministries. between MTBF ministries (individual ministry share of total ceilings unchanged). Budget Execution [No measures yet taken to increase level of delegated authorities to MTBF ministries] 218. From a very basic MTMF in FY06, subsequent years included a more developed discussion of the macroeconomic framework and provided forecasts of key macroeconomic and fiscal aggregates.52 The FD is responsible for developing the macroeconomic forecasts but currently does not have its own budget forecasting model. A unit within FD that has responsibility for macroeconomic and fiscal analysis and forecasting was recently established with the designation of a ‘macro group’ drawn from various departments (Economic Advisers’ Wing, the Budget Wings, and the Resource and Debt Management Wing) within the Ministry. The unit has made tremendous progress since its establishment a few months ago and strengthening its own forecasting capacity remains a priority. 52 These were closely based on the IMF forecasts. Debt servicing forecasts have been based on IMF data as the FD does not yet have a comprehensive and accurate debt management database. 86 219. Budget strategy discussions were introduced in the MTBF document very gradually and remain at an initial stage of development. The first MTBF of FY06 did not include any discussion of the government’s overall budget strategy. Subsequently, the FY07 MTBF introduced a section on revenue and debt/financing strategies and a very broad analysis of the implications of the NSAPR for public spending allocations. This showed that the cross-sectoral thematic structure of the NSAPR and the large number of measures identified did not easily lend itself to the identification of public spending priorities. The FY08 MTBF included an analysis of spending priorities at the sectoral level. As yet the MTBF has not included any analysis of cross-sectoral issues and strategies relating to wage bill management, operations and maintenance allocations, and capital spending. 220. The MTBF initiative has introduced a new strategic phase to budget planning when ministries prepare spending strategies, referred to as Ministry Budget Frameworks (MBFs), that include a set of output targets alongside budget estimates. The FD considered that the development of spending strategies at sector level was not feasible in the Bangladesh context. Not surprisingly, there has been considerable variation in the quality of the MBF submissions, particularly with the information provided at department/agency level. The MBFs are prepared during the first part of the fiscal year and submitted to the FD prior to preparation of the detailed budget proposals (Box 5-3). An edited version of the MBF is then included in the MTBF document. Considerable slippage occurred in the timetables for the development of the FY07 and FY08 MTBFs, which resulted in the work on finalizing the MBFs overlapping with the preparation of the detailed annual budget estimates. Box 5-3: Evaluation and Finalization of Ministry Budget Frameworks Evaluation of MBF Submissions This follows a two-step process:  Step One: Review of the MBFs by desk officers in the Budget Wings of FD to check for realism and consistency with NSAPR priorities and with the preliminary indicative resource ceilings. This review, which makes use of an evaluation matrix, provides the agenda for discussion at a subsequent meeting with each ministry. The review focuses on the narrative section of the MBF, the proposed internal allocations of resources between departments and agencies, and the forecast of planned revenue receipts.  Step Two: An MBF review meeting, chaired by the Secretary of FD, is then held with each of the MTBF ministries to discuss the review findings and agree on the key issues to be addressed during budget preparation. The meetings also agree on the further work required to revise and finalize the MBFs for inclusion in the MTBF document that accompanies the budget. Finalization of the MBFs Subsequent revision and finalization of the MBF submissions involves:  Ministries providing the additional information and clarification requested at the review meetings, and revising their draft MBFs. This should be undertaken during January before ministries start the detailed preparation of their annual budget estimates.  Ensuring the MBFs reflect the conclusions of the subsequent budget discussions and editing for inclusion in the MTBF document. This is led by the Budget Wings in FD and takes place in late April/early May. 221. Establishment of three-year resource ceilings has been one of the most important achievements of the MTBF approach towards improving fiscal discipline and better integration of the NDB and DB. Budget ceilings, which were first introduced in FY06, are set at ministry level and are not disaggregated from broader sectoral ceilings. Resource ceilings are applied only to the ministries preparing MBF submissions and are issued twice during the budget preparation cycle. Preliminary 87 indicative resource ceilings are provided to ministries as part of the First Budget Call Circular (BCC1) covering the preparation of MBFs. These are based on the outer-year forecasts from the previous MTBF with the ceilings rolled forward for a further year. Indicative resource ceilings are provided to ministries in the Second Budget Call Circular (BCC2) that requests the preparation of detailed annual budget proposals. These ceilings take into account the updated macroeconomic and fiscal framework and the conclusions from the evaluation of the ministry MBF submissions. For the FY08 MTBF, the ceilings reflected some prioritization between sectors. There is, as yet, no explicit methodology applied in determining ceilings that, for example, takes into account the impact of centrally made decisions on pay increases or subsidies, or of the financing requirements of ongoing major infrastructure projects. 222. In introducing the MTBF, the FD acknowledged that complementary reforms were required to facilitate the timely and efficient execution and implementation of ministry budgets. The MTBF ministries emphasized that the greater flexibility they were being allowed in the planning of their budgets needed to be matched by a relaxation of what they saw as inappropriate limitations on their authority in implementing the budget and excessive bureaucracy. Their main areas of concern were: (i) delays in release of funds, particularly for the fourth quarter; (ii) the length of time taken to get funding reallocations approved when these have to be authorized by both the PC and FD; (iii) restrictive limits on ministries being able to make certain types of expenditure and other financial authorizations without prior sanction from the PC and FD; and (iv) the significant delays that occur in obtaining required authorizations from PC and FD. So far, little progress has been made towards addressing this agenda. However, its importance is recognized by FD, both for facilitating more timely execution of the budget and for ensuring continued ownership of the MTBF reforms by ministries. 5.2 Achievements and Impact of MTBF on Incentives 223. Although at an early stage, the introduction of a more strategic and performance-oriented approach to budgeting such as the MTBF, has recorded a number of achievements. It is showing a positive impact in a number of areas, including improvements on the demand side for better budgeting within the executive through: (i) engaging higher-level officials in the budgeting exercise; (ii) motivating LMs to prepare better budget submissions; and (iii) highlighting the need for better budget execution and procurement systems. The introduction of the MTBF has introduced a strategic phase to budget development that takes place during October-December and which focuses on the implications of strategic policy priorities for the allocation of resources under both NDB and DB (Figure 5-1). 88 Figure 5-1: MTBF and Annual Budget Preparation Parliament Approved Approved by Cabinet Cabinet/NEC by NEC Sent to Parliament Draft Budget Call Finalisation of Macro - MTBF & Finance Circular 1 Fiscal Framework Budget Division (Preliminary Indicative Indicative Updated Fiscal Forecast Resource Budget Call with Planning Resource & Budget Strategy Estimates Ceilings Circular 2 Commission Ceilings) Review Editing of Ministry Budget Frameworks Ministries / Preparation of Ministry Preparation of Detailed Divisions Budget Frameworks Budget Estimates Phase 1: September -December Phase 2: January –April Phase 3: May –June Medium -Term Budget Strategy Planning Budget Preparation and Review Budget Finalization Source: Ministry of Finance. 224. Higher level officials are more engaged in the budget process through their participation in a number of committees and working groups in central ministries (MoF and PC) and line ministries (Figure 5-2).53 Perhaps most notable is the re-activation of the Budget Management Committees (BMC) in the line ministries to oversee the MBF and budget preparation, and the subsequent monitoring of MTBF and budget implementation. BMCs are chaired by the secretaries who had little involvement with the budget before the MTBF. At a more technical level, the Budget Working Group (BWG) involves key officers involved in MBF and budget preparation from the ministry secretariat and from the departments and agencies. While these committees have been able to attract high-level officials in the budget discussions, they have not been very effective as they lack technical capacity and secretariat support. More importantly, they operate in a very fragmented financial management organizational structure within the line ministries and have, to some extent, substituted for that. In the medium term, however, building a sustainable and inclusive MTBF process will require greater attention to institutionalization of this structure and better integration of financial management functions within the line ministries. 53 The BMRC, which is chaired by the Minister for Finance, oversees the MTBF process. The Medium-Term Budget Coordination Technical Committee, chaired by the Secretary of Finance, exercises overall coordination over MTBF preparation. Two technical working groups-the MTMF Working Group and the Resource Ceilings and Estimates Evaluation Working Group --provide support to the FD with the preparation of the MTMF analysis and the development of the budget strategy and spending proposals. 89 Figure 5-2: MTBF Committees and Technical Working Groups CENTRAL MINISTRIES/AGENCIES LINE MINISTRY Budget Monitoring and Resources Cttee. (BMRC) Medium-Term Budget Coordination Technical Cttee. (MTBCTC) Budget Management Cttee (BMC) Medium-Term Macro Resource Ceilings and Framework Working Estimates Evaluation Budget Working Group Working Group Group (MTMFWG) (RCEEWG) (BWG) Finance Division - Budget Wings SAS (Budget), Planning Cell 225. The introduction of the MTBF approach has been well received by line ministries and it has made possible the introduction of unified budget ceilings (for DB and NDB) at the strategic phase, where line ministries are gaining some flexibility to allocate resources among projects and activities. The approach is seen as having brought about a more strategic and forward-looking perspective and is encouraging ministries to plan their programs for the medium term and to link policy priorities to resource allocations. It has allowed ministries to plan their budgets as a whole and given them greater control over how budget resources are allocated. It has also enhanced the role of ministry secretariats in the budget process from that of forwarding departmental budgets to the FD and PC to being actively engaged in determining budget priorities and resource allocations. 226. The interaction with the FD in finalizing the draft budget proposals is seen as more constructive. However, the ministries, while strongly supportive, have indicated some concerns about the present state of development of the MTBF process. These include: (i) the complexity of the MBF exercise and the tendency for the formats for MBF submissions to be changed each year; (ii) the continued requirement to provide a separate submission to the PC for the ADP despite this submission asking for the same information as the one included in the budget guidelines submission; and (iii) the slow progress being made in addressing requirements for parallel reforms to provide for greater delegation of authority in budget implementation procedures (Box 5-4). 227. The proper preparation of the MTBF has been hampered by data limitations as comprehensive data on spending by economic item is not currently available from the government accounts. Analysis of public spending at sector and subsector levels is hampered by the absence of a program classification (or of a crosswalk table for mapping departmental/agency spending to programs). The budget release process does not allow for enough flexibility in the implementation of programs. More importantly, the introduction of the MTBF has highlighted the ineffectiveness of the financial management functions at the line ministries, where there is no single authority in charge of managing and overseeing the budget. Furthermore, for a proper implementation of the MTBF, the budget and procurement cycles will need to be much better integrated. Current practices often hamper the finalization and implementation of the procurement contracts that usually extend beyond the annual budget. 90 Box 5-4: Line Ministry Reactions to the MTBF – the Local Government Division A recent World Bank study to assess requirements for further deepening of the MTBF process in the water and sanitation sector highlighted the very positive response to the MTBF initiative in the LGD. Senior officials spoke knowledgably about the underlying principles of the MTBF approach and of the opportunity that it provided for forward planning. The resource constrained approach of the MTBF was seen as a positive rather than restrictive development. Individual comments emphasized that:  The MTBF initiative was empowering because “the Ministry can take part in budget preparation and planning well ahead…and can decide where to put the thrust of its expenditure in terms of the ideas of the ministry and the priorities of the NSAPR”.  The MTBF “had built bridges both between FD and LGD and between LGD and its departments/agencies. Before MTBF, ministries just made its budget submission to FD and had it approved, or not, without discussion. The rules of business have been streamlined”.  Another senior official in LGD was “totally convinced of the process” emphasizing the need to move away from a project culture, which decreases the capacity of government since it separates the regular institutions of government from implementation.  In the Department of Public Health Engineering, the MTBF was seen as providing the “liberty to fix priorities” and “to distribute money where needed”. This was particularly important as priorities were moving to sanitation, water quality, arsenic control and capacity-building, whereas previously they had mainly been concentrated on reducing levels of arsenic in drinking water supplies. These comments mirror the feedback obtained by the FD from other ministries and also comments made to the PEIR mission. 5.3 Building a Sustainable MTBF: The Anchor for PFM Reforms 228. The initial progress with the introduction of the MTBF has been promising and the implementation of reforms has received considerable support from the line ministries and other stakeholders. However, this initial phase of reform has filled the institutional space that existed to introduce reforms in the budget management process. Further development and institutionalization of the MTBF will require more fundamental reforms to institutional mandates, organizational structures and processes for budget planning and management. It will also be important to ensure that the necessary incentives are in place to maintain the reform’s momentum and ensure further development and institutionalization of the MTBF procedures. 229. Further progress will require that the MTBF reforms are perceived by stakeholders as bringing about improvements in budget planning and management. Particular attention will be needed in: (i) strengthening the management role of line ministries and gradually giving them more autonomy; (ii) ensuring that the FD is responsive to the feedback from line ministries, particularly in providing predictability of resources and streamlining budget execution procedures; and (iii) reducing the present duplication in budget planning and management procedures. Ultimately the sustainability of the MTBF reforms will depend on the continued sense of ownership by key participants and stakeholders. This applies to the central ministries and agencies (principally FD and PC), to the line ministries and at the political level. 5.3.1 Strengthening Political Engagement in the MTBF/Budget Process 230. Political engagement at key points in the MTBF and budget planning process can play an important role in embedding the MTBF reforms by mainstreaming the analysis undertaken for the MTBF into the government’s decision-making processes (Box 5-5). In Bangladesh, strong political 91 engagement in the MTBF and budget has yet to be achieved. Two key priorities will need to be addressed: (i) approval of the MTBF resource ceilings at the highest decision-making level of the government; and (ii) consultation at the political level at the outset of the MTBF/budget preparation process. 231. The MTBF calendar provides for National Economic Council (NEC) approval of the indicative resource ceilings–an important decision point in the budget process at which critical choices have to be made. In practice, the involvement of NEC has not yet been achieved and instead the ceilings have been approved by the Minister of Finance. Starting with the FY10 MTBF, the NEC should discuss and approve the indicative resource ceilings. The ceilings proposal should be accompanied by the Medium Term Budget Outlook (MTBO) as the strategy paper, setting out the preliminary macro-fiscal framework and identifying the priorities to be addressed under the coming MTBF and budget. 232. The experience of other countries suggests the desirability of having an earlier political-level consultation at the outset of the MTBF/budget preparation process in order to help achieve consensus around key medium-term budget issues and budget strategies. As the MTBF process matures further, consideration should be given to holding an earlier discussion with the NEC about the key issues for analysis during preparation of the coming MTBF and budget. This could be facilitated by the preparation of an MTBF concept paper. Box 5-5: Political Engagement in the Budget Process – the South African Example South Africa’s MTBF and budget process involves a very high level of political engagement at key stages in the MTBF and budget planning process. The Ministers’ Committee on the Budget (MinComBud), a cabinet subcommittee, meets regularly through the year to oversee the process. Cabinet is also directly involved at key decision-making stages.  The FY in South Africa starts in April. Treasury (the finance ministry) initiates the preparation of the next MTBF and budget immediately after the start of the new FY. Between April and September, it undertakes the strategic analysis which focuses on budget prioritization, development of the overall budget framework, and division of revenue between national, provincial and local governments. During this period there are regular meetings of MinComBud.  An extended meeting of cabinet at the end of September agrees on the strategy. The Medium-Term Budget Policy Statement (MTBPS), the MTBF document, is then prepared and tabled in parliament at the end of October. The MTBPS sets the strategic context and directions for the development of the budget for the next FY.  The Treasury then prepares the final budget ceilings for national votes, and provincial and local governments. These are reviewed with MinComBud and approved by cabinet in early November.  Thereafter, the national ministries prepare their final submissions and the national budget and estimates are finalized. The budget is approved by cabinet and tabled in parliament in February. 5.3.2 Strengthening the Management Role and Flexibility of Line Ministries– a Priority for Improving Operational Efficiency 233. The introduction of the MTBF has provided line ministries with greater discretion in the setting of their budgets and in prioritizing competing spending demands against their resource ceilings. The institutional assessment (Chapter Four) showed that autonomy of line ministries/spending agencies enhances operational efficiency as it provides them the flexibility to respond to the specific conditions under which they carry out their tasks. At the same time, a more flexible approach such as the MTBF has placed additional demands on them with regard to budget planning and management. If 92 these new processes are to be sustainable, they should be seen as contributing to more effective decision making on resource allocation and to facilitating subsequent implementation of the annual budget and better budget outcomes. 234. It will be important to look beyond these initial reforms to ensure that the necessary institutional framework and incentives are in place to sustain the engagement of line ministries in the MTBF process: • MBFs should play a more important role in the budget planning and decision-making process. Preparation of the MBFs represents a significant new task for line ministries involving additional workload demands. However, the MBFs currently play only a limited role in the budget discussions and decision-making process. While the finalized MBFs form part of the MTBF document, they do not, as yet, play an integral role in the presentation of the ministry’s budget to parliament. ∼ The FD should, at an appropriate time, review the content and format of the MBF submissions to ensure that the MBFs contribute more effectively to the MTBF and budget discussions and decision making. ∼ As the MTBF process further matures, consideration should be given to submitting the MBFs for approval by the BMRC prior to the issuing of BCC2 (second budget preparation guideline). This would help to raise the profile of the MBFs within the budget decision- making process. ∼ The MBFs should subsequently form the basis for an improved presentation of the annual budget (see paragraph 258 below). • MBFs need to be more integrated with the internal ministry budget processes. Ultimately, the success of the MTBF initiative will depend on its integration with the internal budgeting and budget management processes at departmental/agency level and below. This goes deeper than the level at which information is provided by the ministry in the MBFs. The recent study into the implementation of the MTBF in the water and sanitation sector has revealed considerable interest in seeing how the MTBF initiative should be rolled out further within LGD to the level of its subsidiary budget agencies. ∼ The FD should review with line ministries the implications of the MTBF initiative for internal budget planning and management processes at department and agency level. A key issue here is how to drill down the ceilings process at departmental level linked to the realistic costing of the major service delivery programs. ∼ FD should organize training for departments and agencies on the deepening and strengthening of MTBF and budget planning processes. • Complementary reforms to budget execution processes need to be addressed. Line ministries continue to emphasize the importance of complementary reforms to budget execution processes. Of particular concern is what they consider to be excessive central control over even minor reallocations between budget lines and the length of time taken when approval has to be sought from the FD and PC. Addressing this issue will be critical to ensuring the continued involvement and support of line ministries to the MTBF initiative. ∼ The FD should undertake a study of the extent to which fund release procedures, re- appropriation procedures and financial limits are too restrictive and develop proposals for reform. The study should consider the requirements for: (i) revising budget release requirements and procedures in order to strengthen cash flow management while facilitating timely implementation of the budget; (ii) eliminating the requirement for dual 93 approval of reappropriations on the DB; and (iii) raising financial limits on delegated authority and eliminating those limits which are no longer appropriate. ∼ FD and CGA should complete the computerization of Treasury and accounting processes in order to improve the quality and timeliness of the budget data that is used for planning and evaluation. ∼ In the longer term, the FD should consider the scope for delegating additional financial powers to those ministries that meet benchmark PFM standards with the aim of moving from ex-ante controls towards accountability backed up by strengthened ex- post inspection and audit. ∼ Both FD and PC should introduce targets for maximum turnaround time for handling requests for reappropriations and financial authorizations, and monitor performance against these targets. ∼ Both FD and PC should work on a new budget classification structure in tandem with the computerization reforms. • Complementary measures are needed to integrate the budget and the procurement cycles, which are currently disconnected. Contracts that cannot be signed during the stipulated fiscal year lose their financing in the next cycle. Current practices of budget management within the annual cycle do not accommodate the realities of the procurement cycle; yet procuring entities are held liable and have to struggle hard to get the budget reallocated in order to sign contracts. A similar experience avails to those that may have signed contracts and are under implementation. Addressing these issues will be important for the proper implementation of the budget programs, and especially the ADP. ∼ The realities of the procurement cycle and the influence of seasonality will need to be factored in while preparing procurement plans. Formal covenants will need to be included in the project documentation to accommodate cycles that are not aligned with budgeting. ∼ Consideration should be given to some changes at the start of the fiscal year, or issue regulations that allow the government to begin advanced procurement steps before the beginning of the fiscal year. ∼ Improve project planning in order to provide more accurate estimates of costs and disbursements on the project and incorporate them in the budget projections. 5.4 Addressing the Budget Integration Agenda – The Cornerstone of Improving Expenditure Outcomes 235. Measures to strengthen the autonomy of the line ministries should be taken simultaneously with measures to better integrate the two sides of the budget, DB and NDB. In the FY06 MTBF document it was stated that a key objective of the GoB in introducing the MTBF was ‘removing the demarcation between the development and non-development budgets gradually by joint programming of development and non-development spending’. The importance of moving forward on this agenda is well-known and has been emphasized by line ministries who express frustration at having to deal with two central agencies in both the preparation and implementation of their budgets. It is also at the heart of several other deficiencies in the expenditure management process that were highlighted earlier in this report (Chapter Two). It will become increasingly difficult to take forward the MTBF unless this budget integration agenda is addressed. 94 236. For several decades, the Bangladeshi budget operated through two separate and unrelated components, prepared independently by two separate central agencies–the Revenue Budget (non- development budget) by the Ministry of Finance (MoF) and the Development Budget by the Planning Commission (PC). This fragmentation has seriously constrained the effectiveness of public spending and undermined budget outcomes. The introduction of the MTBF, however, has opened a new window of opportunity to gradually address the fragmentation. The question is how the roles of the MoF, PC and line ministries should be adjusted to better serve the MTBF objectives and improvements in the performance orientation of budgeting. The immediate requirement should be to resolve outstanding issues surrounding the way the budget is prepared, presented and implemented. In the longer term, the GoB will also need to address the challenge of moving to a single integrated budget. 5.4.1 Short-Term Priority Actions 237. The most urgent issues to facilitate budget preparation and implementation relate to eliminating the duplication of procedures that currently occurs in the planning and implementation of the DB. This applies at the preparation stage where the MTBF ministries continue to be required to make separate DB and ADP submissions, even though for government-funded projects these should be one and the same. It also applies to the presentation of the budget where there are differences in the economic categorization of expenditure used for the NDB and the DB. In the implementation of the budget, requirements for ‘double approval’ of re-appropriations and authorization of expenditures that exceed financial limits can result in significant delays. • PC should, starting with the FY10 Budget, use the DB estimates in compiling the ADP for the MTBF ministries and eliminate the requirements for line ministries to submit separate ADP submissions for these projects. ADP submissions for projects that are not financed through the DB should be sought separately. • FD should require expenditure estimates submissions for the NDB and DB to be submitted concurrently and for DB expenditures to be broken down by economic category.54 This requirement should be introduced with the FY10 budget. The FD should discuss with Economic Relations Division (ERD) and the development partners (DPs) how projects financed by direct project aid (DPA) can be shown in the budget using at least an aggregated version of the economic classification. • FD and PC should develop an action plan for eliminating the requirement for double approval of expenditure re-appropriations and authorizations during project implementation. 5.4.2 Longer-Term Challenges in Developing an Integrated Budget 238. In the long term, the PC should refocus its role on serving the objectives of the MTBF with its four objectives: (i) establishing a realistic macro framework; (ii) linking policies, strategies and resource allocations; (iii) establishing procedures for capital investment planning and management; and (iv) establishing a framework for performance monitoring. The move to a fully integrated budget will require that the DB and NDB are ultimately brought together within a single budget. An intermediate stage could be for the DB to become a capital budget with spending of a recurrent nature 54 It is already a requirement that the DB estimates show expenditure broken down by economic category. However, because the ADP is not finalized until late May, there is insufficient time to prepare such a breakdown before the DB is finalized. 95 (currently being met from the DB) being transferred to the NDB. This applies particularly to sectors, such as health and education, where DP assistance has covered significant levels of recurrent spending. Establishing a Realistic Macro Framework 239. A key objective of GoB in introducing the MTBF has been ‘improving macroeconomic stability by matching expenditures and resources as projected in the Medium-Term Macroeconomic Framework (MTMF)’.55 Achieving this objective requires sound economic policies and strategies that provide a sustainable foundation for economic growth and a realistic and prudent macro-fiscal forecast for the MTBF and budget. It is common for a number of government agencies and independent organizations to prepare macroeconomic forecasts and there are processes for consultation between forecasters and other stakeholders around the development of the MTMF. In countries such as Bangladesh, where the economic policy and PFM functions fall under separate ministries, the finance ministry usually has its own capacity to support the development of the medium-term macro-fiscal forecasts. The finance ministry will often adopt a more conservative scenario than other forecasters in order to facilitate prudent management of the budget. 240. The arrangements established in Bangladesh reflect good practice, with the FD responsible for the development of the macro-fiscal framework and processes in place for consultation with other stakeholders. The main challenges are: (i) to build the professional capacities in FD for macro-fiscal analysis and forecasting; and (ii) for the PC to develop its own policy analysis and forecasting capacities to support its long-term strategic policy management role. There is also scope for further strengthening and expanding the consultation process around the MTMF to include independent institutions involved in macroeconomic forecasting. Ensuring Linkage between Policies, Strategies and Resource Allocations 241. The GoB has emphasized the importance of developing a policy-focused budget planning process that ‘ensures proper, efficient and effective use of resources by establishing a more explicit linkage between the NSAPR and other policy documents and line ministries’ objectives, policies and resource allocations.’ This is seen as requiring removal of ‘the demarcation between development and non-development expenditure gradually by joint programming of development and non-development spending’. The introduction under the MTBF of a strategic phase to budget preparation is resulting in a policy-driven approach to budget planning which reflects international good practice. 242. There remains, however, a lack of coherence between the processes for planning and management of strategic plans, the MTBF and the ADP. This will need to be addressed if the goal of an integrated policy-led budget process is to be achieved. This will require that: (i) the new procedures for linking planning and budgeting under the MTBF are backed up by strengthened capacities for policy development and management in PC and in LMs; and (ii) FD builds its capacity to undertake budget policy analysis and review budget submissions from a policy perspective. These changes will require closer integration between economic planning and budgeting skills in both the FD and LMs. The move towards a comprehensive budget process covering both the DB and ADP also implies reform of the ADP and, in particular, the phasing out of those procedures that duplicate the LM budget preparation procedures under the MTBF. Building Robust Procedures for Capital Investment Planning and Management 55 This and subsequent references are taken from the list of objectives listed on page 3 of the Medium-Term Budget Framework 2005/06- 2007/08. 96 243. An MTBF is usually complemented by procedures for capital investment management that provide an interface between the MTBF/budget cycle and the project cycle. There are two points at which these procedures coincide: (i) when the project concept has been identified and the decision is taken on whether to proceed to design and appraisal, and for it to be included in the pipeline of projects for funding in the outer years of the MTBF; and (ii) following a positive appraisal when the project can be selected and approved for funding under the next year’s budget. An important aim of these capital investment management procedures is to ensure that the initial screening and subsequent approval of investment projects are based on technical and economic appraisal criteria and protect against inappropriate political interference. 244. Under the MTBF reforms, Bangladesh is moving towards a unified budget that distinguishes between investment and recurrent spending rather than between development and non- development spending. The budget process for both capital and recurrent spending should be overseen by the FD. This means that the budget preparation aspects of the ADP will become redundant and should instead be replaced by improved procedures for capital investment management focusing on the identification, screening, preparation, appraisal, selection and approval and monitoring of capital investment projects. In reviewing the existing procedures, one objective should be to promote a more decentralized approach to investment management that mirrors the MTBF reforms.56 The financing of capital investment projects, once approved, should be handled through the budget process, but subject to good practice requirements, such as ensuring that ongoing capital investment projects are fully financed, before new projects are considered for inclusion in the budget. 245. The location of institutional responsibility for overseeing capital investment management should be revisited. Locating responsibility within the finance ministry has some advantages in emphasizing the link between capital investment management and the budget process. However, there are examples where responsibility has been successfully located in an economic planning ministry, but closely coordinated with the budget process. Establishing a Framework for Performance Monitoring 246. In introducing the MTBF, GoB has set out its aim of ‘establishing a system for measurement of performance of ministries/departments and subordinate offices, so that a clear indication of the desired output of the inputs provided through the budget can be shown’. In the preparation of the MBFs, considerable emphasis has been given to the identification of outcome indicators for measuring the realization of ministry strategic objectives, and output indicators through which the performance of ministry spending programs can be assessed. Not surprisingly, ministries have faced considerable difficulties in elaborating appropriate performance indicators and this remains very much a work in progress. 247. Performance monitoring arrangements for the MTBF are still at an early stage of development. These arrangements will have implications for the respective monitoring roles of both the PC and FD. Performance monitoring in the PC should focus on policy outcomes and the implementation of the strategies through which policy objectives are to be achieved, while performance monitoring in the FD should focus on the monitoring of budget outputs and the activities through which these outputs are to be achieved. Both the policy and budget monitoring processes should be well- linked to allow for the sharing of information and prevention of duplication. This is important since 56 An initial step was taken in March 2007 when the PC issued new guidelines on project preparation, processing and approval. These emphasized a broader approach to appraisal within the context of national and sectoral priorities. They also emphasized the assessment of the sustainability of proposed investments and their downstream recurrent costs. 97 sector policy and strategies should guide the design of ministry spending programs, while budget outputs contribute to sector strategy implementation and the realization of policy objectives. 5.4.3 Institutional Roles and Capacity-Building Requirements 248. The previous section highlighted the requirements for closer integration between policy planning and budgeting processes under the MTBF. It addressed the need to build economic planning skills capacities, particularly in the FD and LMs, in order to support the further rollout of the MTBF initiative. It also noted that existing economic planning skills capacities in both the FD and LMs have not been sufficiently mainstreamed into the MTBF and budget planning process. In part, this reflects issues relating to the operation of the cadre system in Bangladesh (Box 5-6). Box 5-6: Cadre System Issues The way in which the cadre system in Bangladesh operates and the rivalry that exists between cadres is widely acknowledged as a significant constraint to effective public administration. The planning and budget system is staffed principally by two types of cadres: (i) the Economics Cadre from which economics and planning-related positions are mostly staffed; and (ii) the Administrative Cadre from which budget-related positions in the Budget Wings of the FD and LM secretariats are staffed. Cadre issues contribute significantly to the lack of integration between planning and budgeting processes. This occurs in two main ways:  First, frequent transfers of staff in the Administrative Cadre to positions not related to the budget process significantly undermine efforts to develop an experienced and well-trained group of budget specialists in the FD and the LMs. While in recent years the FD has had some success in retaining key personnel, the lack of continuity in staffing remains a serious issue, particularly in LM secretariats.  Second, it has contributed to separation between economic planning and budgeting functions, the former being staffed from the Economist Cadre and the latter from the Administrative Cadre. This is reflected in the organizational structure of the FD and LMs, where economics cadres’ positions are located in separate units and are not sufficiently mainstreamed into the core work processes of the ministry. The MTBF reforms imply the breaking down of this separation with the creation of an integrated planning and budgeting function. The proposals made in this paper require at the very least that officers from the Administrative and Economics Cadres should work alongside each other within integrated planning and budgeting structures in both the FD and LMs. This should, however, be seen as a stop-gap measure pending a fundamental review and reform of the cadre system as a whole. A small number of senior positions in the FD are filled by officers from the Audit and Accounts Cadre. These officers have been assigned to the Senior Service Pool and, although they remain members of the cadre, they do not have the right to return to cadre posts. 249. This section brings together the main conclusions from the preceding analysis to provide recommendations regarding the implications of the MTBF for the roles and capacity-building requirements in the PC, FD and LMs. It should be emphasized that there are many different ways in which these issues can be addressed and that the proposals made here should, therefore, be seen as a point of departure for further discussion and elaboration. 98 The Planning Commission 250. The move towards a unified budget under the MTBF implies a significant change in the participation of the PC in the budget process. This should emphasize a higher-level role that reflects the primary responsibility of the PC in overseeing the management of the GoB’s policy and strategy agenda. At the same time, the PC should become less engaged in the detail of budget planning and monitoring. 251. In relation to the MTBF process, four main functions of the PC can be highlighted: • Policy coordination: The PC has the lead responsibility for developing and managing the GoB’s overall national development strategy and for ensuring that a robust set of cross- sectoral (for example poverty, gender and environment) and sector policies/strategies are in place. While the PC does not have primary responsibility for sector policy development, it has an important role in supporting the policy development process, in reviewing policy and strategy proposals and in ensuring their consistency with overall macro-level policies and strategies. It should also ensure that fiscal realities are taken into account during policy development and that policy initiatives are framed in such a way that their budgetary implications can be clearly spelt out. • Oversight of the capital investment program: The PC could continue to oversee the identification, appraisal and approval of projects in the GoB’s capital investment program. It could also undertake monitoring and evaluation of large and complex investment projects and of the capital investment program as a whole.57 Procedures should be made more rigorous so that capital investment projects are only considered for approval for financing from the budget when they have been fully prepared and subject to an appropriate level of appraisal. • However, responsibility of the PC for overseeing the preparation of the budget estimates for capital investment projects and the subsequent management of capital budget execution should be transferred to the FD. These changes would require substantial reorientation and rationalization of the role and functions of the Programming Division. • Monitoring policy and strategy implementation: The PC should oversee the monitoring and evaluation of the implementation of GoB policies and strategies. This responsibility can be considered as comprising two core elements: (i) undertaking monitoring of a limited number of performance indicators linked to the realization of the GoB’s strategic objectives and major strategies/programs; and (ii) providing guidance to LMs on the development of their own internal policy and strategy monitoring and evaluation systems. The PC should also develop its capacities for the evaluation of the government strategies and programs. • The present emphasis of monitoring in the PC–which has been on ADP implementation– should be shifted to the FD as part of the integrated budget monitoring procedures being developed for the MTBF. It should be replaced by a strengthened strategic monitoring process that focuses on policy and the measures necessary to achieve those outcomes. This requires a shift away from very detailed monitoring of activities towards monitoring a small number of strategic performance indicators and to ensuring that adequate procedures are in place in LMs for more detailed monitoring of sector policy outcomes and strategy implementation. 57 An alternative would be to transfer management of the GoB’s capital investment to the FD with the PC being responsible for determining the broad intersectoral priorities within which the capital investment program is developed. However, taking into account existing mandates it has been assumed in this note that the PC will retain primary responsibility for oversight of the capital investment program. 99 • Supporting the policy development and monitoring function in LMs: The PC has an important role in specifying these new functions and in supporting capacity-building and training for the LM staff responsible for carrying them out. The MTBF is creating new demands on the planning cells in the LMs and moving the focus of their role away from the ADP towards managing the strategic policy and programming and monitoring functions. 252. These changes will generate a requirement for technical assistance (TA) in specifying new and improved work processes and providing capacity-building and training support in each of these areas. However, the need for such assistance should be seen in the context of a requirement for a fundamental review of the role and functions of the PC and how it might evolve into a more strategic policy and planning agency. The findings of such a review could form the basis for the specification and design of a broader TA operation. The Finance Division 253. As the MTBF reforms are further rolled-out, the FD should take responsibility for overseeing the entire budget process. This will enable the GoB to move progressively towards a single integrated budget with the distinction between capital and recurrent spending replacing that between development and non-development expenditure. 254. Achieving this will require the FD to integrate economics and planning skills into its planning and management of the budget by: • strengthening capacities for macro-fiscal analysis and forecasting to support the development of the MTMF. As noted above, the FD has established a new Macro-fiscal Wing which includes a separate Macroeconomics Branch. The staff assigned to this branch will require strong economics skills. The new branch will require significant support in developing its work program and in introducing improved analytical and forecasting techniques. • strengthening budget policy and strategy analysis. The structure of the FD has not previously facilitated budget policy and strategy development or the analysis of cross- sectoral budget issues, such as poverty and gender impacts. The new Macro-fiscal Wing provides for this function within a separate Fiscal Branch. Staff assigned to this unit will require strong analytical and planning skills. The unit will require support in developing its functions and work program linked to the MTBF timetable. • strengthening analytical and planning capacities within the sectoral Budget Wings to enable desk officers to assess budget requests and performance from a more strategic policy and program delivery perspective. This will require stronger economics/planning skills than currently exist within the Budget Wings and which could be met by a combination of re-assignment of staff from the Economic Advisers’ Wing and training of existing staff. It should also be accompanied by a more detailed review of the functions and working of the Budget Wing with the aim of placing greater emphasis on analytical work and less on transactional responsibilities. • creating a separate monitoring branch within the Macro-fiscal Wing, which will oversee the development and implementation of the performance monitoring system for the MTBF/budget. To date, the FD has played a relatively limited monitoring role and its skills in 100 this area are not well developed. Ideally, staff appointed to this new unit should have strong monitoring skills and experience. 255. These requirements emphasize the need for the FD to review its present assignment of staff to see how the new skill requirements can be most effectively met. In particular, there may be scope for re-assigning economists currently attached to the Economic Advisers’ Wing to the new Macro-fiscal Branch and to the Budget Wings. The possibility of further strengthening economic planning skills in FD through transfers from the PC, particularly from its Program Division and Implementation Monitoring Evaluation Department (IMED), could also be considered. The FD has already received substantial TA support for the introduction of the MTBF. This is expected to be continued under the proposed Multi- Donor Public Finance Management Trust Fund. Support coverage in the areas identified above should be considered as part of this operation. Line Ministry Secretariats 256. The MTBF requires a more proactive role by ministry secretariats in the planning and budgeting process. As noted above, Budget Management Committees (BMCs) are being established to provide a dedicated budgeting capacity that currently does not exist in the LMs. 257. The MTBF has a number of implications for the future role of Planning Cells: • a stronger emphasis on sector policy and strategy development and monitoring. This should also include taking the lead responsibility for preparation of the sections of the MBFs relating to policy priorities and strategies and intrasectoral resource allocation. As noted above, many Planning Cells have limited skills and experience in this area; • overseeing the implementation of the capital investment management procedures in the ministry. This would involve a shift in emphasis away from the focus on ADP preparation and monitoring; and • assisting with the assessment of MBF and budget spending proposals from a policy perspective. 258. Further deepening of the MTBF in LMs will, therefore, require mainstreaming of the Planning Cells into the broader policy and budgeting functions of the LM. One option might be to broaden the remit of the BMB into a Policy, Planning, and Budgeting Wing into which the existing Planning Cells would be integrated. This would have the advantage of bringing the existing Planning Cells more directly into the planning and management structure of the ministry, while emphasizing the very close linkage between planning and budgeting skills. 5.5 Strengthening the Demand Side for Better Budget Outcomes 5.5.1 Involving Parliament and Civil Society 259. MTBFs have also been used as a vehicle for stronger consultation with parliament and civil society on key budget issues. This has significantly increased the transparency of the budget process. To date, the focus in Bangladesh has been on building understanding of the MTBF reforms within government. Engagement of external stakeholders on the MTBF has been limited and has not been directly sought by the FD. For example, there has been little coverage of the MTBF document in the 101 press. Discussion between FD and business leaders and civil society has been limited to the final stages of the budget preparation and has, therefore, not significantly influenced the planning of the budget. 260. With the MTBF becoming firmly established as an integral part of the budget planning cycle, the opportunity exists to use it in providing a platform for an improved dialogue with parliament and civil society as part of the annual budget planning process, in particular by: • enhancing parliamentary discussion of the MTBF. Following the experience of other countries it would be desirable to refocus parliamentary discussion of the budget towards issues of strategy and performance. ∼ The FD should mainstream the MTBF document into the presentation and discussion of the annual budget to parliament. This applies both to the presentation of the MTBF and the MBFs. As noted above, the MBFs could, in time, provide the basis for an improved presentation of ministry spending estimates. ∼ In the longer term, the FD should consider whether it would be appropriate to make a presentation to parliament on the MTBF at the conclusion of the strategic phase of MTBF/budget development. The purpose of such a presentation would be to inform parliament of the basis on which the budget is being developed. This is common practice in a number of other countries.58 • engaging in a dialogue with civil society. Since the MTBF provides a more strategic presentation of the government’s spending plans, it can be used as a basis for enhancing dialogue with civil society on budgetary issues and, more generally, for increasing the accountability of the budget process. ∼ As the MTBF becomes further embedded, a citizens’ version of the MTBF should be prepared and launched concurrently with the budget. This should set out the MTBF and budget proposals in a form more relevant and meaningful to the general public. ∼ The FD should also consider the scope for consulting with civil society as part of the MTBF and budget process. This could most appropriately be done at the beginning of the financial year so that the findings and conclusions can be incorporated into the subsequent development of the MTBF proposals for the coming three years. 5.5.2 Improving the Presentation of the Budget and Increasing Transparency 261. The introduction of an MTBF is commonly associated with improvements in the presentation of the budget and typically involves two main changes. These are: (i) an improved presentation of the overall budget strategy and priorities; and (ii) changes to the format of ministry budget estimates involving an improved explanation of the activities to be undertaken and of the related outputs. Bangladesh’s MTBF document addresses both of these requirements. However, rather than replacing some of the existing budget documentation, it has provided an additional presentation of the budget, resulting in significant overlap with the existing budget documentation. • As the MTBF exercise is taken forward, the FD should consider how to improve the presentation of the MTBF and budget: 58 Examples of governments that make MTBF presentations to parliament at the conclusion of the strategic phase of the budget preparation process include: South Africa (the Medium-Term Budget Policy Statement); the United Kingdom (the Pre-Budget Report); and Sweden (the Fiscal Policy Spring Bill). 102 ∼ The MTBF documentation should in future form the basis for: (i) the main summary presentation of the budget; and (ii) the detailed presentation of ministry budgets. The format of the existing Economic Review document should be revised so that it supports and complements the MTBF document and avoids duplicating analysis that is contained in the MTBF. Consideration should be given to integrating the Demand for Grants document with the detailed presentation of ministry budgets. ∼ FD should also review the format of the MBFs to see how it can be more concise and reader-friendly. It will be important to ensure that the MBF documentation does not become too unwieldy as the MTBF initiative is expanded to cover all ministries.59 • The FD should consider making the MTBO a public document, setting out the framework and strategy for the next budget. Consideration could also be given to presenting the MTBO in parliament for information purposes. In many other countries, the conclusions of the strategic phase of MTBF/budget preparation are incorporated into a public document, setting out the basis under which the government is developing its budget plans. 5.5.3 Strengthening the Role of the C&AG 262. While a lot of progress has been made, a lot more is needed to make the C&AG the key institution of the accountability framework in Bangladesh. Audit reporting is improving both in terms of content, relevance and timeliness. However, its reports are not yet timely and not published. The Public Accounts Committee struggles to discharge its duties effectively and there is no public debate of critical financial management issues reported on by the C&AG. The reporting process followed by the C&AG remains largely symbolic rather than substantive in fulfilling an effective constitutional role in holding government and the executive to account. New audit approaches are being applied, including performance audits, within a ministry-wide audit program. However, the coverage of the audit work remains constrained by the limited number of qualified audit staff available. The technical work remains heavily dependent on external expertise as there is not yet a ‘critical mass’ of skilled resources to carry out adequate audit coverage. • The C&AG should gradually gain its full independent status. As a start, the institution should be allowed to decide its own staffing and budget, independent of the executive. • Audit reports will need to be made public and a public discussion organized around the issues raised. • Staff should be provided with enhanced training and incentives. More importantly, the issue of the separation of accounting and auditing cadres to avoid potential conflicts of interest should be at the top of the agenda. 59 The MBFs for the 14 ministries contained in the FY08 MTBF document totaled 440 pages. There is thus a risk of the MBFs becoming overly lengthy as the MTBF initiative is progressively extended to all 51 ministries/divisions. 103 CHAPTER SIX INCREASING DEMAND FOR BETTER BUDGET OUTCOMES THROUGH FISCAL DECENTRALIZATION Bangladesh’s administrative structure and expenditure management framework remain highly centralized for a country of its size and per capita income. There is a complex system of various types of subnational government entities with largely differing fiscal and organizational properties, but most of them play a very limited role in raising revenues and delivering public services. Intergovernmental transfers account for a small share of total government expenditures, and for the most part are allocated through ad hoc annual negotiations rather than being driven by formulas. With most transfers being earmarked and restricted to own- revenue sources, discretion at the subnational level to allocate resources to address local needs is limited, undermining responsiveness and operational efficiency that is typically associated with more decentralized public service delivery mechanisms. Regular local elections are held at the lowest levels of local governments, but constraints of politically elected officials to command fiscal resources and appoint and control their staff lessen their ability to affect local government operations. Despite a checkered history of attempted local government reforms, effective devolution of state power has been hampered, largely due to resistance of parts of the national leadership and centralized bureaucratic institutions opposing substantive jurisdictional and fiscal concessions to local governments. As elsewhere, the subnational reform agenda has gained renewed momentum among both policy makers and civil society. Sustaining these efforts in Bangladesh’s volatile political environment is critical to improve public services and enhance accountability to citizens. Key Policy Priorities • Develop a common vision among the various stakeholders on the role that different levels of government can play in service delivery. • Strengthen existing capacities for service delivery at the local level under existing arrangements. • Improve the intergovernmental fiscal framework over time through: (i) better definition of administrative functions and responsibilities; and (ii) better assignment of revenue and expenditure responsibilities. 6.1 Subnational Government Structures in Bangladesh 263. Bangladesh’s eventful history has shaped a complex subnational administration system with various types of overlapping deconcentrated and devolved local administrative entities (Figure 6-1). These entities largely differ in terms of their organizational setup, financing and functional responsibilities. Bangladesh’s constitution established a unitary form of government and despite various local government reforms, the tradition of a centralized state bureaucracy has prevailed until today. The central government is territorially divided into six administrative divisions, which are further subdivided into 64 Zila Parishads (districts). In rural areas, districts are further organized into 508 Upazila Parishads (Subdistrict Councils). 104 Figure 6-1: Subnational Government Structure Central Government (MLGRD&C) Divisions (6) Districts (64) Deconcentrated agencies Rural local govts. Urban local govts. Local Govt Engg Dept. Upazilas (481) Pourashavas (309) City Corporations (6) Dir. of Public Health Engg. WASAs Union Parishads (4498) 264. Divisions, districts and subdistricts do not constitute separate autonomous levels of government but rather are upwardly accountable administrative extensions of the central government at the regional level. Major line ministries operate deconcentrated offices at all three levels performing their service delivery functions, but these are not horizontally integrated into subnational administrations. There are no directly elected officials at these levels and civil servants in regional offices are appointed by, and report, to their respective central line ministries. A reform in 1991 established the so-called Upazila Development Coordination Committee (UDCC) at the subdistrict level, which comprises elected chairs of the lower-level union councils. This constitutes some indirect form of representation. However, the authority of these committees is limited in practice and subdistrict operations remain largely vertically driven by policies within the respective line ministries. 265. A more devolved structure with some rudimentary forms of elected representation exists at the lowest levels of government. In rural areas, there are about 4,498 Union Parishads (Union Councils), which are further subdivided into wards. Urban areas have two alternative structures: City Corporations in the six largest cities and Pourashavas (municipalities) in the rest of the country that are further subdivided into wards. The Union Parishad chairpersons and Pourashava and City Corporation mayors are directly elected by popular vote of the entire constituency, while the ward members/commissioners are elected by their respective constituencies. To ensure female representation, there are reserved seats for women. The term of local bodies is five years, but there are often delays in calling the local elections by the national government. Administratively, however, all employees and senior officials at the devolved government levels are appointed by the central government directly or its representatives at the district level. Employees of local bodies thus owe their primary allegiance to the center. This significantly impairs the operational independence and accountability of the elected local governments.60 266. The number of subnational government tiers (including the deconcentrated government units) is high by international standards. Administrative and operational costs are likely to rise with the number of layers of government which, together with a lack of clarity regarding functional assignments, undermines government accountability as constituents will find it difficult to determine which level is responsible for delivering specific services. Moreover, frequent and irresolute changes in the territorial 60 The Pourashava and Union Parishad governments hire a limited number of lower-level staff locally. 105 division and the structure of subnational administrations have generally undercut the development of strong and accountable local institutions. For example, the Upazilas were called Thanas for a period of time. Gram Sarkars or village governments were formed at the ward level in 2004 but these were very quickly eliminated following complaints that they were simply another opportunity for political patronage. 267. In addition, there are a set of central institutions under the overall umbrella of the LGD that exert significant influence on subnational administrations. The LGD sets basic rules for operation, determines the number of employees that local governments can have, sets the intergovernmental transfers, and structures other aspects of the overall system. The LGD has broad oversight over a range of areas including, local governments, drinking water, sanitation and sewerage, regional and local roads, village police and courts, local economic development as well as the development, operation, and maintenance of small-scale water plants. LGD also directs two national-level implementation agencies— the Local Government Engineering Department (LGED) and the Department of Public Health Engineering (DPHE)—as well as water supply and sewerage authorities (WASAs) in Dhaka and Chittagong, and a local government training academy—the National Institute for Local Government (NILG). 268. The LGED has evolved from a cell in the ministry in 1982 to become the largest and most influential line agency of the government, responsible for a portfolio of more than US$1.5 billion in 2006/07. It is the principal agency through which a large share of development assistance is channeled for a variety of programs and projects, ranging from infrastructure to social protection and capacity building. These works are undertaken by an extensive deconcentrated network of LGD offices at the Upazila and Zila levels. The LGED is also involved in the development of local level planning systems. It is responsible for the preparation of the Union Plan Book and the Upazila Plan Book procedures, and with its extensive resources and field presence, LGED has substantial influence on the planning, execution and monitoring of public expenditures at the local level. 269. At the headquarters, the Monitoring, Inspection and Evaluation (MIE) wing of LGD leads the supervision of local governments. It is charged with monitoring and evaluating the performance of UPs, Pourashavas and Zilas. It is not, however, empowered to supervise or monitor the LGED, the DPHE and the City Corporations, which immediately excludes the bulk of LGD’s budget from its direct scrutiny. The LGD is represented by a Deputy Director (DDLG) at the division level and by an Assistant Director (ADLG) at the district level. They supervise and support the functioning of local bodies at these levels and conduct inspections of local governments. Inspections are carried out together with the Upazila level officials, and the UP chairman and secretary. Inspections include an examination of the UP accounts (although this does not constitute a formal audit), registers, minutes of meetings, revenue performance, payment of local contributions to the salaries of the secretary and other UP staff, and a review of budget utilization and development activities. Inspection reports are prepared and shared with the MIE wing, the District Council (DC), the Upazila Nirbahi Officer (UNO) and the UP chairman. In addition, monitoring is also based on reports filed by local governments. The MIE wing has defined 14 indicators for monitoring, which include, among others: payment of salaries and allowances; tax collection and resource mobilization; budget performance; and progress with development projects. Based on these reports, the MIE wing produces an annual report on local government performance. These assessments form the basis for annual performance grants to local governments. 6.2 Subnational Government Finance 270. In terms of fiscal structure, Bangladesh is highly centralized (Figure 6-2). It lags behind other countries of its size and per capita income on both aggregate expenditure and revenue measures of 106 fiscal decentralization. Subnational expenditures as a percentage of total consolidated government expenditures are estimated to be in the range of 3-4 percent.61 The comparable figures for Indonesia or South Africa, two unitary countries that decentralized in the last 15 years or less, are 34 percent and 52 percent respectively. 271. On the revenue side, most major tax bases remain under the control of the central government. While own-source revenue potential varies vertically across the different levels as well as horizontally across different entities at each level, less than 2 percent of total government revenue is collected at subnational levels, again placing Bangladesh at the lowest end internationally. Local revenues are collected from property taxes, various user fees and lease income on community properties. However, tax rates are uniformly set at the central government level. Besides the restricted revenue base, revenue collection is undermined by weak municipal and local government tax administration. There are many reasons for this, including corruption and capacity constraints, in particular a poor assessment system for property taxes. Consequently, transfers from the central government are the main source of financing for many local governments, especially the UPs and B and C category Pourashavas. Figure 6-2: Bangladesh is One of the Most Fiscally Centralized Countries in the World 35.00 2 R = 0.6239 N = 33 (Developing and Transition Countries) nue 30.00 datedReve 25.00 otalConsoli 20.00 sShareofT 15.00 lRevenuesa 10.00 ub-Nationa 5.00 S Bangladesh 0.00 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 Sub-National Expenditure as Share of Total Consolidated Expenditures Source: IMF GFS 2001 (latest available year). Local Government Revenues 272. Revenue assignment for local governments is regulated by the Local Government Ordinance of 1976 and respective amendments. The available options and own-source revenues differ across the various kinds of local governments but each government, in principle, has a long list of potential revenue instruments comprising of various taxes, fees, and other income sources. However, in reality, many of these can generate little revenue and local governments rely heavily on transfers from higher levels of government. 273. UP revenues in general are very low and far from sufficient to meet minimum standards of service delivery. The Local Government Ordinance of 1976 includes 28 revenue sources that could be used by UPs, but in reality few of these sources are used (Figure 6-3). UPs can raise revenues from 61 This is based on randomized but nonrepresentative sample of 30 UPs and CCs. 107 holding tax, business and entertainment tax, fees on various licenses and services, lease of markets and ferry terminals and other structures, rentals of UP property, and income from other miscellaneous sources. Of these, the holding tax on the annual value of land is the most significant, followed by tax on land transfers and income from leasing of various UP properties. A model tax schedule exists and sets upper limits on the tax rates that UPs can assess, but they may have modest ability to control some rates. UPs have no collection or assessment staff; so their revenues are either collected by the chairman and Council or on a sharing basis. Figure 6-3: UP Revenue Sources 100% Other Grants, 2.5% Other Grants, 2.7% 90% Wage Grants, 15.5% Wage Grants, 17.3% 80% Block Grant, 5.9% Block Grant, 10.2% 70% Transfers 60% Sectoral Development Grants , 33.3% Sectoral Development Grants , 31.6% 50% 40% Program Grants , 4.9% Program Grants , 10.2% 30% Other Own Revenues, 10.0% Other Own Revenues, 8.1% Property Transfer Tax , 7.9% 20% Own Source Property Transfer Tax , 7.1% Lease Income, 4.4% Revenues Lease Income, 2.8% 10% Property Tax, 13.8% Property Tax, 11.8% 0% FY 06/07 FY 07/08 Source: Survey of 30 UPs. 274. UPs rely heavily on transfers for their financing with between 64 and 70 percent of total revenues coming from various kinds of transfers including salary subventions for UP staff. It is notable that, of these, direct block grants to UPs were less than 6 percent, while earmarked grants of various kinds accounted for the rest. Between 22 to 27 percent of the transfers are used to pay for staff salaries and honoraria for the chairman and the members of the Council. 275. Pourashavas are granted 26 revenue sources but many are believed to generate very little revenue. The main sources are property tax, land transfer tax, building permits, tax on professions and trades, advertisements tax, amusements tax, nonmotorized vehicles tax, and several fees. A model tax schedule has been created that establishes rates, although municipalities may be able to alter the effective tax rate by upholding appeals. Own-source revenues are less than one-third of total Pourashava revenues for a sample of 30 Pourashavas around the country (Figure 6-4), with the rest coming from transfers. Pourashavas depend heavily on project-specific funding, followed by general block grants. What is markedly different from UPs is the per capita level of revenues for Pourashavas. The total per capita revenue was nearly Tk. 1,220 in 2006/7, of which about Tk. 405 came from own source, rising to Tk. 1,930 and Tk. 580 respectively in 2007/8. Also, notable is the 58 percent jump in total Pourashava revenues over this period. 108 Figure 6-4: Pourashava Revenue Sources 100% Other Transfers, 0.9% Other Transfers, 0.9% 90% 80% Specific Projects, 44.2% Specific Projects, 53.3% 70% Tra nsfers 60% 50% General Government Grants, 21.7% 40% General Government Grants, 15.7% Other Own Revenues, 3.7% 30% Fees, 2.9% Other Own Revenues, 7.2% Rates, 12.0% Fees, 2.7% 20% Own Source Rates, 7.5% Revenues Leases Income, 5.2% 10% Leases Income, 4.7% Taxes, 9.4% Taxes, 8.0% 0% FY 06/07 FY 07/08 Source: Staff estimates based on a survey of Pourashavas conducted in 2008. 276. The main sources of revenue for City Corporations (CCs) are transfers and grants; taxes (property, entertainment, vehicle, professional); fees and fines; rents and leases; and loans (Figure 6- 5). Even though CCs have much higher local revenue potential than Pourashavas or UPs, just over one- third of their revenues came from own sources. The rest were various kinds of transfers, including general block grants and support for specific projects. The largest share came from financing for donor- supported projects, followed by financing for government-supported projects. Together, they constituted 50 to 60 percent of the revenues of CCs for 2006/7 and 2007/8. There has also been a small decline in total revenues from 2006/7 to 2007/8 (by a little over 1 percent in nominal terms), which is in complete contrast to Pourashavas. This is especially marked in the case of transfers like the general block grant and donor-supported projects. 277. There is a widespread feeling that the current levels of own-source revenue do not accurately reflect the potential of these cities. In particular, the large cities have not expanded the property tax rolls or undertaken timely assessments. Moreover, there are reports of large-scale evasion and corruption in local tax administration evidenced by collection rates that are often significantly below the estimates. Revenue composition of individual cities shows much variation (Table 6-1). Chittagong, notably, had a very high share of own-source revenues compared to the other five CCs surveyed, while Barisal and Sylhet were almost entirely dependent on various types of transfers. Donor funding for various projects also showed notable variations among cities. Whereas Khulna and Chittagong were least dependent on foreign aid projects, Barisal and Sylhet were heavily dependent. 109 Figure 6-5: City Corporation Revenues 100% Non-governmental , 5.2% 90% Specific Foreign 80% Specific Foreign Projects, 39.2% Projects, 31.1% 70% Transfers 60% Specific Projects Gov't, 21.2% Specific Projects Gov't, 18.3% 50% 40% General Government General Government Grants, 8.3% Grants, 6.6% 30% Other Own Revenues, 9.5% Other Own Revenues, 10.9% Fees, 5.1% 20% Fees, 5.6% Rates, 4.4% Own Source Rates, 4.5% Revenue 10% Taxes, 16.9% Taxes, 13.3% 0% FY 06/07 FY 07/08 Source: Survey of 6 CCs. Table 6-1: Revenue Composition of Individual City Corporations Dhaka (%) Khulna (%) Chittagong (%) Rajshahi (%) Barisal (%) Sylhet (%) Source 06-07 07-08 06-07 07-08 06-07 07-08 06-07 07-08 06-07 07-08 06-07 07-08 Share of Own-source 20.51 25.95 12.35 25.95 65.42 57.41 44.16 16.68 14.39 9.16 15.24 8.69 Share of Transfers 79.49 74.05 87.65 74.05 34.58 42.59 55.84 83.32 85.61 90.84 84.76 91.31 Foreign Aid Project 34.14 27.83 11.24 8.48 12.18 19.19 13.62 22.30 42.11 42.87 41.77 43.32 Source: Survey of 6 CCs. Note: All numbers are percentages of total revenues. 278. Transfers from higher levels of government are the most important source of financing for most local governments in Bangladesh–-accounting for 50-60 percent of local government income. There are some variations among the different types of local governments and among different entities of each type. The transfer system is fragmented, with a large number of different mechanisms through which resources are channeled to subnational levels. Broadly, there are four major types of grants: (i) earmarked sectoral project grants funded either by donors or by the central government; (ii) program grants, such as grants in grains for Test Relief, Food for Works, Vulnerable Group Development, Vulnerable Group Feeding, Rural Infrastructure Maintenance Program, as well as Age Pensions and allowances for widows and Muktijuddha (freedom fighters) in cash; (iii) block development grants; and (iv) recurrent expenditure grants, including grants towards salaries and allowances for the elected officials and staff. The lion’s share of transfers support vertically driven sector programs and are earmarked for specific development projects with local governments having very little discretion in the prioritization of these grants. 279. A smaller share is remitted as block transfers, which are allocated through the LGD ADP. Block transfers are determined according to a formula based on population, physical area and state of development of the area for which the grant is being made. The state of development is based on other criteria, such as poor communication levels (subjective judgment), literacy (using 1990 census statistics), 110 nutrition level (on the basis of expert opinion of the health officer) and rate of unemployment (1990 census statistics). Block grants in 2007 amounted to Tk. 17.9 billion, which is small–-both as a share of the national budget and as a revenue source of local governments. 6.3 Expenditure Assignment – Subnational Government Functions 280. The expenditure assignments for local governments are determined by the local government ordinance promulgated in 1976 and its subsequent amendments. However, there is no general framework for the assignment of functions and different forms of delegation seem to have evolved in different sectors, as well as among the three different types of local government. As a result, the delineation between national and local government roles is unclear, undermining incentives for local governments to perform their functions well. Moreover, the expenditure assignments that have been made to elected local governments are relatively narrow, at least by comparison with the responsibilities held by local governments in many parts of the world. 281. Bangladesh’s local governments have little responsibility for primary and secondary education, fire protection and primary health care. In principle, local governments have jurisdiction over public service providers in their area, including schools and hospitals. Despite their limited functions, local governments often fail to deliver many of the services for which they are responsible. Instead, they appear to deliver services to parts but not all of their constituencies. Very low resource levels and very limited institutional capacity for service delivery hamper local government service delivery. Local governments do not have the ability to hire and fire employees, which significantly inhibits their ability to set service levels and make decisions. 282. At the UP level, the focus of government responsibility appears to be on small infrastructure projects rather than on ongoing service delivery. This tends to result in disjointed and spread out infrastructure expansions with little emphasis on maintenance of assets (Figure 6-6). The average UP spends about Tk. 400,000 on recurrent expenditures, including salaries and allowances, which was 38 percent of total expenditures in 2006-07. The share devoted to salaries has been falling because of the rapid growth in development expenditures. About one-third of the salary budget is spent for honoraria of the chairman/members of the UP and the rest are spent for salary support for the UP staff. Some of the general establishment expenditure growth may be attributable to increases in the salaries and allowances of the UP staff due to the alignment of their compensation to the pay scale of the public employees. About 60 percent of UP expenditures are for development purposes, which have risen three times faster than recurrent expenditures. The major heads of development expenditures include agriculture, road construction and maintenance, and health and sanitation. These categories account for 35-40 percent of total expenditures of the UPs and constitute more than one-half of total development expenditures of the UP. 111 Figure 6-6: UP Expenditures 100% Other Development, 14.8% 90% Other Development, 18.9% Education, 1.6% 80% Education, 3.0% Road construction and maintenance, 14.8% Development Expenditures 70% Road construction and maintenance, 18.0% Health and sanitation, 5.0% 60% Agriculture, 14.2% Health and sanitation, 9.7% 50% Other Re-current, 5.8% Agriculture, 10.8% 40% Other Re-current, 6.3% 30% Recurrent Establishment Costs, 43.9% Expenditures 20% Establishment Costs, 32.8% 10% 0% FY 03/04 FY 06/07 Source: Survey of 30 UPs. 283. Legally, responsibilities of Pourashavas are divided into compulsory and optional functions. The compulsory functions include waste management, water supply and sanitation and drainage control. The optional services are provided as extended amenities to their inhabitants; some of these services include maintenance of educational institutions, control over private structures and establishment of charitable dispensaries. Current expenditures represented 30.5 percent of total expenditures in 2006/7, down dramatically from 2003/4 (Figure 6-7). The major factor has been the rapid increase in development expenditures, not slow growth in the operating budget. Municipal expenditures in Bangladesh are low relative to the assigned responsibilities due to the small resource base at their disposal from which to collect taxes and the limited absolute amounts of central government transfers. The functions of CCs and Pourashavas are similar, although CC functions are not categorized into compulsory and optional. 284. Pourashavas and city corporations account for nearly 70 percent of subnational government expenditures while representing only about 28 percent of the total population in FY06 (Figure 6-8). In contrast, Union Parishads accounted for about 20 percent of the total local government transfers. While this pattern partly reflects cost differentials of providing similar service levels in urban and rural areas, it is likely to be influenced by the weak equalization properties of the current transfer system that is rewarding own-source revenue collection, which is higher in larger cities. The distribution is also potentially vulnerable to political pressures that might work in favor of better organized and politically stronger urban populations. However, in line with the government’s renewed emphasis on rural development (analyzed in Chapter Two) some shifts away from the largest cities to municipal and rural areas can be observed between FY06 and FY07. 112 Figure 6-7: Pourashava Expenditures 100% 90% 80% Development, 54.73% 70% Development, 69.51% 60% 50% 40% Other Recurrent, 21.79% 30% Recurrent Expenditures Other Recurrent, 16.86% 20% Wages, 23.48% 10% Wages, 13.63% 0% FY 03/04 FY 06/07 Source: Survey of 30 Pourashavas. Figure 6-8: Expenditure by Type of Local Government 100% Wards, 7.64% Wards, 6.09% 90% Union Parishad, 20.59% 80% Union Parishad, 31.27% 70% Pourashava, 20.69% 60% 50% Pourashava, 25.11% 40% 30% City Corporations, 51.09% 20% City Corporations, 37.53% 10% 0% FY 06-07 FY 07-08 Source: Survey of 30 Pourashavas. 113 Deconcentrated Expenditures 285. In the centralized fiscal and administrative environment, service delivery responsibilities are primarily performed by the deconcentrated offices at the Zila and, in particular, the Upazila level.62 Upazilas manage service providers, including schools and hospitals and are used as conduits for several grant programs, including some grants provided by the LGD. Upazilas are also the main decision makers for transfers through the Food for Work and Volunteer Group Development Project, but line ministry staff, rather than local government officials, deliver these services. 286. Deconcentrated government spending across Bangladesh’s 64 districts is fairly equalized, but the allocation of development spending to poor districts is lower. Total expenditures in the district with the highest per capita spending, Rangamati, were more than seven times higher than in Kishoreganj, the district with the lowest per capita spending. However, variation is driven mostly by outliers, and spending in the average district is relatively close to the mean. Coefficients of variations are consequently below 1 for both development and recurrent expenditures, which is low in comparison to other developing countries with more devolved fiscal structures. Variation in recurrent expenditures is less pronounced than for development expenditure and driven mostly by higher spending in the six division capitals due to higher administrative costs related to division administration. While aggregate spending levels respond to a variety of spending needs and capacity that might not necessarily correspond with poverty levels, spending needs on pro-poor social infrastructure, such as health and education as well as spending under social welfare programs, are presumably higher in districts with a higher incidence of poverty. The largely discretionary ADP allocation system– at least at the aggregate level– does not seem to accommodate pro-poor spending considerations (Table 6-2 and Figure 6-9). Table 6-2: Fiscal Equality Measures for Deconcentrated Expenditures (2006/7) Level Recurrent Expenditures Development Expenditures Total Expenditures Minimum 1,422 660 2,471 Max 9,091 9,885 18,976 Max/Min Ratio 6.39 14.97 7.68 Coefficient of Variation 0.56 0.72 0.57 Source: MoF. World Bank staff estimates. 62 Lower level local governments often enhance service levels (such as with local education) where the responsibility lies with national government institutions. 114 Figure 6-9: Per capita Development Spending Varies Across Districts and Districts With Higher Poverty Rates Spend Less 10 4 y = -1.833x + 2.318 R² = 0.176 9 3.5 8 a k a T 3 7 d n Per Capita Spending in Thousand Taka a s u o 6 h T 2.5 n i g n 5 i d n e p 2 S 4 a t i p a C 3 r e 1.5 P 2 1 1 0 0.5 A NA I J AJ T E I R T I LJ J HJ R R R HR I TJ L A I J GR R I RRRL I JI I NI RR HARN NA R AR AA NNA L IR AAA UI NTILI RJNAM N N LERA AN A NN U A H A RALAR A UU A D A RTTE AAT HOGA DA NM UGI P NHGON A H OB U BG PRHH A O RA A UGP R AU RPG A R ARAGAU PUPJ UA U A P DP HT ONR SN HA I L E ANJPZH AKAU P OGA Z PPI L GNA HA HB KL NGA GI AB R I KGJ GGNI AOKH S AL GL T DA I DK A U R UF GO R HOB I I OY R M STKRP S S JRE RB A A GGIA A A I GE G K GOAB SDHE ZM N I G I HL A A R DA N U NA O A P J A N A E RH R O TA A U N A A NR K M M I B R AHP I N AR AINNA A R I A R N GNB K ATH N AI I J R VA H ' B B M A S HGC S OR SHC A S A G B UA H I A E Y S O O I A M D A A E U A AA Y TPL R E A D 0 PR S H W T J A NKUHC OA H DHF T B B X S A NR J H A N AK GH UL A HI MS I AJ O M L NS N AME U A NGSHC J A M AI RH U M OK A N UG N AR 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 C N Y K N A M H N S P AC M A C L M A B T M L A P N H District Poverty Headcount R K B Source: FY 2006/07. MoF. World Bank staff estimates. Note: Scatter Graph excludes districts with Chittagong Hill Tribes: Rangamati, Bandarban, Khagrachhari, which represent outliers in terms of development spending. 6.4 Planning and Budgeting at Local Level 287. Expenditure planning in Bangladesh is undertaken at different levels by different agencies. Much of the planning is done vertically and the fund flows reflect this vertical orientation. At the national level, the PC has responsibility for macro and strategic planning. At the sectoral level, various ministries and line departments, through their planning cells, undertake sector-specific planning. Sectoral planning under vertically integrated administrative structures is highly centralized and compartmentalized. Officers at the deconcentrated level provide inputs to sectoral agencies on development priorities and identify programs and projects, often without any local inputs. The practice is to rely on the “local knowledge” of the field level officer. While there are forums for consultation and coordination at the Zila (DDCC) and Upazila (UDCC) levels, in practice neither of these forums facilitates inputs from local bodies or citizens. 288. Currently, there are three defined procedures for local level area-based planning. • UPs are required to produce annual and five-year development plans based on guidelines prepared by LGD. However, UP planning is limited by strict guidelines that set out allocations across sectors and limit local discretion. Local planning becomes an ineffectual exercise in this scenario. • UPs are also required to produce and update a Union Plan Book with the help of LGED, identifying local resources and priorities. In practice, only a small number of UPs have prepared Plan Books, and where they have been prepared, they look more like wish-lists of infrastructure projects that may not reflect public priorities and have no bearing on the actual resources available. • Finally, at the Upazila level, there is provision for a Plan Book. However, in the absence of a planning body at the Upazila level, the planning process, as noted earlier, is a sectoral and bureaucratic exercise led by the UNO and the Upazila Engineer. 115 289. Since there is no coherent area-based and integrated local-level planning and expenditure prioritization process at the present time, local budgeting becomes an ineffectual paper exercise. Local bodies are required to balance budgets, and therefore, in that sense there is a hard budget constraint. Current revenue patterns are used to forecast future revenues. These forecasts show “taxes in arrears,” which leads to a major case of revenue overestimation. In the case of UPs and Pourashavas, the budget is prepared by the chairman and the secretary. The previous year’s budget becomes the starting point, and then some increments are added where necessary. 290. Budgets are required to be prepared and submitted for approval to higher authorities by the end of June, however, in reality, this often extends to July or later. UPs are required to hold an open budget meeting in order to take the community into confidence. In most cases, it appears that the budget is finalized with some inputs from members, but without active input from the community which, in most cases, is unaware of the budget. The situation is not very different in Pourashavas. Without any transparency, it is not unusual for budgeting to become an exercise in dispensing patronage and rents. In the case of UPs, the final budget is then sent to the DC’s office for approval, while in the case of Pourashavas, the approving authority is the LGD, which must then approve or amend the budget within 30 days. The budget is automatically approved if LGD does not respond in the 30-day period, and the perception exists that the budget is often accepted without LGD review or comment. 6.5 Key Policy Directions 291. Reforms of local government institutions and intergovernmental fiscal relations can be a critical element supporting more effective and efficient public spending in Bangladesh. The government has successfully maintained fiscal prudence and expanded service access in a number of sectors, including education and health and, more recently, rural roads. Now the challenge is to improve quality and efficiency in service provision, which is more difficult to achieve in the overly centralized framework. Greater devolution of government would offer significant advantages to Bangladesh by enhancing the quality of local public service delivery, encouraging local resource mobilization, allowing for greater local voice in public service decisions, and generally increasing satisfaction with government. 292. Significant re-organization of the government structure would be necessary to create a more devolved system that operates with incentives for efficient behavior at all levels of government. Greater local control would be necessary on expenditure choices and staffing in conjunction with greater ability to raise own-source revenues. The intergovernmental revenue structure would need to be re-examined and restructured in light of different expectations from local government, with the likelihood that more transfers are necessary. Furthermore, major reforms would be necessary in a number of national government institutions, such as the DPHE and the LGED. These organizations would need to serve as facilitators and monitors rather than being highly integrated into local government functions, as they are today. While this reform agenda has gained some momentum, actual change has lagged behind. 293. The prospect of successful implementation of local government reforms would increase with a common vision to provide direction to, and avoid confusion among, the multiple stakeholders involved. In addition, there is a need to plan and systematically implement capacity and skills development at all organizational levels to ensure the reforms can be properly implemented. In addition, a high level oversight and steering committee for budget reforms could be envisaged to ensure consistent high level attention to, and coordination of, reform implementation and timely resolution of 116 issues that might arise. The commission would oversee and guide the required legal, institutional and fiscal changes required to achieve an adequate framework for meaningful local governance. 294. Building accountable local governments that can deliver public goods and services to local communities more efficiently warrants reforms in three critical areas of intergovernmental systems. These include: (i) reforming intergovernmental fiscal frameworks to clearly define responsibilities of subnational governments and sectoral functions; (ii) matching revenue assignments to service delivery responsibilities and rationalize the system of intergovernmental fiscal transfers; and (iii) strengthening accountability mechanisms between policy makers and service providers. 295. Firstly, the current framework of revenue and expenditure assignments–the cornerstone of any intergovernmental fiscal system–remains overly centralized and lacks clarity, resulting in diffused accountability, poor efficiency, and suboptimal outcomes in service delivery. Although there is some broad demarcation of assignments and responsibilities, the current regulatory framework and practice has resulted in overlapping functions and responsibilities. This diffuses accountability to the public and weakens real autonomy and authority of local governments. Thus, there is a need to clearly define the expenditure responsibilities of different tiers and structures of local governments’ service delivery sectors and specific functions within these sectors. This also necessitates redefining the role of various deconcentrated service delivery entities, such as the LGED, the DPHE and the WASAs that operate in urban and rural areas. 296. Secondly, the revenue assignments of the local bodies should match the service delivery responsibilities of the local governments. Local bodies currently have limited revenue autonomy in base and rate setting as well as authority in administration. The current mix of revenues does not promote accountability and efficiency because there is no matching of preferences in service delivery with a willingness to pay, especially in urban areas. Revenue assignments should support expenditure responsibilities. The revenue regime should promote accountability and transparency by identifying sources of financing best suited for each type of expenditure responsibility. To the extent possible, public services with a local distribution of costs and benefits should be funded through local tax payments to ensure efficient and accountable use of public resources. If local tax rates are flexible, they can signal the costs of local services at least at the margin, and local residents can choose the level of services they desire. At the same time, local residents have a greater incentive to monitor performance if services are directly funded from their tax payments. By the same token, excessive use of transfers might undermine efficiency, since local governments can shift part of their expenditures onto higher levels of government. 297. In addition, there is a need to move towards a more rational, transparent and predictable system of intergovernmental fiscal transfers. This would build on the experience of the World Bank- supported Local Governance Support Project and other similar initiatives, as well as learning from other comparable international examples. Fundamental to this is the need to enhance direct fiscal transfers to local bodies in line with their assigned functions. This would call for a comprehensive review of the current intergovernmental fiscal architecture, especially the vertical and horizontal revenue sharing mechanisms, as well as the institutional arrangements for intergovernmental fiscal relations. In particular, the fiscal transfers through ADP need to be more transparent and timely, and based on rational criteria. Further, there is also a need to determine the incentives that should be built into the system to enhance the own-source revenue performance of the local bodies. 298. Thirdly, the impact of decentralization on local service delivery is largely contingent upon the responsiveness of local governments to their constituencies and citizens. This, in turn, is largely predicated on the extent to which they are held accountable. Strengthening accountability mechanisms between service providers and policymakers will require ex-ante controls and ex-post audits. The 117 foundations of such an accountability framework should lie in: (i) performance rather than mere compliance; (ii) outcomes and outputs rather than only inputs; and (iii) strengthening client focus within the government and service providers rather than adding additional layers of bureaucratic controls. Thus, greater fiscal resources and enhanced discretion to local bodies need to be supported with more robust systems of fiduciary controls, and improved local capacities in planning, budgeting, procurement and financial management. Experience has shown that a range of public sector and social accountability mechanisms, which go beyond elections, can play a vital role in enhancing the accountability of local governments towards their constituents–for example by financial reporting and audits, citizen report cards and town hall meetings. 299. In principle, these reforms could extend to different elected tiers of local bodies, such as UPs, municipalities, the City Corporations and the Upazilas. Supporting and strengthening the newly elected Upazilas would become critical if they are to play a vital role in service delivery, especially for those services that require a certain degree of scale economies. Strengthening Upazilas is also important for UP level reforms to be sustainable in the longer term. This will call for long-term commitment and greater resources from the government. With regard to Pourashavas and City Corporations, extending the principles of enhanced fiscal resources in the context of a predictable and transparent intergovernmental framework, along with strengthened systems of local accountability and capacities, will be vital to improving their accountability and efficiency in service delivery. Concurrently, greater emphasis should also be given to enhancing own-source revenues, in particular for urban local bodies that have strong revenue potential, and strengthening sustainable institutional mechanisms for capital financing of urban infrastructure. 118 APPENDIX ONE THE STRUCTURE OF THE PUBLIC SECTOR IN BANGLADESH Nonfinancial Public Enterprises Overall net losses of SoEs increased more than three times, from around Tk. 7 billion in FY04 to Tk. 26 billion (0.6 percent of GDP) in FY06. SoE value added declined by around 78 percent, from 2.3 percent of GDP in FY95 to 0.5 percent of GDP in FY06, while their gross saving reverted to a dismal -0.3 percent of GDP in FY06 after reaching a historic high of 0.5 percent of GDP in FY03. A decline in their capital expenditure from 1.9 percent of GDP in FY01 to 1.2 percent in FY06 contributed to an improvement in the SoE fiscal position as their saving-investment gap decreased from 2.1 percent of GDP in FY01 to 1.1 percent of GDP in FY04. This, however, was not sustained as the deficit rose to 1.5 percent in FY06, an amount equal to the central government’s saving-investment gap (see Box). Public Financial Institutions Public financial institutions consist of SCBs and development financial institutions (DFIs) which constitute a large part of Bangladesh’s banking system. Although net profits of SCBs have been close to zero throughout the last decade and a half, reported profits continue to be seriously overstated because of significant shortfalls in loan loss provisioning. Despite the over-reporting, the total SCB and DFI dissaving in FY05 was 0.4 percent of GDP. The cumulative provisioning shortfall for SCBs as of the end of June 2005 was estimated at Tk. 45.7 billion or 1.2 percent of GDP.63 The public banking sector has thus actually lost substantially throughout the last decade. Note that these are partial estimates of loss, based on shortfalls in provisioning for loss on only classified loans. If shortfalls for provisioning for loss on other investments are taken into account, the financial losses of PFIs would be even higher. The SCBs and DFIs continue to have very high numbers of nonperforming loans mainly due to provision of loans by them on considerations other than commercial and under directed credit programs. However, their NPL ratios have declined somewhat from 29 percent in FY03 to 26 percent in FY07 due to some progress in recovery of long outstanding loans and write-offs of loans classified as “bad” or “loss” for over five years. The dominant market share of the PFIs has declined since the early 1990s, but the public banks still account for a significant 45 percent share of the market.64 On the asset side, the market share of SCBs and DFIs declined respectively from 57 percent FY01 to 47 percent in FY05. The largest share of PFI loans goes to the private sector. However, because of large shares of nonperforming loans, total assets (and capital) are inflated by under provisioned bad loans. On the liability side, PFI dominance declined from 56.1 percent in FY02 to 48 percent in FY05.65 Like assets, most PFI deposits come from the private sector. Due to dismal financial performance and notwithstanding large government backing, SCBs and DFIs are undercapitalized. The capital adequacy ratios (CAR) of SCBs and DFIs continue to remain below the minimum regulatory CAR of 9 percent of their risk weighted assets. Bangladesh Bank 63 However, efforts have been directed at raising bad loan provisioning practice in Bangladesh to international standards. 64 Bangladesh Bank Annual Report, 2006-2007, p. 28 and World Bank, Bangladesh: Increasing Access to Rural Finance in Bangladesh: The Forgotten Missing Middle, June 2007. p. 2 65 Bangladesh Bank, Annual Report 2003-2004, Table 5.1, p. 27. 119 The Bangladesh Bank’s (BB) total income averaged a steady 0.5 percent of GDP during the last ten years. The central bank’s gross income is largely determined by its seigniorage revenue collected from issuing money, amounting on average to 0.5 percent of GDP. At total expenditure of 0.2 percent of GDP, BB generated systematic net profit close to 0.35 percent of GDP per annum on average during FY95-05. Of this, 80 percent is transferred to the central government as dividend payments. The negligible remainder is BB’s gross savings. Total BB assets grew proportionately less than GDP in recent years, declining from 12 percent of GDP in FY00 to 9.5 percent in FY07. The central government is the largest debtor to BB, whose outstanding loans constituted 5.2 percent of GDP in FY07, compared with 3.4 percent in FY00. Gross foreign assets of BB increased from 3.4 percent of GDP in FY00 to 6.3 percent in FY04. On the liability side, currency and bank reserves at BB have grown consistently with the rise in demand for money in response to increasing GDP under conditions of modest inflation. Reserve money–the sum of currency and bank reserves–as a ratio to GDP has fluctuated around 7 percent in recent years. BB’s foreign debt, after declining to 0.9 percent of GDP in FY00, has remained stable at around 3 percent of GDP in recent years. Other There are large hidden subsidies and contingent liabilities which are not included in the budget. Direct budgetary subsidies currently amount to less than 0.5 percent of GDP and are provided for only a few items, including school textbooks, fertilizer distribution (with budgetary allocations only for imported urea), and several nontraditional exports (such as knitwear, leather, jute goods, flowers, frozen food, and some agro-processed products). Indirect subsidies, however, are estimated at 2.6 percent of GDP. These are provided through distorted gas and electricity prices as well as preferential interest rates on loans to SoEs. Similarly, large contingent liabilities have been accumulated on account of SoEs and the banking sector. Part of the increase comes from government-guaranteed nonconcessional foreign suppliers’ credit. The government started disclosing the explicit guarantees and counter-guarantees with the FY07 budget. The outstanding amount of explicit guarantees and counter-guarantees provided by the central government against loans negotiated by various state-owned financial and nonfinancial enterprises has declined to US$1.9 billion (2.4 percent of GDP) compared with US$2.5 billion (4.1 percent of GDP) in FY07. The government has also provided Tk. 15 billion in the FY07 revised budget and Tk. 78.23 billion in the FY08 budget to internalize quasi-fiscal costs arising from BPC debt and Bangladesh Power Development Board BPDB deficits. The continued misalignment of administered prices of fertilizer, diesel, power and natural gas with their cost of production and/or import cost could potentially exert considerable pressure on budgetary resources in the medium term. This would make it difficult to meet the potential costs related to the structural reforms in the banking sector and SoEs, and result in increased budgetary outlays on the infrastructure and social expenditures needed to sustain growth and poverty reduction goals presented in the PRSP. 120 APPENDIX TWO: CONSOLIDATED PUBLIC SECTOR DEFICIT – CONCEPTS AND DEFINITIONS66 How the fiscal deficit is measured has an important bearing on an analysis of the sustainability and macroeconomic implications of deficits. The composition of the public sector and the economic relevance (or quantifiability) is key dimensions of various types of deficit measures. The composition of the public sector can be defined in three different ways: (i) central government only; (ii) consolidated nonfinancial public sector, which adds local government, social security, and nonfinancial public enterprises; and (iii) consolidated total public sector, which adds the central bank and public financial institutions. Deficits based on the most inclusive definition of public sector are more informative of a country’s fiscal position and public sector resource transfers if they are not subject to arbitrary accounting conventions. Practitioners often limit the deficit measure to nonfinancial public sector deficit. This leaves out an important fiscal element, the profit/losses of the central bank and other public financial institutions from quasi-fiscal operations that subsidize activities in private and state enterprise sectors. A comparison of quasi-fiscal deficits and conventional nonfinancial public sector deficits illustrates how misleading nonfinancial public sector deficits are as indicators of overall fiscal policy when quasi-fiscal operations are large. In FY99, for instance, the public financial sector deficit was larger than the nonfinancial sector deficit. There are also several ways of measuring the deficit that are more or less economically relevant. The nominal cash approach permits broad comparability of deficits across countries while a variant, the operational deficit, deducts the inflationary component from nominal interest payments on public debt. The deduction, reflecting the compensation of debt holders for erosion of the real value of public debt caused by inflation, is an important correction for high inflation, high domestic debt economies. Despite recent increases, Bangladesh is still a low domestic debt economy and the inflation rate has been in the low single digits. Consequently, the inflation correction is not critical for Bangladesh. An accrual or, payments order, approach measures income and spending actions when they occur, even if they do not immediately involve cash flows. Deficits measured on an accrual basis would be larger than those measured on a cash basis when arrears have been allowed to accumulate on government payments of interest, wages, or purchases of goods. Accrual-based deficits open the door to a whole set of unconventional deficit measures based on considerations of public net worth or inter-temporal budget constraints. Such measures constitute the most meaningful gauge of a government’s fiscal position, but they are not observable. There are other economically meaningful measures. Despite their usefulness for assessing overall fiscal stance and issues of sustainability and solvency, the questions addressed in the analysis of Bangladesh’s fiscal deficit require the use of cash-based nominal deficit measures with the widest available coverage of the public sector. The Consolidated Public Sector (CPuS) comprises five entities: (i) the central government (GOV); (ii) state-owned enterprises (SoEs); (iii) State Commercial Banks (SCBs); (iv) Development Financial Institutions (DFIs); and (v) the Bangladesh Bank (BB).  Central Government: The overall deficit is measured by the central government’s saving- investment gap. Central government savings are calculated by subtracting the current expenditure and food account deficit from the sum of revenues and foreign grants. Central 66 This appendix draws heavily from Easterly and Schmidt-Hebbel (1993). 121 government investment is the part of ADP which the central government spends directly and not through SoEs.  State-owned Enterprises: Gross saving is calculated by adding depreciation to retained income. The latter is the part of net profit retained after subtracting tax and dividend payments. SoE investment is their capital expenditures.  State Commercial Banks: Saving of the SCBs is their net profit where net profit is the difference between total income and total expenditure.  Development Financial Institutions: DFI saving is defined in the same manner as that of SCBs.  Bangladesh Bank: The central bank pays a significant portion of its profit to the government as a dividend. Thus, the saving of Bangladesh Bank is calculated by subtracting the dividend payment from net profit. Consolidated Public Sector (CPuS) Deficit is defined as the saving-investment gap where the CPuS saving comprises the saving of all five entities of the public sector and CPuS Investment is the sum of the investment of the central government and SoEs. This treatment effectively defines public sector deficit as the difference between aggregate public sector saving and aggregate public sector investment which is financed by surpluses from other sectors. The CPuS can, therefore, be rewritten in terms of the economy’s aggregate resource or saving-investment constraint: CPuS Deficit = Public Investment - Public Saving = (Private Saving - Private Investment) + Foreign Savings A larger CPuS deficit must lead to some combination of lower private consumption (at a given level of income), lower private investment, and higher foreign saving (current account deficit). Which of the three components bears the burden of higher CPuS depends broadly on factors that influence the private domestic and foreign response to public deficits: the flexibility and sophistication of domestic financial markets, access to external financing, the source of domestic financing (bank and nonbank), planning horizons of consumers and investors, and the composition of the deficit. 122 APPENDIX THREE FRAMEWORK FOR LONG-TERM SUSTAINABILITY ASSESSMENT The government’s flow budget constraint can be written as: bt = (1 + r )bt −1 − xt − σ t (1) Where bt is the end-of-period stock of real debt, xt is the real primary balance and σ t is the real value of seigniorage revenue. Forward iteration on this equation combined with the condition lim(1 + r ) − ( j +1) bt + j = 0 (2) j→∞ implies ∞ bt −1 =  (1 + r ) −(i +1) ( xt +i + σ t +i ) (3) i =0 This is the government’s lifetime budget constraint. It states that the government finances its debt at the end of period t-1 by, from date t forward, raising seigniorage revenue and running primary surpluses with an equal present value. The most basic tool for fiscal sustainability analysis uses a steady state version of this lifetime budget constraint. To begin, consistent with the presentation in most World Bank and IMF documents, it is useful to rewrite (3) in terms of stocks and flows expressed as fractions of GDP. Let y t represent real GDP, define b t = _ bt _ ,xt = xt _ and σ t = σt . Given this notation, equation 3 can be re-written as yt yt yt − ∞ − − bt −1 y t −1 =  (1 + r ) −( i +1) ( x t +i + σ t +i ) y t +i or i =0 _ ∞ _ _ y t +i b t −1 =  (1 + r ) −(i +1) ( x t +i + σ t +i ) (4) i =0 y t −1 yt Imagine a steady state in which (i) real GDP grows at a constant rate g, so that = 1 + g , (ii) the y t −1 _ primary surplus as a fraction of GDP is a constant x , and (iii) seigniorage as a fraction of GDP is a _ constant σ . In this case, equation (4) reduces to 123 _ ∞ 1 + g i +1 _ _ b t −1 =  ( ) ( x+ σ ) (5) t =0 1+ r Assuming r>g, (5) reduces to _ _ _ _ _ b t −1 = b ≡ ( x + σ ) / r (6) _ Where r = (r − g ) . (1 + g ) Equation (6) can be used in two ways. First, one could make reasonable assumptions about the values of _ _ x, σ , r and g based on historical trends in the country’s fiscal accounts as well as typical historical values of seigniorage revenue, the real interest rate and the real growth rate. Together these assumptions can _ be mapped into an estimate of b using equation (6). If the government’s actual stock of debt exceeded this estimate, then the government’s finances could be argued to be unsustainable. Alternatively, equation (6) can be re-written as _ _ _ x = r b t −1 − σ (7) _ _ Given estimates of r and σ and data on the size of the government’s actual debt stock, bt −1 , equation (7) can be used to determine the necessary size of the primary balance to ensure fiscal sustainability. _ _ That is, rather than setting x equal to some historical average, one can determine the value that x would need to take in the future to maintain sustainable finances. _ A final interpretation of equation (7) is that if x were set consistent with it, then the debt-GDP ratio _ _ _ would remain constant in the steady state. In other words, when x , σ , r and g are constant and x is _ given by (7), not only will the government’s finances be sustainable, but it will also be true that b t will _ be constant and equal to b . _ _ Mt How to calculate σ ? Assume base money is a constant fraction of GDP: yt = m ; inflation is Pt constant at some rate π , and real growth is constant at the rate g, so that _ M t − M t −1 Mt P y M t −1 σ= = − t −1 t −1 Pt y t Pt y t Pt y t Pt −1 y t −1 _ 1 _ π + g + πg _ _ = m− m= m ≡σ (8) (1 + π )(1 + g ) (1 + π )(1 + g ) 124 With assumptions and projections regarding inflation, growth and the size of base money, one gets a _ benchmark value for σ . Calculations of sustainable primary balances for a range of real GDP growth and interest rates are presented in the table below: Sustainable Primary Balances (% of GDP) Nominal Interest GDP Growth Rates (%) Rates (%) 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 6 -1.1 -1.4 -1.6 -1.9 -2.1 -2.4 -2.6 -2.9 -3.1 7 -0.7 -0.9 -1.2 -1.5 -1.7 -2.0 -2.2 -2.5 -2.7 8 -0.3 -0.5 -0.8 -1.0 -1.3 -1.5 -1.8 -2.0 -2.3 9 0.2 -0.1 -0.4 -0.6 -0.9 -1.1 -1.4 -1.6 -1.9 10 0.6 0.3 0.0 -0.2 -0.5 -0.7 -1.0 -1.2 -1.5 11 1.0 0.7 0.5 0.2 -0.1 -0.3 -0.6 -0.8 -1.1 12 1.4 1.1 0.9 0.6 0.4 0.1 -0.2 -0.4 -0.7 13 1.8 1.6 1.3 1.0 0.8 0.5 0.2 0.0 -0.3 14 2.3 2.0 1.7 1.4 1.2 0.9 0.7 0.4 0.1 15 2.7 2.4 2.1 1.9 1.6 1.3 1.1 0.8 0.5 16 3.1 2.8 2.5 2.3 2.0 1.7 1.5 1.2 0.9 Note: Benchmark values for other parameters: Total debt-GDP ratio = 48.3%; Inflation rate (GDP Deflator) 4%; Base money-GDP = 7.1%. Fiscal Sustainability: An Extended Accounting Framework We start with the central government (GOV). Equation (1) below describes the current-price budgetary identity for the GOV deficit concept consistent with the conventional Overall Budget Balance, expressed as a ratio to current-price GDP. The left-hand side defines the total budget deficit ratio to GDP as the sum of the primary deficit, domestic interest payments, and foreign interest payments. Interest payments are separated into shadow interest payments (the amount of interest that the government would have to pay at a market interest rate) and the implicit interest rate subsidy (the difference between actual and market interest payments). The right-hand side of the identity defines the three sources of funds that are available to GOV: domestic debt financing, foreign debt financing, and foreign grants. Therefore: 125 where D is primary deficit, P is the GDP deflator, y is real GDP, si (si*) is the domestic (foreign) nominal shadow interest rate, i (i*) is the domestic (foreign) nominal interest rate, and B (B*) is domestic (foreign) debt. E is the nominal exchange rate expressed as domestic currency units per unit of foreign currency and FG is foreign grants. All capital letters denote current-price variables. Foreign variables (si*, i*, and B*) are denoted by asterisks and expressed in foreign-currency units. Stock variables are dated at the end of the corresponding period denoted by subindexes t and t-1. Simple manipulation of equation (1) yields the following identity for the nominal primary deficit ratio to nominal GDP. Note that the primary deficit ratio can be financed by either of the following sources: increasing the ratios to GDP of the outstanding stocks of domestic debt or foreign debt; the net growth effect (from the difference between real GDP growth and the real domestic interest rate) on domestic debt holdings; the net growth effect (from the difference between real GDP growth and the real foreign interest rate adjusted for real exchange rate depreciation) on foreign debt holdings; the domestic and foreign real interest rate subsidies arising when actual rates are below their market values; and foreign grants. Hence: where all lower-case variables (with the exception of all rates) denote the ratio of the corresponding nominal variable to nominal GDP. The following additional notation is used for variables in rates: n is real GDP growth, r (r*) is the domestic (foreign) real interest rate, and sr (sr*) is the domestic (foreign) real shadow interest rate, ε is the rate of real exchange rate depreciation, defined as the nominal exchange rate depreciation adjusted for foreign and domestic inflation. All t-period rates of change refer to changes in variables between periods t-1 and t. Now consider the corresponding equations for the Consolidated Public Sector. The following equation describes the current-price budgetary identity for the CPuS deficit (as a ratio to current-price GDP). This deficit measure is consistent with the saving-investment difference. It is not consistent with the Overall Budget Balance measure applied above to GOV. The left-hand side of equation (3) is equivalent to equation (1): the total deficit ratio to GDP (now for the saving-investment difference of the CPuS) is the sum of the primary deficit, domestic interest payments, and foreign interest payments, where the 126 interest payments are separated again into shadow interest payments and the implicit interest rate subsidies. The right-hand side of the identity defines three sources of funds that are available to CPuS: monetary financing, domestic debt financing, and foreign debt financing.67 where M is reserve money outstanding with the private sector. Similar to equation (1), equation (3) can be easily manipulated to yield the following identity for the nominal primary CPuS deficit ratio to nominal GDP. On the right-hand-side of the following equation we find six of the seven primary deficit financing sources that were included in equation (2). The difference with equation (2) is the absence of foreign grants from equation (4) and the presence of three new monetary financing sources: the increase in the money stock ratio, the growth effect on the money stock, and the inflation tax on money–the sum of the latter two sources comprise seigniorage revenue that accrues to the CPuS and is collected by the central bank. where ε is domestic inflation. 67 Monetary financing is available to the CPuS (but not to GOV) because of its inclusion of the central bank (BB) that issues reserve money held by the private sector. Foreign grants are not included as a source of financing, because the deficit concept applied to the CPuS–consistent with the saving-investment definition–includes foreign grants as a source of government income, and are therefore, reflected in CPuS saving. 127 Now consider two alternative concepts of fiscal sustainability that will be applied to both deficit measures–for the GOV and the CPuS. The narrow version of a sustainable primary deficit ratio is the level that is consistent with a constant debt to output ratio. According to equations (2) or (4), this is the ratio obtained when the first right-hand side terms are set to zero, that is when the nonmonetary domestic and foreign debt ratios (in equation 2) and the monetary debt ratio (in equation 4) are constant. However, current fiscal policy–as reflected by the current levels of all other financing sources– remains unchanged. Alternatively, a wider version of sustainability derives a primary deficit to output ratio consistent with permanent or longer-term values of the right-hand side financing sources. This implies adding to the constant debt ratios the substitution of longer-term values for actual levels of all relevant right-hand side variables in equations (2) and (4). 128 APPENDIX FOUR BUDGETARY PROCEDURES AND INSTITUTIONS IN BANGLADESH Methodological Framework for Comparative Analysis68 This appendix presents the analytical framework and the methodology used for assessing the effectiveness of budget procedures in ensuring efficient allocation of fiscal resources in Bangladesh. The methodology is designed to generate measures that are comparable over time and across countries and, as a result, make it possible to identify key weaknesses and strengths in a country's fiscal institutions and procedures. Section 1 presents the analytical framework and Section 2 describes the methodology. 1.1 Analytical Framework The goal of budgeting is to allocate public resources among alternative activities that the government undertakes. An efficient budget equalizes the marginal returns on all government activities with the marginal cost of public funds. However, as the literature on budget procedures has shown, solving this problem in practice faces two basic difficulties. First, the costs and returns of various activities are not commonly known. Second, strategic behavior of diverse beneficiaries of government spending in their attempts to increase their share of the resources can give rise to a great deal of waste and misallocation (known as the "tragedy of the commons"). This difficulty is, of course, exacerbated by the first one, as beneficiary groups tend to use their private information to their advantage. To achieve efficiency, the government needs appropriate coordination and conflict resolution mechanisms, which must be embedded in budgetary procedures and broader political and economic institutions. Budgetary procedures are all the rules and regulations that constrain the ways in which budgets are formulated, approved, and implemented. These rules and regulations may have formal sources (for example constitution, common laws, and bureaucratic or legislative by-laws) or follow from informal social norms, understandings, and traditions. To act as constraints on the process of budget making, budgetary procedures must be costly to change so that they have a longer life than the budget itself. Therefore, rules that are made along with the budget cannot be part of budget procedures because they cannot act as constraints on budget making. The aspects of budget procedures that are relevant to fiscal outcomes are those that structure information flows, create incentives, shape coordination, or restrict collective decision-making. For intertemporal efficiency and macroeconomic stability (which are jointly referred to as the "aggregate fiscal discipline") we follow the literature by considering the main institutional aspects under two topics: ex-ante constraints and coordination mechanisms for restricting the overexpenditure of the interest groups involved in budget making at the expense of their collective interests (in the form of inflationary financing) or at the expense of future generations (in the form of public debt). We then discuss the institutional features that are important for the efficiency of cross-sectional allocation and for the operational efficiency of activities undertaken by government agencies. Finally, we examine the interaction of budget procedures with a number of country characteristics. 68 This appendix draws heavily on Hadi Salehi Esfahani (1999a). 129 A. Ex-Ante Constraints Ex-ante constraints are laws, social norms, constitutional clauses, and the like that restrict the decision powers of policy-makers and predetermine certain budgetary outcomes by placing bounds on spending or deficit. For example, a constitutional clause or a law that limits budget deficits or public debt may act as a mechanism to force decision-makers to internalize the costs of their spending programs. Such rules are common among subnational governments in federations (for example, many states of the United States have adopted a balanced budget rule or a "golden" rule, which limits borrowing to the amount of investment). The debt and deficit limits imposed by the Maastricht Treaty on the countries of the European Union can also be viewed as a form of ex-ante constraint. Most countries have constitutional requirements that government spending must be properly financed. But, this is generally interpreted as indicating that the government should not simply cover its deficit by printing money, although even that restriction is not always observed.69 Another important constraint mechanism are the laws that require spending or deficits to be constrained by projections based on a multi-annual macroeconomic program. The use of macro models is, of course, rather widespread for forecasting purposes, however, in some countries the exercise is more detailed and the forecasts are used as constraints on future budgets. When ex-ante constraints are effective, they can help maintain aggregate budget discipline. In particular, they may prevent dominant executives or legislative majorities that face turnover from overspending and from leaving large debts for future generations. But, the constraints can also impose nontrivial costs on the economy because they reduce policy flexibility and may even entail adverse effects. Rigid restrictions on budget deficits, in particular, force the government to cut expenditure when negative economic shocks lower income and tax revenues, thus exacerbating economic fluctuations and causing intertemporal misallocation (Barro 1979).70 Multi-annual macroeconomic plans can offer greater flexibility, but tend to be institutionally and technically more demanding to become effective. Ensuring adequate effectiveness is a general problem for the use of ex-ante constraints. When faced with binding constraints, politicians try to evade them through creative accounting or collective defection. In situations where an ex-ante limit on deficit proves restrictive, politicians may manage to redefine budget categories to evade the rule or may collectively agree to disregard it. This is, in fact, what many governments do when they use their regulatory or ownership powers over the banking system to channel the system's resources toward activities that they would like to finance while keeping the activity off-budget. Similarly, politicians may ask private investors to implement parts of the government's programs with implicit public guarantees, but without explicit acknowledgment of the potential costs. Another common example is the use of "extra budgetary funds" (EBFs) to circumvent spending and deficit limits. Below, we further elaborate on the role of such off-budget expenditures under the rubric of comprehensiveness of budget. There are other factors that affect the effectiveness of ex-ante constraints as well. Even when the budget is comprehensive, the constraints may prove ineffective if the budget contains significant autonomous segments. That is, if there are rules that restrict policy-makers' control over parts of the budget and allow those parts to grow autonomously. For example, in some countries, laws or past political promises commit the government to maintain the interest rate or food and energy prices at certain levels through subsidization. When inflation rises for whatever reason, these subsidies 69 Another example is Indonesia's constitutional prohibition on domestic borrowing, which was intended to curb money printing. The effectiveness of such rules, of course, depends on a variety of factors, which we discuss below. 70 Eichengreen and Bayoumi (1994) show that balanced budget rules are associated with pro-cyclical fiscal behavior among the 50 states of the US. 130 automatically grow and force the ex-ante constraints to be compromised. This can further fuel inflation and create a vicious circle.71 Another factor that may undermine ex-ante constraints is the absence of hard budget constraints on public enterprises, subnational governments, or other public entities. If these units have access to channels to finance their expenditures with borrowings that cause liabilities for the national government, then ex-ante constraints may be violated ex-post. Also, if there are few ex- post restraints on revising the budget or ensuring that line ministries stay within their specified budgets, ex-ante constraints may have little consequence. Finally, when policy-makers are not accountable for the outcome of the budget, circumventing the constraints is likely to be a natural outcome. Despite these potential weaknesses, some countries have used multi-annual plans as a means of maintaining budgetary discipline. Most prominently Denmark and New Zealand, which faced critical budgetary conditions in the early 1980s, managed to raise the effectiveness of their macroeconomic planning arrangements and successfully brought their deficits under control. This does not mean that all other countries can achieve the same results with multi-annual plans, but it does indicate the possibility under certain conditions. We will discuss this issue and the reforms in New Zealand and Denmark below as part of our discussion on the nature of effectiveness characteristics and their interaction with country characteristics. To sum up, the presence of ex-ante constraints in a country should be associated with better budget discipline if the necessary mechanisms for ensuring the effectiveness of such constraints are in place. Thus, in analyzing budget institutions in a country, one must search for various types of ex-ante constraints that may exist in the country and examine their effectiveness characteristics. B. Coordination Mechanisms There are two different institutional arrangements for achieving coordination in the budget process when ex-ante constraints are not binding. The first one is the delegation of authority over the design and execution of the budget to an individual (or a few individuals) in charge of the "central budget agencies" (for example, Ministry of Finance, Ministry of Planning, Budget Office, or any other agency in charge of the budget and plan) that are independent of the line agencies. We will refer to this authority with the acronym, CBA. The second arrangement is an ex-ante agreement among the decision-makers (cabinet ministers, political parties participating in coalition governments, or members of the legislature as the case may be) concerning the levels of deficit and spending. In actuality, governments may use a combination of these arrangements and may rely on them to different degrees. The effectiveness of the two mechanisms also may vary according to the conditions under which they are applied. In particular, both arrangements require monitoring and enforcement to ensure that delegation is not abused and ex- ante agreements are not violated. We will examine the factors that can help satisfy these requirements after we discuss the two arrangements in more detail. Delegation Under the delegation approach, the budget drafted by the CBA is typically subject to approval by executive and legislative bodies. But, the CBA is given a great deal of power over the budget through "closed rule" voting arrangements. That is, the CBA controls the agenda and other decision-making bodies can only accept or reject its budget proposal as a whole. If the proposal is rejected, only the CBA can revise the detail items. This allows the CBA to prevent other decision-makers with differentiated 71 Such a process in Egypt drove the subsidy budget to about 10 percent of GDP in the 1980s. In Iran, the figure is much higher, but most of the subsidy is paid in hidden forms, such as zero valuation of raw oil sold domestically and the use of highly overvalued exchange rates for subsidized items. 131 interests from creating coalitions that overexploit public resources. The position of the CBA is further strengthened if the procedures are more hierarchical and, in the case of conflicts and repeated rejections, an interim solution is implemented that favors the CBA. Such powers, of course, create the potential for abuse. Moreover, concentration of decision-making power may distort the budget if the necessary information is not properly communicated and used by the CBA. Below, we examine the institutional characteristics that affect the effectiveness of the delegation arrangement from these perspectives. The use of delegation is common among developing countries. In all Latin American countries except the Dominican Republic, finance ministers have considerably greater authority than other ministers (IDB 1998, Table 2.6).72 Also, in most cases, legislatures are barred from making any changes in the budget that increase spending or deficit above those proposed by the government (IDB 1998, Table 2.7). Moreover, in many cases, such as Colombia, Chile, Costa Rica, Ecuador, and Peru, the government's position is further enhanced by an automatic enactment of its proposal in case the legislature fails to approve a budget (IDB 1998, Table 2.8). In most of the other Latin American countries, the government also remains largely in control of the agenda because a rejection by the legislature implies that the previous year's budget continues to be implemented on a pro rata basis while the government prepares a new budget. In no case is the legislature in a position to impose a budget of its own design on the central budget agencies. The same is true in most Asian and Middle Eastern countries. In Europe, France has the strongest form of delegation where the prime minister and the finance minister unilaterally decide the overall budget targets, determine the terms of voting in the legislature, and veto any amendments proposed by the legislature (von Hagen and Harden 1995, Tables A2, A3, and A4). In Germany, the finance minister has veto power in the cabinet and can change budget bids after consultation with the line minister concerned. At the legislative stage, changes in spending and deficit ceilings require the consent of government, which also has the power to postpone votes and call for a repeated vote. The budget procedure in the United Kingdom also strengthens delegation by giving the Treasury a special status and assigning the government the prerogative to restrict amendments and to make budget resolution a vote of confidence.73 As von Hagen and Harden (1995) show, these three countries have managed to adjust to macroeconomic shocks relatively fast. In most other European countries, delegation is weaker. However, only the ones that have used ex-ante agreement procedures effectively have succeeded in maintaining fiscal discipline (von Hagen and Harden 1995; Hallerberg and von Hagen 1997) Ex-Ante Agreements Ex-ante agreements can act as a substitute for delegation when a multitude of decision-makers (cabinet ministers or members of legislature) participate in the design of the budget. The simplest type of ex- ante agreement that can diminish the commons problem is a vote on total spending and deficit limits before debating and voting on the details of the budget. At that early stage, each decision-maker realizes that, although higher spending and deficit ceilings allow him/her to enjoy larger budgets, his/her share of each additional dollar in the aggregate ceiling is likely to be less than one and possibly uncertain. As a result, the expected marginal benefits of the aggregate budget for each coalition of decision makers are likely to be much closer to the expected marginal costs than when each coalition sees itself as the residual claimant, which is the case when the budget is voted on item by item. This 72 The IDB report indicates that the relative authority of the finance minister in Brazil is formal, but not practical. However, in a more detailed study of Brazil, Esfahani (1999b) finds that the Minister of Planning and Budget at the design stage and the Minister of Finance at the implementation stage have quite strong positions vis-à-vis other ministers. 73 In the UK, only government ministers may propose spending and taxing measures in the parliament. 132 effect allows the decision makers to act in a coordinated manner and move towards aggregate budget discipline. Of course, once the aggregate targets are agreed upon, the decision makers should find it in their interest to honor the implied constraints and should have assurances that the limits and the approved budget itself will be implemented. These require effective monitoring and enforcement mechanisms in the implementation process and a long-term perspective on the part of the coalition that backs the agreements. Ex-ante agreements often take the form of negotiation and voting over the broad characteristics of the budget before the details are discussed. This may happen in the process of coalition formation, when a multiparty government is formed, in the cabinet, when the cabinet is the main authority over the budget, or in the legislature, when the legislature has the prerogative to redesign the budget. The agreement may include limits or shares of the budget assigned to some general spending categories, as well as the aggregate figures. This is particularly the case in coalition governments where the participating parties need to have some idea about how the budget will be divided while maintaining some cap on total spending and the deficit. The inclusion of disaggregated limits and shares obviously weakens the coordinating effect of ex-ante agreements, but the outcome is still likely to be much better than collective decision-making without such a procedure. Another cost of disaggregate limits is, of course, allocational inefficiencies that they entail because the limits are set before the budget is drafted and all the possibilities and information about the resources and needs are explored in detail. Finally, ex- ante agreements have an inherent feature that reduces their effectiveness: in drafting the agreement, the decision makers anticipate an impact on their own shares of the budget and, compared to a dominant CBA, are less inclined to vote for full discipline. The Netherlands offers an example of a government that has used ex-ante agreements among coalition partners as a key discipline mechanism. The Netherlands' parliamentary system consists of a multitude of small parties that can participate in government only in the form of coalitions. The parties, participating in the government, start with an agreement over the budget and typically maintain the agreement because failing to do so is synonymous with the dissolution of the coalition, which creates a chance for each one to be replaced. The practice of two-stage voting on the budget as an ex-ante agreement is more common across countries. In many parliamentary systems (for example in Denmark, Ireland, and Portugal), the cabinet votes on budget aggregates at the start of the annual budget process. Often, the parliament also uses an initial vote to approve or disapprove the budget aggregates proposed by the government (as in New Zealand). In some presidential systems where amendments in the legislature are not restricted (for example, Iran), some discipline is reached via an initial vote on revenue and spending estimates. Such restrictions force legislators who want to propose amendments to various items to identify the expenditures that must be cut (or the revenues that must be raised) in order to fund their proposals. C. Effectiveness Characteristics of Aggregate Budget Discipline Institutions There is a host of factors that impact the effectiveness of delegation and ex-ante constraints. We categorize them into seven groups discussed below. Some of these characteristics could have arguably been included as part of the coordination or constraint mechanisms themselves. However, in our analysis of institutional arrangements, we consider the interactions of various factors, whether they are viewed as part of the arrangement or as effectiveness characteristics. As a result, the outcome is not sensitive to the method of classification and we, therefore, do not go into the details of such possibilities here. 133 Autonomous Budget Segments Governments usually allow some automatic adjustment in some categories of public expenditure. This may be desirable for stabilization or uncertainty reduction purposes. But, to ensure fiscal discipline, policy makers must have the choice to maintain or change such automatic adjustment. If laws, constitutional provisions, informal norms, or past political commitments constrain the government's choice and allow some spending categories to grow autonomously, the budget may become uncontrollable. Governments try to deal with such situations by cutting expenditures elsewhere. But, as those other parts shrink, the marginal cost of cutting them further rises and the government is eventually forced to abandon discipline or evade ex-ante constraints. Thus, the greater the degree of autonomy and the larger the share of autonomous expenditures in the budget, the less effective are ex- ante constraints and coordination mechanisms as a means of fiscal discipline. There are four expenditure categories that are most susceptible to becoming autonomous: subsidies, social spending, salaries, and defense. An example on subsidies was given above. Social spending (pensions, welfare, unemployment benefits, and other social payments) becomes autonomous if, for example, the constitutional clauses or common laws oblige the government to provide a nontrivial minimum level of real income to large segments of the population under relatively easy eligibility conditions. This has been a major issue in most developed countries that have built welfare states in the post-World War II era and had to strengthen their commitments in the face of adverse economic shocks in the 1970s. Ever since, social spending has been by far the largest public expenditure category in these countries. In most cases, the government has had to go through a long process of adjustment and institutional reform in the 1980s and 1990s to tackle the fiscal problems generated by such commitments. Nevertheless, the problem persists in some countries (for example Belgium, Spain, Portugal, and Sweden), and is likely to resurface in some others that have not brought about sufficient flexibility in social spending. Reform in this area has been a key ingredient of success in countries that have gained lasting fiscal discipline, such as New Zealand. Similar trends have been experienced in some developing countries where the government's obligations as a provider of social security have expanded. In Brazil, for example, the 1988 constitution committed the government to expensive pension and social assistance programs that were not linked to the fiscal resources of the country. Government salaries typically comprise another large segment of the budget with an autonomous component. Most governments implicitly or explicitly commit themselves not to reduce nominal salaries. In fact, in many cases the government is expected to raise nominal salaries every year. In some countries, however, the commitment goes further and the salaries are indexed to ensure that their real value does not decline. Salaries in some parts of the bureaucracy may also become more or less autonomous as a result of rules that allow the personnel in those parts to propose their own salaries (typically subject to executive or legislative approval). Such rules give the personnel the agenda power and impede the ability of the CBA or ex-ante constraints to impose discipline on the budget. Another source of autonomy in salary budget is government commitments to act as an employer of last resort. This type of obligation may be limited to small groups in the labor force. In some countries, however, it covers large portions of the labor force (for example, all university graduates and draftees in Egypt) and can become a major burden during economic slowdowns. Gaining flexibility in public employment and salaries is an important ingredient of reaching fiscal sustainability. It has certainly been a major contributor to New Zealand's successful adjustment (Scott 1996). Defense expenditures also have the potential for gaining autonomy because of the typical secrecy associated with them. This is particularly the case in countries where the military is politically strong and manages to determine the size of the defense budget by itself, outside the normal budget process. 134 For other segments of the budget, autonomy is often associated with the way budget baselines are established. In countries where budget baselines are extrapolations of past budgets simply adjusted for inflation, as in Australia and the UK before the 1980s, running costs have a life of their own and new programs or expanding ones have to be financed through new taxes or deficit financing. An alternative is to let the baselines grow slower than the rate of inflation and have the ongoing activities compete with the new programs over the cost savings (or "efficiency gains," as they have come to be known). Since the 1980s, a number of countries, including Australia, Denmark, and the UK, have adopted such an approach. New Zealand has addressed the problem by establishing a process of regular program evaluation and by exchanging agency autonomy for reduced budget requests (Scott 1996). Note that the rules that give autonomy to parts of public expenditure are in a sense similar to ex-ante constraints, although they act as floors rather than ceilings. This distinction makes a great deal of difference in terms of viability of the two sets of rules. Unlike ex-ante constraints, autonomy rules are more likely to be observed because such rules play important distributive roles and benefit specific interest groups that provide political support for their enforcement. Comprehensiveness of the Budget National budgets differ widely with respect to their coverage of subnational government budgets, subsidies to, and contributions by, public enterprises, and contingent liabilities of the government (for example, explicit and implicit guarantees of public or private enterprise debt). There is also a great deal of difference in the role of EBFs and hidden subsidies (such as rents distributed through multiple exchange rate systems, bank regulations, and below market pricing of natural resources owned by the government). Although keeping some funds and expenditures off-budget provides policymakers with some flexibility, it can easily undermine fiscal discipline because it creates strong incentives for the politicians to channel resources towards more discretionary types of expenditure and fill in the shortages created elsewhere by deficit financing. EBFs may also cause serious allocational inefficiencies because the costs and benefits of off-budget expenditures are not directly compared with those of the funds included in the budget. As a result, the effectiveness of coordination mechanisms and ex-ante constraints is likely to decline as the comprehensiveness of the budget decreases. This is particularly important in the case of ex-ante agreements and ex-ante constraints where the whole process may be undermined by funds that are allocated outside the budget. An example of a country that has suffered macroeconomic instability partly as a result of EBFs is Turkey, where the expenditures of such funds reached about 11 percent of the GDP in 1990 (Önis and Webb 1994). With the possibility of spending public resources through EBFs without prior legislative authorization, the Turkish government had a strong incentive to reduce taxes in some areas (for example, by reducing tariffs in the name of trade liberalization) and then re-imposing new taxes on the same bases to fund EBFs. The result was that in the late 1980s, EBFs absorbed about one-half of all government revenues. This phenomenon contributed to the continuation of large budget deficits, which have been the main force behind high inflation rates in Turkey. As part of its reform attempts, Turkey has been eliminating EBFs. Complete removal of EBFs has been an integral part of New Zealand's reform process. Inclusion of subnational governments and public enterprises in the budget is important because the activities of those entities have implications for the national government that must be taken into consideration. Even in federal systems where states have autonomy over their budgets, the national government cannot ignore the financial difficulties that the state governments may face and the public service needs that may go unattended. Rather than waiting to be faced with difficult situations and 135 "unanticipated" expenditure needs, the national government can do better by closely monitoring the activities of lower level governments and by making provision and taking preventive measures that preserve the integrity of its fiscal policy. Naturally, it is best to consider these issues as part of the budget exercise. There are many other contingent liabilities that should also be carefully evaluated for budgetary purposes. Most governments set aside certain emergency funds to deal with various risks. However, this is often done in an ad hoc manner without systematically assessing financial risks involved in all government activities, including guarantees and indemnities. Politicians may prefer the laxity of budget procedures in this respect because it gives them a leeway to propose and approve budgets that look balanced, but frequently require significant adjustment or yield deficits during implementation because of "unforeseen events beyond the government's control." The experience of fiscal reforms in New Zealand, which require systematic assessment of risks for all programs and policies in bilateral negotiations between the CBA and line ministries, indicates that accounting for contingencies can be quite effective in improving fiscal performance (Scott 1996). Ex-Post Restraints When the budget is approved at the beginning of a fiscal year, it may be difficult to foresee all possible contingencies. As a result, during the year, as uncertainties are resolved, the government may find that the approved budget is no longer desirable. Budget laws often allow limited changes in the budget during the year, especially if revenues fall short of the predicted levels and spending cuts are necessary to contain the deficit. However, if the government has easy ways of increasing spending during the year through supplementary budgets, then the coordination and allocational efficiency of the budget will be lost. The reason is that the possibility to change spending invites pressure on the government to expand or reallocate the budget ex-post. More importantly, anticipation of such changes makes it difficult for those who participate in the budget design to take the exercise seriously and weakens their incentives to make difficult choices about discipline and priority. While the problems that expenditure increases pose for fiscal discipline are well recognized, lack of restraint on expenditure cuts is often viewed as a requirement for ensuring discipline in case of revenue shortfalls or emergency expenditure needs. This view, however, ignores the fact that when the option to cut expenditures ex-post exists, policymakers may have an incentive to overestimate revenues and set the expenditure ceiling high. In that case, a "revenue shortfall" is always assured and the policymakers will have their options open to spend as they see fit ex-post, which effectively weakens any restriction on ex-post expenditure increases. Besides, line agencies that anticipate expenditure cuts will have little incentive to plan ahead and will be spending part of their energies to lobby for the "release" of the funds allocated to them in the budget. Obviously, this will undermine a major purpose of budgeting and lead to considerable misallocation and operational inefficiency. One may, of course, argue that removing the ability of the government to cut spending in case of resource shortages could lead to macroeconomic instability, especially in countries where the government faces a borrowing constraint and excess spending has to be financed by printing money. But, if that is the problem, it should be solved by creating revenue stabilization funds and by using prespecified rules rather than discretion to cut spending. Budget modifications, with or without revenue shortfalls, are not uncommon across countries but the stringency of the rules varies greatly. In most countries, the CBA is given the authority to impose expenditure cuts (sometimes according to prespecified formulas) in the case of revenue shortfalls, typically with a legislative approval requirement if the cuts are large. However, some countries (for example Italy, New Zealand, and the United States) restrict the authority of the CBA in this respect and 136 give spending agencies more control over their appropriations. The common practice for "non- emergency" situations is to authorize the CBA to carry out small modifications, but require a new law for larger changes. Margins of up to 10 percent of the appropriations for changes without legislative approval are not unusual (as in El Salvador, Iran, and New Zealand). However, higher margins, such as the 20 percent margin in Brazil, tend to be excessive. In some countries (for example, Mexico), the government has even more room to maneuver and budget modifications do not require a new law. The polar case is where no modification is allowed during the year (as in Uruguay). Another dimension that differentiates countries in terms of ex-post restraints is whether modifications that require a new law are initiated by the executive or by the legislature. While the former arrangement is quite common, some countries, such as Germany, Bolivia, and Guatemala, have opted for the latter, which is more restrictive because coordination in the legislature is more difficult than in the executive. Other conditions imposed on budget modification may also be important. In particular, a very sensible requirement is that supplementary budgets be approved together with the annual or multi-annual budget plan for the following years. Such a joint consideration forces the government and the legislature to be more careful with adjustments that have longer-term consequences. Most countries ignore this issue and consider supplementary budgets by themselves. New Zealand has introduced the requirement as part of its recent reforms. Finally, the extent of ex-post restraint depends on the presence of spending limits on line ministries. Many countries (as with Denmark, Germany, Greece, France, Portugal, and the UK) impose cash limits on the spending departments. Some, like New Zealand, however, let the line ministries overspend as long as they can justify it to the legislature. The extent of accountability of public officials, of course, is a major factor in determining whether such powers are abused or not. Hard Budget Constraints on Subnational Government and Public Enterprises In many countries, subnational government and public enterprises have some authority to borrow and spend independently of the national government budget. Because these entities carry out activities that are of interest to the national government, they may be in a position to oblige the national government to bail them out when they run into financial difficulties. When this possibility exists, the government may have to shoulder continually large debts over which it has little control. The severity of this effect depends on the borrowing options open to subnational governments and public enterprises. The problem is more serious when they have access to state-owned banks or they can easily obtain explicit or implicit guarantees for their debts from the national government. Large amounts of debt passed on to the government by public enterprises in the late 1980s in Brazil and in the 1990s in Iran have been a major source of fiscal instability in these two countries. Brazil's current fiscal difficulties partly originate from the debt that states borrowed from their own banks and transferred to the national government when the country started to face the prospect of financial collapse as state banks became insolvent. In all such cases, "softness" in the budget constraint facing the subnational governments and public enterprises destroys budget discipline and causes misallocation because the capability to overspend overtakes considerations of social costs and benefits. Lack of hard budget constraints also diminishes the value of the initial budget design and undermines any coordination mechanism, whether it relies on delegation or on ex-ante constraints. 137 Expected Tenure of Decision-Makers A delegation system loses its effectiveness if the CBA leaders expect their term in office to be short, because they may be able to avoid the full consequences of their actions. A natural tendency in that case is for the CBA leaders to use their power to overspend in areas of their interest, which are often politically motivated. Expected tenure also matters under an ex-ante agreement system because the coalition that backs the agreement may fall apart more easily if the partners don't expect to be in their decision-making positions for long. However, the effectiveness of ex-ante constraints is less likely to be affected by the tenure of CBA leaders. The expected tenure of CBA leaders is a function of the rate of turnover in the government, which itself depends on the electoral system and the underlying political stability of the country. Typically, parliamentary systems with proportional representation produce coalition governments that do not last very long. Parliamentary systems with plurality electoral rules and presidential systems tend to generate longer expected tenures for the CBA leaders. Although no country seems to have institutional arrangements that give the CBA leaders long-term tenure and independence from politics, in principle it is feasible and may offer a useful option for obtaining fiscal stability. The idea of such an arrangement has been recently proposed by a number of authors in the form of an independent "National Debt Board," which sets the limits on public debt and budget deficit (von Hagen and Harden 1995, IDB 1998). Educational Levels of Decision-Makers Lack of sufficient education, especially in economics-related areas, can make it difficult for the policy makers to reach informed decisions. Under a delegation approach, usually this is not an important problem because the CBA leaders and their associates are often selected from among those with appropriate levels of education. However, ex-ante agreements involve large groups of decision makers and can work well if educational levels are generally high among the country's policy makers. The use of macroeconomic programs can also be more effective if the policy makers are better educated and can work with more sophisticated models. Accountability For any budgetary procedure to work effectively, those who are responsible for the design and implementation of the budget must be accountable for the outcomes. Fiscal accountability has a number of dimensions. To begin with, the budget must be based on sound accounting principles and must be audited properly. In this respect, the quality of the audits, the length of time that it takes for the audit reports to be prepared, and the independence of the auditors from the government are key factors in identifying financial outcomes and determining responsibility. Accountability also requires clear sanctions that would apply to those responsible for the outcomes in case the budget performs badly. Publication of the government's explanations about the deviations of the implemented budget from the approved one is another factor that enhances accountability, especially if the budget information is easily accessible and the government's reconciliations are accompanied by public debates. Freedom of the press, democratic political institutions, and a higher average level of education among the public can further help raise the degree of accountability for policy makers. Finally, openness of financial markets in the country can serve as a quick punishment mechanism for policy makers because it allows investors to respond to unsustainable fiscal policies by denying the government an easy source of borrowing. This also produces an immediate signal about the nature of government policies and puts further pressure on policy makers as the withdrawal of funds weakens the economy. 138 Enhancing fiscal discipline through increased accountability has been a key part of New Zealand's successful reform program. The reforms have included a move to commonly-accepted accounting practices, an efficient system of audit with independent auditors, clear definition of responsibilities with performance measurement and well-specified sanctions in case of failure, and extensive reporting and proactive publication of the outcomes. D. Allocative Efficiency As pointed out above, ex-ante institutional constraints that precommit some parts of the budget can have serious adverse effects on allocation. Delegation and ex-ante agreements may also have adverse effects on the allocative efficiency of the budget, which can be reduced by a number of institutional mechanisms. The main mechanisms of this kind can be grouped into four categories. As in the case of institutional arrangements affecting the aggregate budget discipline, the performance of the allocation mechanisms is also affected by effectiveness characteristics similar to those listed above. Prioritization through Objective Criteria It is well-known that allowing the systematic use of cost-benefit analysis with a well-established set of assumptions and methodology to guide prioritization among projects and expenditure items can greatly improve the allocation of the budget. However, the technical and institutional requirements of performing such analysis discourage widespread use of such mechanisms for budget design. But reforming countries such as Australia and New Zealand have found it essential to extend the use of objective criteria in program and activity assessment. Another tool, which is complementary to cost- benefit analysis, is a national development plan with targeted macro and sectoral level outcomes–for example, the infant mortality rate, school attendance rate, and percent elementary graduates scoring 75 percent or better on national exams. The result can be particularly helpful if both tools are used with medium- to long-term horizons. The use of objective criteria is likely to be less effective when the policy makers are less accountable, the budget is less comprehensive, and autonomous segments play a larger role in the budget. Among other factors that tend to reduce the effectiveness of prioritization are lack of ex-post program evaluation mechanisms and insufficiently educated line ministry administrators. There are also a number of procedural aspects that some governments use to ensure minimum shares for some programs or to facilitate the task of drafting the budget, but have detrimental impacts on prioritization. For example, many governments earmark certain revenues for specific activities or separate capital and current budgets, which sometimes allow line agencies to make their capital budgets without much regard to their future operational costs. Like off-budget and autonomous expenditures, such lack of unity in the budget can increase the divergence in the returns across budget items and weaken prioritization. Another important characteristic of the budget process that can cause a similar disunity problem is the ability of the legislature to introduce new projects into the budget without the consent of the CBA. Since the new projects do not meet the same criteria as those in the proposed budget, they give rise to some inconsistency in the budget. Moreover, the possibility of adding new projects creates an incentive for the legislators to allocate part of the budget to start up their pork-barrel projects in order to increase the likelihood of funding such projects in future years. This divides up the budget into inefficiently small pieces and prevents the high-return projects from receiving sufficient funding and finishing on time. In addition, it reduces the accountability of the executive, who may be able to blame any implementation problem on the introduction of inappropriate projects by the legislature. Even if new projects require the consent of the executive, the possibility of additions creates pressures in the legislature that may be 139 passed onto the executive to extract its approval. Of course, the legislators can have a legitimate role in the selection of projects when they act as a check on the power of the CBA. But this does not mean that they should be able to force projects into the budget at the legislative stage. Participation of Line Agencies and Ministries in Planning Although involvement of line agencies and other claimants in the budget decision can create difficulties for the aggregate budget, their participation in the planning process can help policy makers tap on their information without losing control. This can be done, for example, by asking line agencies and ministries to support their cases for spending via multiyear forecasts of the full costs of their programs, including the cost of contingent liabilities, capital use, and subsidies and transfers. Another example is when line agencies/ministries are allowed to negotiate multiyear agreements with the CBA for their spending programs. Such arrangements can be reinforced by asking line agencies to present cutbacks in their spending on existing programs to potentially offset any proposal they make for new spending. Australia, Denmark, and New Zealand have successfully incorporated such features in their reformed fiscal institutions. The participation of line agencies is, of course, more productive when the budget is more comprehensive, more unified, and less autonomous. Technical capability of line agency administrators also raises effectiveness. Flexibility of Line Ministries While the CBA, the executive, and the legislature are the main players in strategic priority-setting and determining budget aggregates, their information about the best way to allocate resources within programs undertaken by ministries is limited. As a result, detailed control of input in each ministry's activities is bound to cause allocative inefficiency. In other words, much efficiency can be gained by giving line ministries flexibility over how they want to allocate their budgets subject to total limits. Over the past couple of decades, Australia, Denmark, and New Zealand have introduced such arrangements with satisfactory results. Careful program evaluation, accountability, and capability of line ministry administrators are key factors that determine the effectiveness of line ministry flexibility. Breadth of Consultations Participation can be much broader than the line agencies. The government can encourage participation from civil society (business community, workers’ associations, nongovernment organizations) to provide input into the planning process. The information that those groups provide, and their interactions with each other and with the government, can better identify the priorities and build broader consensus for them. This can facilitate the task of choosing among projects and programs subject to the aggregate restrictions imposed by ex-ante agreements or by a strong CBA. It is easy to see that all the effectiveness characteristics that affect other allocation mechanisms (accountability, comprehensiveness, unity, low autonomy, program evaluation, and the extent of the polity's education) matter for the efficacy of consultations as well. 140 E. Operational Efficiency Operational efficiency implies that the resources allocated to government agencies are used productively in the implementation process to generate the intended public services.74 The institutional arrangements that influence aggregate discipline and budget allocations affect the possibilities for operational efficiency as well. In particular, gross misallocations and uncoordinated spending are likely to give rise to waste in the implementation process. However, the outcome is also conditioned by the bureaucratic institutions of the country. The relevant administrative arrangements can be summarized under two categories: the autonomy of line agencies and merit-based recruitment/promotion. Autonomy of Line Agencies As in any other principle-agent problem, the government can motivate line agency managers to perform well by evaluating their output and rewarding them accordingly, or it can control the inputs to make sure a minimum level of service is delivered. The two approaches are to a large extent mutually exclusive because if the government controls the inputs (for example equipment and material required or the number of workers and their promotion and reward decisions), it will have a hard time blaming the manager in the case of diminished performance. The problem lies in the difficulty of distinguishing the contribution of the manager from the role of government controls, which are necessarily imposed with imperfect information about the needs and conditions of the agency. Moreover, government control of inputs often entails day-to-day political interference, political appointments, and budget manipulations that greatly complicate the task of assessing the manager's performance. As a result, the government has to choose between giving a line agency autonomy and controlling it tightly on the input side. There is plenty of evidence that the first option generates better outcomes if the government develops reliable mechanisms for assessing the output of the agencies and holding the managers accountable for the outcomes.75 These mechanisms consist of procedures, such as client surveys, audits, performance reports, and clear rules for ensuring the competitiveness of procurements. All of these mechanisms can be further enhanced by the availability of information about the agency's activities and by freedom of the press. Other important factors for the effectiveness of agency autonomy are availability of support resources (equipment, travel and training funds), the predictability of resource flows to the agency, and the duration of the manager's tenure, which must be sufficiently long to allow proper assessment and reward for the manager. The switch to agency autonomy with performance-based accountability has been a hallmark of fiscal reforms in Denmark and New Zealand. Merit-based Recruitment/Promotion To encourage professionalism among public employees, the government must provide them with long- term careers, restrict the range of political appointments, and use objective performance criteria to recruit, reward, and promote personnel. Such arrangements are more effective when the salaries of public employees are competitive with those in the private sector and the general level of education is higher (so that there are better human resources for the government to recruit from). Availability of support resources and predictability of resource flows to line agencies may also help the incentives remain stronger within the bureaucracy. F. Budget Procedures and Country Characteristics 74 Efficiency can be measured, for example, as the inverse of the unit costs for quantifiable dimensions of services. 75 This is, for example, what often happens when state enterprises are corporatizated. 141 All budgetary procedures that help aggregate discipline or allocative and operational efficiency entail some costs as well. For example, both delegation and ex-ante agreement procedures that help solve the common-pool problem may limit the possibilities for the exchange of information and equalization of returns across budget items. An important implication of this observation is that the effectiveness and optimality of various procedures may vary across countries because the tradeoffs between the costs and benefits of each procedure are likely to depend on some country characteristics. Current theoretical and empirical knowledge in this area is still quite limited and the following analysis of interactions between country characteristics and budget procedure is necessarily tentative. We focus on two country characteristics that seem to play important roles in the tradeoffs of budget procedures–namely, the significance of economic shocks (for example, terms of trade fluctuations and major climatic effects) and the heterogeneity of the polity in terms of characteristics, such as ethnic backgrounds and economic assets. For the sake of this analysis, let us consider a benchmark country, B, which has a uniform polity and does not suffer from frequent and strong economic shocks. In such a country, reaching coordination through ex-ante agreements should be relatively easy because the homogeneity of interests reduces the likelihood of disagreements. Moreover, the absence of strong fluctuations should allow the agreements to remain reasonably efficient ex-post. The delegation approach is likely to be relatively less attractive in such an environment because it restricts effective participation in decision-making, while some of its important benefits– quick and effective decision-making in response to shocks –are not of much value. Ex-ante constraints are also likely to be relatively less beneficial in such a situation, because their function can be replicated by agreements, which can be adjusted annually in a more flexible way. As a result, the conditions of the benchmark country are the best for the ex-ante agreements to work and such procedures are more likely to survive in that environment. An actual example that seems to resemble this benchmark case and does rely on agreements is Denmark. Now, suppose there is another country, A, which is identical to B in every respect except that it faces more significant economic shocks. Since A has a greater need for adjustment, it can benefit more from delegation than B because that procedure reduces the costs of repeated negotiations over budget allocations as conditions change. Stated differently, if country A chooses to rely on ex-ante agreements, it will experience greater difficulty than country B. This is, indeed, consistent with the empirical findings. For example, von Hagen and Harden (1994) and Alesina et al (1999) found that European and Latin American countries that had more hierarchical budget procedures have responded more quickly to the economic shocks of the 1970s. Next, suppose there is a third country, C, which is again similar to B, except that its population is more heterogeneous in terms of economic and other interests. In country C, bargaining among interest groups is likely to be more costly. As a result, moving towards procedures that reduce the need for interest group bargaining should be more beneficial in country C than in country B. That is, the increase in polity diversity tends to increase the payoffs from delegation. This observation helps shed light on the pattern of budget procedures and fiscal performance across Europe. In particular, among large countries with more diversified polities, the ones that use stronger delegation– France, Germany, and the UK –have conducted more successful fiscal policies than those that avoided delegation; for example, Italy (von Hagen and Harden 1994). On the other hand, the countries that have reached fiscal discipline via reliance on ex-ante agreements are all small countries, such as Denmark and New Zealand. To reach allocative efficiency, heterogeneity and frequent economic shocks raise the contributions of all four institutional arrangements. In more heterogeneous and shock-prone countries, prioritization 142 through objective criteria can play more important roles in helping policy makers reach agreements over programs and allocate public funds efficiently. In addition, in such countries decision-making requires more information and more flexible responses, which raises the benefits of broad consultation and line ministry participation and flexibility. Similar effects help increase the operational efficiency gains from line agency autonomy as heterogeneity or shock intensity rise, but the effects of merit-based recruitment and promotion on institutional arrangements are unlikely to be systematically related to such country characteristics. 1.2 Methodology For empirical assessment of a country's budget institutions, three sets of variables must be measured: (i) measures of budget institutions and their effectiveness; (ii) indicators of fiscal performance, such as government spending and deficit relative to GDP, as well as the structure of spending on various programs and their outcomes; and (iii) indicators of country characteristics, such as heterogeneity and economic shock frequency. For country characteristics, a variety of sources offer useful data. To make comparisons across countries in terms of diversity of the polity, we employ the degree of ethnolinguistic heterogeneity available from Easterly and Levine (1997) and the income distribution Gini coefficient collected by Deininger and Squire (1996). The former, which measures the probability that two randomly selected individuals in a country speak different first languages, reflects ethnolinguistic differences that can potentially hamper cooperation. The latter variable shows the diversity of access to economic resources, which again can intensify distributional tensions in the budget process. The share of nonagricultural sectors in GDP can also be a useful indicator of the extent of variety in economic activities and assets. This measure, in combination with the size of GDP and population, also reflects the degree of specialization and potential divergence in interests across interest groups. The data for these variables are directly available from the World Bank's World Development Indicators. Finally, data for a variety of institutional characteristics other than budget procedures can be obtained from the Cross-national Time Series Data Archive (Computer Solutions Unlimited 1998), Polity III (Jaggers and Gurr 1996), and International Country Risk Guide (Political Risk Services Inc. 2000). The first source provides (mostly factual) panel data on a series of variables characterizing the political system, such as the type of government (presidential, parliamentary, monarchy), the size of legislative majorities, and the like. The second source offers a set of cross-country, time-series rankings about political openness, the extent of constraints on the executive, degree of centralization, and so on. The sources for the second group of variables are country macroeconomic databases and the IMF's Government Financial Statistics. Additional data concerning the proposed and approved budgets are directly collected from the country under study. To measure outcomes, each case study makes an effort to use the existing studies of government services and combine them with the broadly available data on health and education outcomes across countries. The third source contains survey-based panel data about rankings of countries with respect to risk of contract repudiation by the government, expropriation risk, rule of law, and bureaucratic quality. To build measures of budget institutions based on the analytical framework, we use a questionnaire that helps rank various procedures in terms of their contributions to budget discipline and its allocational and operational efficiency. The information required for the questionnaire is obtained from studies of the budget or from country experts, knowledgeable about the budget or involved in the process. The answer to each question is obtained from more than one source to cross-check the factual information, 143 which is the bulk of the required data, and reduces the possibility of bias in case of questions that require some judgment. Below, we review the questions and the indexes based on them. A. The Questionnaire The questionnaire is a diagnostic check on the budget process and consists of three sections that correspond to procedures relevant for aggregate discipline, allocative efficiency, and operational efficiency. There are a total of 44 questions, most of which consist of multiple parts. The first 24 questions focus on aggregate discipline; the next 10 are concerned with allocative efficiency, and the last 10 examine operational efficiency. Each question produces an index that ranges between 0 and 1, with higher scores representing greater procedural capability for reaching discipline and efficiency. The indicator column in the questionnaire is where the answers to the questions are recorded by entering 1 next to the most appropriate answer. The column to the left of the indicators shows the weights assigned to each answer and the score column shows the results of the aggregation calculations. In the case of questions with multiple parts, the score for the question is typically a weighted average of the score for the individual parts. However, there are some cases where the calculation is somewhat involved and will be discussed as we examine those questions. The scores for the questions are then combined to calculate indexes for various institutional arrangements and effectiveness characteristics. The questionnaire has been placed in a computer workbook that makes all these calculations automatic. Aggregate Fiscal Discipline Questions The first three questions ask to what extent the CBA dominates the budget process; that is what delegation mechanism is used. Question 1 examines the procedures before the legislative stage. It consists of three parts. The first part inspects the negotiation process between the CBA and line ministries and returns a higher score when the CBA has the upper hand. The second part determines whether the CBA has a strategic position in the cabinet, and the third part checks whether there are major changes in the budget aggregates as it passes through the cabinet. This last piece of information essentially verifies whether the actual process is contentious or has a dominant side. The results of the three parts are aggregated with weights of 0.40, 0.40, and 0.20, respectively, to come up with the score for Question 1. Question 2 explores the legislative stage of the budget process. It consists of four parts. The first part documents the extent of agenda control by the CBA which, at this stage, often represents the executive. The second and third parts ask what happens in the case of conflict between the CBA and the legislature over the budget. They cover both the interim events while the conflict is being resolved and the ultimate powers of the two sides. The last part checks the extent to which total spending changes in the legislative process. As in Question 1, this information verifies the extent to which the CBA-legislature relationship is actually one-sided. When aggregating these four parts, the last one is assigned a weight of 0.20 and the other three equally share the rest of the weight, 0.80. To arrive at a measure of dominance of CBA in both government and legislative stages, we combine the scores for Questions 1-2 by forming their product. The reason for this procedure is that if dominance is weak in either stage, discipline may be lost. The use of ex-ante agreements is measured by Question 3. It has two parts that examine the arrangements in the executive and the legislature. The question produces a higher score when there are definite procedures that produce explicit agreements over budget aggregates. Question 4 evaluates the use of macroeconomic programs in two dimensions: the time-horizon and the extent to which the borrowing and spending are constrained by the model's estimates. Question 5 documents the spending 144 or deficit limits commonly accepted through informal rules or specified in the constitution or other laws. It consists of four parts, the first two of which ask whether ex-ante limits exist at all on deficit and spending and, when limits exist, whether they are explicit or implicit. The third part checks whether there are limits on program spending to support the aggregate limits on spending to ensure that politicians do not purposefully underestimate the budgets of urgent programs to create room for their favored projects, and to force the violation of the aggregate limits due to the funding needs of the latter group. The fourth part inquires about the time horizon and the frequency of adjustments in the limits, with the view that too frequent revisions undermine the role that ex-ante constraints can possibly play. To calculate the score for this question, we view an explicit limit on deficit and debt as the best option, with implicit limits and constraints on aggregate and program spending as less effective options. Therefore, we give half as much weight to the latter two options than to limits on deficit and debt and score the question according to the highest outcome between the first option and the sum of the latter two options. We discount the total score for Question 5 as the time horizon of the ex-ante limits shrink. The remaining aggregate discipline questions assess the effectiveness characteristics of the above arrangements. Question 6 analyzes the nature of baseline calculations for the running costs and Questions 7-10 document the extent of autonomy in subsidies, wages, social security spending, and the defense budget. Each of these four questions has a part that determines the weight of its corresponding expenditure category in the budget. These weights are used to calculate a score for each question that reflects the importance of its autonomy in the sum of the four categories. The formula that calculates the score for Question i, Si, i = 7,…, 10, is Si = [mi − wi (1 − si)]/(Σj mj) j = 7,…, 10, where wi and mi are, respectively, the actual and maximum shares of the expenditure category corresponding to Question i and si is the aggregate score for parts of Question i that assess freedom from expenditure autonomy. The overall score for the four questions is the sum of their Si's. This formulation ensures that both the extent of autonomy, 1 − si, and the share in the budget, wi, reduce the score and that the total for the four questions weighs the four categories appropriately, while remaining between 0 and 1. This overall score is combined with the score for Question 6 with weights of 0.85 and 0.15 to form the score for the extent to which autonomous budget segments allow the fiscal discipline to work. Question 11 deals with the comprehensiveness of the budget and has eight parts that examine the inclusion of subnational and public enterprise budgets, the extent of extra budgetary funds, the size of hidden subsidies, the assessment and use of contingent liabilities of the government, and the extent to which implicit and accrued expenditures are recorded. Parts F and G both deal with the issue of contingent liabilities and are not independent of each other. We set their joint weight in the total score for Question 11 equal to the weight of each of the other five parts. Question 12 attends the ex-post revision restrictions in five parts. The first four parts examine the possibility of revisions in the absence of revenue shortfalls or extraordinary conditions. Part A is concerned about the CBA's ability to make budget cuts, part B focuses on increases and other modifications, and Part C checks the timing of the revisions. Part D is concerned about the existence of spending limits on line ministries. The last part documents whether the CBA has the ability to impose cuts in case of revenue shortfall or extraordinary conditions. The score for Question 12 is a weighted average of the five components. 145 Questions 13 and 14 evaluate the strength of budget constraints for subnational governments and public enterprises. Various parts of these questions investigate the possibility that some public agencies may find mechanisms to spend beyond the constraints specified in the approved budget, with the national government accommodating them or being unable to close the loopholes. Question 13 captures the fact that the easiest means of constraint evasion for subnational governments is to have their own banks to borrow from, especially when there are deposit insurance schemes combined with poor regulations or some expectation that the national government may bail out provincial banks. Being able to borrow from the national government itself or from multilateral and private sources are other channels of expanding the budget, although more difficult ones. Multilateral and private sources would be demanding if the loans are not guaranteed by the national government. If the national government provides such guarantees but does not properly estimate, debate, and fund them in its budget, then the potential for runaway spending may be significant. This aspect will be captured by the interaction between the scores for hard budget constraints and budget comprehensiveness (see below). Question 14 examines the strength of budget constraints for public enterprises by asking how deficits of such enterprises are financed. The scoring of the arrears of public enterprises in the last part of the question supplements the institutional assessment of the first part. The size of the public sector (as measured by its investment share in the second part of Question 14) determines the weight that must be placed on the risks the public enterprises may pose. Questions 15 and 16 document the information regarding expected tenure of the CBA, and the educational level of the decision makers. Assessing accountability is more involved because it has eight components. Question 17 evaluates the financial audit system in terms of independence of auditors, the time it takes for the audits to be performed, and the quality of the reports. Question 18 asks whether the government has to provide explanations for the deviation of various aspects of the implemented budget from the approved one. Question 19 deals with what happens if the deviations are not justifiable. Question 20 captures the extent to which information about the performance of the budget is available to the public. Questions 21-24 record the freedom of the press, extent of democracy, average years of schooling in the total population, and a measure of financial openness. The data for these questions is included in a workbook along with the questionnaire. The ranking regarding press freedom is provided by the World Human Rights Guide (Humana 1992). The extent of democracy index is from the Polity III data set. The data for the average years of schooling is based on Barro and Lee (1996) and the score used in index calculations is calculated by dividing the log of the average years of schooling by its maximum value (log 14) across all countries. Financial openness is measured by the stock market turnover rate (that is, the value of the trades of domestic shares on domestic exchanges over the year, divided by the average value of domestic shares listed on domestic exchanges in that year) analyzed by Levine and Zervos (1998) and available on the World Bank website. The score for overall accountability is an average of the scores for the Questions 17-24. To calculate the effective scores for the coordination and constraint mechanisms, the scores obtained for the presence of arrangements in the budgetary process are combined with the relevant effectiveness characteristics through a matrix of weights. The weights show which effectiveness characteristics matter for which institutional arrangement. For simplicity, all the effectiveness characteristics that affect an institutional arrangement are assigned equal weight, although one may want to adjust the weights to see the sensitivity of the results with respect to them. Since the characteristics are complementary in enhancing the effectiveness of the arrangements, they are combined by calculating their geometric mean with the weights acting as the exponents. This aggregate effectiveness measure is then multiplied by the score for the institution itself to arrive at effective scores for the institutional arrangements. 146 The aggregate discipline calculations discussed above is summarized in Table 1, which provides an example of the computed indicators for New Zealand based on the information provided in Scott and Ball (1996) and Campos and Pradhan (1996). The first column in Table 1 shows the scores for the presence of various arrangements. The first row shows the scores for the effectiveness characteristics, which are listed in columns 2 through 8. The next eight columns show the components of accountability (audit, ex-post reconciliation), which are important pieces of information in their own right. The numbers below the first row in columns 2-8 are the weights assigned to various effectiveness characteristics when the effective scores of coordination mechanisms are calculated. Finally, the last column shows the effective scores. The example in Table 1 shows that New Zealand essentially relies on macroeconomic programming and ex-ante agreement to ensure fiscal discipline. It has no institutional limit on borrowing or spending and does not place the CBAs in a position of dominance. Allocational Efficiency Questions Question 25 asks whether objective criteria are used for prioritization. It combines with equal weight the scores for the use of macroeconomic framework, systematic cost-benefit analysis, and the time horizon of public expenditure plans. Question 26 is concerned with the participation of line ministries in the planning and budgeting process and documents whether they can make a case for their spending programs and provide rankings for the adoption or deletion of projects. Question 27 examines the control of line ministries over the allocation of their current and capital budgets and produces a higher score when they enjoy greater flexibility. Question 28 gives points for broader consultation during the budget process within the government and legislature and with civil society. The effectiveness characteristics of the arrangements needed for allocative efficiency are partly the same as those applicable for aggregate discipline. In particular, budget autonomy and comprehensiveness are important for allocative efficiency as well. Education level of administrators also matters, although mostly at the level of ministry administrators. Question 31 is designed to provide ranking in this regard. Similarly, accountability characteristics that matter for allocational purposes are akin to those measured for the government as a whole, but they mostly refer to ministry levels. This is done in Questions 32-34, where program-level ex-post reconciliations, explicit sanctions, and availability of information to the public are documented. Freedom of the press, democracy score, and the general education levels of the public are also included in the accountability measure the same way that they are used for aggregate discipline. Two new effectiveness characteristics are used for allocative efficiency: budget unity and program evaluation. Budget unity is considered in Question 29. Part A of this question starts with the extent of earmarking, Part B asks whether full costs are considered in each program, and Part C records whether capital and current expenditures are integrated. Part D deals with the legislature's ability to introduce new projects. There are three additional parts in Question 29 that measure the deviations of the sectoral allocations as the budget goes through various stages. These parts, which receive small weights in the total score, act as a check on the extent of agreement among various decision-making bodies and the closeness of implementation with the plan. Program evaluations, which are the subject of Question 30, represent cost-benefit analysis exercises that measure the effectiveness of the implemented programs ex-post. Carrying out such exercises can help improve the assessment of existing and future programs. Moreover, policy makers may be more careful in their assessments if they know that such evaluations will be made in the future. The calculations for arrangements that are relevant to allocative efficiency are summarized in Table 2. As in the case of Table 1, the numbers in the first column show the scores for the presence of the 147 arrangements and the numbers in the first row show the scores for the effectiveness characteristics. Again, the components of accountability are shown separately and the numbers below the first row in columns 2-7 are the weights placed on the effectiveness characteristics for the calculation of effective scores. Also, as before, the calculation involves the product of the arrangement score and the geometric mean of the effectiveness characteristics. The only new feature is the calculation of an overall index for allocative efficiency reported in the first row of the penultimate column in Table 2. Because all the four arrangements considered here constitute the country's ability to reach allocative efficiency, the overall measure is simply an average of the four disaggregated effective scores. However, for the purpose of obtaining a summary score for allocative efficiency, it is also possible to use different weights for the effective scores of different arrangements. The last column of Table 2 shows the weights that we have used in our calculations, which are all equal. The example presented in Table 2 suggests that since 1994 New Zealand has created a great deal of potential for reaching allocative efficiency through the participation of line ministries and the use of objective criteria in the prioritization of fiscal goals. Operational Efficiency Questions To gauge the extent of merit-based recruitment and promotion, Question 35 documents the civil service rules and arrangements for hiring and career building. This question is partly based on the work of Poterba et al (1997) and in a similar fashion assesses the "Weberian" characteristics of the bureaucracy. Competitiveness of public sector salaries is a factor that Rauch and Evans include as part of the institutional arrangement, but here it is applied as an effectiveness measure (see Question 39). In addition, we use the measure of educational attainment of the population as another effectiveness measure. This variable highlights the quality of the pool from which the bureaucracy is recruiting. Finally, indexes of predictability of resource flow and availability of support resources for line agencies are applied as other effectiveness measures. These indexes capture the difficulties of maintaining staff incentives when the environment is unpredictable or there is a shortage of support equipment and funds necessary for performing agency tasks. These indexes are calculated based on Questions 38 and 40 respectively. The extent of line agency autonomy is assessed by Question 36. Part A asks about the extent of personnel resources that line agency managers can control. Parts B, C, and D document the extent of control over investment and other funds. Parts E, F, and G examine autonomy from a political perspective and document the extent to which political appointments matter and interfere with the careers of line agency managers. Predictability and availability of resources, average tenure of line agency managers and their accountability are important effectiveness characteristics for this arrangement. The average tenure is recorded by Question 37 and accountability is calculated based on five components. As before, information availability (Question 44) and explicit sanctions (Question 43) at the agency level, as well as the freedom of the press and democracy score, matters for accountability. However, for operational efficiency, the government can set up performance measurement exercises and carry out performance audits in addition to financial audits. The existence and role of such exercises is examined by Question 42. The financial audit score is already measured by Question 17. Finally, procurements of supplies for agencies can be a major source of malfeasance in budget implementation. For this reason, we have devised Question 41 to specifically rate the extent of competition induced by government procurement rules. The overall effectiveness calculations for operational efficiency can be organized in a table such as Table 3. As in the previous tables, the scores for the institutional arrangements and their relevant effectiveness characteristics are combined through a weighted geometric mean calculation. As in the case of allocative efficiency, one can aggregate the two effective scores to generate an overall 148 operational efficiency score. In the example of Table 3, the overall score is a simple average of agency autonomy and merit-based recruiting/promotion arrangements. Again, the last column can be used to place weights on such aggregation. The assessment presented in Table 3 indicates relatively strong institutions for operational efficiency in New Zealand. 149 Table 1: Aggregate Fiscal Discipline Indicators of Budgetary Institutions for New Zealand (1994-1998) Effectiveness Characteristics Accountability Education Level of Decision Average Tenure of Central Average Education Score Hard Budget Constraints Information Availability Comprehensiveness of Ex-post Reconciliation Openness of Financial Budget Agency Heads Freedom of the Press Autonomous Budget Arrangement Score Ex-post Restraints for the Population Explicit Sanctions Democracy Score Effective Score Accountability Determinants of Segments Markets Aggregate Fiscal Makers Budget Audit Discipline Institutional Arrangements 0.94 0.94 0.74 0.98 1.00 1.00 0.95 0.89 1.00 1.00 1.00 1.00 1.00 0.91 0.80 Dominance of the Central Budget Agencies (CBAs) 0.32 0.17 0.17 0.17 0.17 0.17 0.17 0.29 Proposal Stage 0.64 0.17 0.17 0.17 0.17 0.17 0.17 0.59 Legislative Stage 0.49 0.17 0.17 0.17 0.17 0.17 0.17 0.45 Ex-Ante Agreements 0.77 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.71 Macroeconomic Program 1.00 0.17 0.17 0.17 0.17 0.17 0.17 0.92 Institutional Limits on Borrowing/ Spending 0.00 0.20 0.20 0.20 0.20 0.20 0.00 150 Table 2: Allocative Efficiency Indicators of Budgetary Institutions for New Zealand (1995-1998) Effectiveness Characteristics Accountability Average Education Score for Makers at the Line Ministry Education Level of Decision Weight in Overall Score Information Availability Comprehensiveness of Ex-post Reconciliation Freedom of the Press Arrangement Score Autonomous Budget Program Evaluation Explicit Sanctions Democracy Score Effective Score Unity of Budget the Population Accountability Segments Budget Determinants of Level Allocative Efficiency Institutional Arrangements 0.94 0.94 1.00 1.00 1.00 0.93 0.67 1.00 1.00 1.00 1.00 0.91 0.79 Prioritization Through Objective Criteria 1.00 0.17 0.17 0.17 0.17 0.17 0.17 0.97 0.25 Participation of Line Ministries in Planning 1.00 0.25 0.25 0.25 0.25 0.97 0.25 Flexibility of Line Ministries 0.67 0.33 0.33 0.33 0.65 0.25 Breadth of Consultations 0.57 0.17 0.17 0.17 0.17 0.17 0.17 0.55 0.25 151 Table 3: Operational Efficiency Indicators of Budgetary Institutions for New Zealand (1995-1998) Effectiveness Characteristics Accountability Average Education Score for Performance Measurement Competitiveness of Salaries Predictability of Resource Weight in Overall Score Information Availability Average Tenure of Line Freedom of the Press Resource Availability Arrangement Score Agency Managers Explicit Sanctions Democracy Score Competitiveness Effective Score the Population Accountability Procurement Determinants of Audit Flow Operational Efficiency Institutional Arrangements 0.91 0.99 1.00 1.00 1.00 0.95 0.89 1.00 0.78 1.00 1.00 1.00 1.00 0.84 Merit-based Recruitment/Promotion 0.82 0.30 0.20 0.40 0.10 0.80 0.50 Autonomy of Line Agencies 0.89 0.30 0.30 0.10 0.30 0.87 0.50 152 APPENDIX FIVE PEFA MATRIX (FROM DFID’S IMPACT ASSESSMENT OF PFM REFORMS IN BANGLADESH, 2007) Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification A. PFM OUT-TURNS: Credibility of the Budget PI-1 Aggregate No score Expenditure data classified by B Aggregate expenditure deviated from the expenditure out- ministry not available: no budget by 6.5 percent in 2002/3, 10.5 turn compared to calculation or comparison percent in 2003/4 and 6.9 percent in original approved possible. 2004/5 (see attached worksheet). This is budget. rated B. PI-2 Composition of No score Ditto B The composition of expenditure deviated expenditure out- from the budget by 3.4 percent in 2002/3, turn compared to 2.6 percent in 2003/4 and 9.7 percent in original approved 2004/5 (see attached worksheet). This is budget. rated B. PI-3 Aggregate revenue No score Ditto C Domestic revenue was 93.2 percent of the out-turn compared budget in 2002/3, 92.2 percent in 2003/4 to original approved and 92.6 percent in 2004/5. This is rated C. budget. PI-4 Stock and B+ Dim (i): Stock of arrears less in B+ Chapter 2. (i) Arrears at June 2006 monitoring of 1992, but believed to be the same include Tk. 358 crore due on power and expenditure percentage range of total water bills (per Monitoring Cell) and 1 payment arrears. expenditure. No change observed month’s salaries (total 1.8 percent of total in pattern of expenditure arrears expenditure) and little change from June over the period. Rated A. 2005. Rated A. Dim (ii): SoE Monitoring Cell Chapter 3. (ii) Reliable annual data providing such data since 1985. on arrears of GoB payments for fuel and Rated B. utilities provided by SoE Monitoring Cell. No age profile. Rated B. B. KEY CROSS-CUTTING ISSUES: Comprehensiveness and Transparency PI-5 Classification of the D Budget formulation and B Budget for FY07 shows administrative and budget. execution is based on GFS-compliant economic classifications 153 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification administrative and economic (albeit with a high proportion of classification using GFS standards unallocated grants), and a standard with exceptions, for example one functional classification close to COFOG. code called ‘contingency’ was Rated B. used for almost all operational expenditures. Not used consistently. No functional classification in the sense of COFOG, nor program classification (CORBEC 4, 8 &11). PI-6 Comprehensiveness D Budget documentation included B Information in budget documentation of information the revised budget for the current comprises: macroeconomic assumptions, included in budget year, but not macroeconomic including GDP growth, inflation and documentation. assumptions (implied in CORBEC exchange rates (MTBF Box 2, 4.20 and 18) nor the other items (interview 4.27); fiscal deficit and its financing per with former Budget Deputy Sec). GFS 1986 (Budget at a Glance); prior year’s One item out of nine is rated D. budget out-turn, but for revenue budget only (Non-development Expenditure Budget); current year’s revised budget (Development and Non-development Budget); summarized budget for revenue and expenditure for the current year, but not the previous year (ditto); impact of some policy changes (Budget in Brief, Explanatory Notes). Debt stock and financial assets are not shown. Total of five out of eight items are rated B. PI-7 Extent of D+ Dim (i): As for 2006. Rated B. B (i): The main extra budgetary funds are unreported Dim (ii): Financial statements of about 150 statutory bodies performing government the CGA do not fully incorporate regulatory, promotional and operations. the receipts and expenditure of developmental functions (Monitoring Cell). project aid, food aid and technical The only ones with significant income assistance. ERD depends mainly other than from CG grants are the 8 (?) on disbursement reports from Education Boards, which receive about Tk. donors but only a few donors 800 crore examination fees a year 154 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification furnish reports on a regular basis. (1.1percent of total expenditure). There is Payments made by donors direct also the PM Relief Fund, for which to suppliers are not reported to expenditure is well below 1 percent of the CGA (CORBEC 22). Rated D. total expenditure (CFAA 3.2 and interviews with former senior MoF officers). There are also hidden subsidies. The level of unreported expenditure is therefore in the range of 1-5 percent. Rated B. (ii): Donor-funded project expenditure is included in budgets (except for some inputs provided in kind by way of grant) and is substantially included in CGA accounts (ERD). However there is significant delay (estimated at 4-6 months) in bringing aid under RPA and DPA to account. Estimates of inclusion within final June accounts range from 50 to 90 percent (FMRP). Rated B. PI-8 Transparency of D+ As in 2006 (CORBEC para. 14). D+ Dim (i): Elected local government bodies intergovernmental (city corporations, municipalities & union M2 fiscal relations. parishad) receive block grants for scorin development and non-development g expenditure according to need as established in budget submissions to Ministry of Local Government. No transparent and rules-based system of allocation. Rated D. (ii): Grants depend on approval of the Ministry of Local Government’s own budget: LG bodies are notified before the start of the FY. Rated C. (iii): Expenditure reports are rarely received from LG bodies. Rated D. 155 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification PI-9 Oversight of C+ Dim (i): Almost all autonomous C+ (i): As for 1992. Contingent liabilities of aggregate fiscal risk bodies submit fiscal reports SoEs are still unknown (PER, 75). Rated C. from other public including audited accounts at (ii): As for 1992. Rated A. sector entities. least annually, but a consolidated risk overview is missing (Monitoring Cell, CORBEC paras. 26/27). Rated C. Dim (ii): LG bodies cannot generate fiscal liabilities for central government (LGD interview, and submit annual receipts and payment accounts (CFAA Chapter 4). Rated A. PI-10 Public access to key C Annual budget documentation B Annual budget documentation, year-end fiscal information. (but up to six months late), year- financial statements (within six months of end financial statements (within audit?) and external audit reports (within six months of audit?) and external six months of audit?) are published, and audit reports (within six months in-year budget execution reports (monthly of audit). Two items out of six fiscal reports on www.mof.gov.bd but six earns a C. months in arrears), and contract awards (www.cptu.gov.bd), but not resources available to primary service units. Four items out of six still earns a B rating. C. BUDGET CYCLE C(i): Policy-Based Budgeting PI-11 Orderliness and D+ Dim (i): Clear budget calendar, B+ (i): Separate budget calendars and participation in the MDAs allowed six weeks to procedures for MTBF MDAs (10 in FY 2007, M2 annual budget prepare detailed estimates, but responsible for 46 percent of primary scorin process. delays often experienced (IMCL: expenditure) and all other MDAs. Both g 4.12.1). Rated B. calendars are clear, are generally adhered Dim (ii): Budget circular does not to, and provide MDAs ample time (>six include ceilings (IMCL 4.13.3), and weeks) to prepare detailed estimates. no cabinet involvement until after Rated A. estimates are prepared. Rated D. (ii): Intended that cabinet approve budget 156 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification Dim (iii): Budget approved late, so ceilings but for FY 2007, Minister of MoF regularly uses vote on Finance approved ministry ceilings before account procedure (IMCL 4.12.2). issue of budget circular (MTBF Rated D. para.2.1).Rated C. (iii): Legislature regularly approves the budget before 1 July. Rated A. PI-12 Multi-year D Dim (i): No forward estimates of C+ (i): Three-year MTBF since FY06. MTBF for perspective in fiscal fiscal aggregates are undertaken FY 07-09 does not explain differences in M2 planning, (IMCL 4.14.2). Rated D. allocations for FY07 from previous MTBF. scorin expenditure policy Dim (ii): No debt sustainability Rated C. g and budgeting. analysis in the last three years (ii): Annual DSA for external and domestic (inference from IMCL 4.19). Rated debt (MTBF Chapter Six). Rated A. D. (iii): Ten MDAs, responsible for 46 percent Dim (iii): No sector strategies of primary expenditure in FY07 budget, have substantially complete were included in the MTBF, but fiscal costing of investments and ceilings and sectoral strategies not recurrent expenditure (IMCL properly reflected in some cases (para. 4.18.4). Rated D. 2.5). No sector strategies costed in the Dim (iv): Budgeting for PRSP, but costs estimated for some investment and recurrent programs are established (FRA 18). Rated expenditure are separate C. processes with no recurrent cost (iv): Single ceiling issued to each MTBF estimates shared (CORBEC 54:16, MDA covering development and recurrent IMCL 4.16). Rated D. costs with effect from FY 2006 (MTBF). Many investment decisions have weak links to sector strategies (WB review, 4.4). Rated C. C(ii): Predictability and Control in Budget Execution PI-13 Transparency of C As in 2006 C Substantial discretionary powers (C), taxpayer obligations information on tax liabilities not and liabilities. comprehensive (C), and tax appeal system in place but needs redesign (C). PI-14 Effectiveness of D Registration system ineffective C Taxpayer registration system not linked to measures for and penalties not usually other systems (C), compulsory penalties 157 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification taxpayer enforced. for nonsubmission of tax returns since registration and tax 2004 (B), and tax audit is ad hoc (D). assessment. PI-15 Effectiveness in D As in 2006. D Income tax debt collection is below 60 collection of tax percent (D), tax collections often payments. transferred less regularly than monthly (D), and there is time lag in reconciliations (D). PI-16 Predictability in the D As in 2006. D (i): Cash management responsibility is availability of funds divided and unsystematically for commitment of implemented. Quarterly revenue expenditures. projections may be made, but no regular expenditure projections during the year (World Bank Review, June 2005, 4.16-18. Rated D. (ii): Controlling and disbursement officers of spending units get detailed allocations from their administrative ministries but these are not reliable as spending authorities (ditto). Rated D. (iii): There are frequent and nontransparent re-appropriations of major head budgets by Finance Division, Planning Commission and administrative ministries (FRA, Table 1). Rated D. PI-17 Recording and C As in 2006. C (i): External debt data are complete, management of updated and reconciled with donors cash balances, debt annually by ERD (one year lag). Domestic and guarantees. debt data are maintained by BB and National Savings Directorate (six-week lag). A new unit established in FD 6/05 to integrate external and domestic debt data is not yet fully operational. Rated C. (ii): Cash balances consolidated daily (single BB account with 100 subaccounts), 158 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification but funds held by Sonali Bank remain outside. Rated B. (iii): ERD approves all loans to GoB with concurrence of FD and PC, and FD Budget Wing approves all issues of GoB guarantees. No clear guidelines, criteria or overall ceilings. Rated D. PI-18 Effectiveness of C+ No change since 1992 except C+ (i): Payroll for Defence and single payroll payroll controls. partial computerization. for all other public servants. No regular reconciliation with personnel records, but changes fully documented and CAOs reconcile with previous month’s payroll, although this is often delayed. Rated B. (ii): Paybills received by CGA up to 20th of month are paid by 1st but administrative delays occur before receipt by CGA. Rated B. (iii): Most public servants paid in cash by DDOs rather than by cheque. Payrolls are mainly manually prepared, and controlled by collective bargaining agents. No independent preparation of payroll bank reconciliations, nor follow up of uncashed cheques (CFAA Chapter Three). Rated C. (iv): Payroll audit is a part of external audit, but this does not include physical identification of payees. No internal audit. Rated D. PI-19 Competition, value No score. As for 2006 except that there was No score. (i): Under Public Procurement Act (7/06), for money and no central technical unit, no all contracts over specified thresholds (for controls in standardization of regulations example goods over Tk. 10 lakhs) should procurement. and bidding documents, general be advertised and awarded through open use of enlisted contractors rather tendering, with alternative methods used than open competition (former only in clearly defined circumstances. senior officer in procurement, & CPTU does not receive information on all 159 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification CPAR 11/99). Insufficient such contracts: only those over Tk. 1 crore information to score. have to be notified and advertised on the CPTU website. Procurement agencies are required to keep records and submit performance indicators annually to CPTU. So far, only LGED and REB do so. Their returns show 100 percent open competition for contracts > Tk. 1 crore (CPTU interviews). Insufficient data to score. (ii): Regulations require justification for use of alternative methods of procurement, but there is insufficient information to rate higher than C. (iii): Independent Review Panels set up to hear complaints that are not settled up to secretary level. Out of 10 complaints to date, four have been awarded in favour of the plaintiff, four against, and two are in process (CPTU information). Rated A. PI-20 Effectiveness of D+ Dim (i): Commitments entered D+ (i): Commitment control abandoned. DAOs internal controls for into vote control registers by and UAOs can commit GoB to spending for nonsalary MDAs, partially effective. Rated C. which there is no budget and pay from expenditure. Dim (ii): Comprehensive rules but following year’s budget (WB review 4.17 scattered through 58 manuals; and FRA 4.24). Rated D. little understood (CORBEC para. (ii): Internal Control Manual and Public 54 (21) & Annexure XXIV). Rated Expenditure Management Manual issued D. by FD, but these do not replace the old Dim (iii): Inadequate enforcement manuals and training only just started. of rules (CORBEC para.30) Rated Rated C. D. (iii): No change since 1992. Rated D. PI-21 Effectiveness of D As in 2006. D No internal audit function established yet; internal audit. it remains a PAO general responsibility. Funds allocated for contracting out internal audit are not yet used. Rated D on 160 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification all three dimensions. C(iii): Accounting, Recording and Reporting PI-22 Timeliness and D Same as in 2006 (CORBEC para. D (i): Monthly bank reconciliation, especially regularity of 36, IMCL 7.3.11). at ministry level, is not yet working accounts effectively, due mainly to the number of reconciliation. coding errors by GoB and BB. Software installed in 64 DAOs by 6/06. The latest reconciliation (in total) is three months in arrears. Rated D. (ii): Problems with clearing temporary advances, personal ledger deposit accounts, suspense and remittance accounts before the year-end final accounts (WB review 4.28). Ninety-five percent are cleared (CGA interview) but with more than two months delay. Some MDAs charge unauthorized expenditures to suspense accounts pending authorization in a later year against a new budget. Suspense accounts are also used to hold up and divert funds released towards the end of the financial year (CFAA). Rated D. PI-23 Availability of D No surveys undertaken in C A survey of primary health care and family information on previous three years. planning was published in 9/05 (FMRP). resources received by service delivery units. PI-24 Quality and D (i): As in 2006. Rated D. D+ (i): Monthly Reports on Fiscal Position timeliness of in-year (ii): Reports issued 10-14 weeks show actual expenditure against budget, budget reports. after month end (IMCL 3.6.2 & but not commitments. Rated D. 7.3.9, Gardener 2.2.1). Rated D. (ii): CGA expenditure reports usually (iii): Exchange accounts (book available five weeks after the end of each adjustments between MDAs) left month and NBR revenue reports within 161 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification unrecorded as expenditure, one month (WB review 3.1 & 4.23). Rated misuse of suspense accounts and B. personal ledger deposit accounts, (iii): There are concerns about the quality absence of timely bank of information, due principally to the lack reconciliation and correction of of bank reconciliation at ministry level, and classification errors and incomplete ADP spending (WB review 3.1 misappropriations, omission of & 4.26-28). Rated C. donor direct disbursements to suppliers. Rated D. PI-25 Quality and D+ Dim (i): As in 2006 (IMCL 4.21.8, D+ (i): Revenue, Appropriation and Finance timeliness of annual 6.1.2, 7.3.9). Rated B. Accounts are prepared annually. They financial Dim (ii): Latest statements were provide full information on revenue and statements. for 1987/8 (five years in arrears). expenditure, but not on assets and Rated D. liabilities, for example food stocks. Rated Dim (iii): As in 2006. Rated D. B. (ii): Annual statements are not available for audit within six months (Gap analysis, six). Rated B. (iii): Statements do not conform to IPSAS cash-based standard, and do not disclose accounting policies (Gap analysis, six). Rated D. C(iv): External Scrutiny and Audit PI-26 Scope, nature and D Dim (i): As for 2006. D (i): C&AG audits cover < 50% of central follow-up of Dim (ii): As for 2006. government entities each year, and do not external audit. Dim (iii): Prompt actions not highlight systemic issues (CFAA Chapter taken and breaches, often of a Six, FMRP, TI report). Some performance serious nature, may be condoned audits undertaken. Government Audit or ignored (CORBEC, 30). Rated D. Standards, based on INTOSAI standards, adopted but not yet met. Rated D. (ii): Audit reports are issued more than 12 months after receipt of financial statements for audit (CAG presented audited accounts for FY 2005 (Railways 162 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification and Defence only) and FY 2004 (other MDAs) to President 19 October 2006. Rated D. (iii): As for 1992. Little evidence of response (FMRP). Rated D. PI-27 Legislative scrutiny D+ As for 2006. D+ (i): Parliament reviews the budget of the annual (revenue and expenditure) only after it is budget law. finalized. Rated C. (ii): Parliamentary procedures are well established and respected. Rated B. (iii): Time allowed for debate is less than one month (WB review 4.14). Rated D. (iv): Rules for in-year amendments to the budget without prior legislative approval exist, but are not always respected. Supplementary estimates prepared to give ex-post authorization. Rated C. PI-28 Legislative scrutiny D+ As for 2006. D+ (i): Over last three years’ audit reports, of external audit PAC usually takes more than 12 months to reports. complete its examination. Recent improvement in timeliness. Rated D. (ii): In-depth hearings take place with PAOs (Secretaries) of audited bodies facing audit objections. Rated B. Actions are recommended by the PAC but are rarely acted on. Even initial actions are delayed for years (TI report). Rated C. D. DONOR PRACTICES D-1 Predictability of Not scored. D+ (i): For FY 2005 and 2006, $300 million Direct Budget direct budget support was forecast by IDA Support. well before finalization of the budget, and this was fully received. Rated A. (NB. FY 2004 data not available). (ii): In FY 2005 and 2006, the DBS was received an average of six months late 163 Chapter 2: Chapter 4: Chapter 1. Score 1992 Chapter 3: Justification Score 2006 Chapter 5: Justification (expected in July, but 200 received in December and 100 in March or early April, due to difficulty with conditionality). This is rated D. (FY 2004 data not available). D-2 Financial D+ (i): Same as 2006. Rated B. B (i): Most donors provide projections on information (ii): Only a few donors provide their project disbursements at a time and provided by donors disbursement reports, and only format suitable for GoB budget for budgeting and on request and with delays requirements (ERD-FABA). Rated B. reporting on project (CORBEC 54 (9), CFAA Chapter (ii): Main donors, responsible for over 70 and program aid. Six). Rated D. percent of all project aid, provide monthly disbursement reports within one week, classified by project in accordance with GoB budget (ERD-FABA). Rated B. D-3 Proportion of aid D Same as in 2006. D It is estimated that 39 percent of foreign that is managed by aid (budget FY2007) is managed by use of national national procedures. This includes all procedures. direct budget support (IDA), commodity aid and food aid counterpart funds, and 25 percent of project support (due to C&AG- FAPD audit of foreign-aided projects). Rated D. 164 REFERENCES Alesina, A., and R. Perotti. 1996. Fiscal Discipline and the Budget Process. American Economic Review, 86 (2): pp. 401-407. Alesina, A., R. Hausmann, R. Hommes and E. Stein. 1999. Budget Institutions and Fiscal Performance in Latin America. Journal of Development Economics, 59 (2): pp. 253-273. Anwar, I., M. Sami. N. Akhtar, M.E. Chowdhury, U. Salma, M. Rahman and M. Koblinsky. 2008. Inequity in maternal health-care services: evidence from home-based skilled-birth-attendant programmes in Bangladesh. Bulletin of the World Health Organization, 86 (4): pp. 252-259. Asiedu, E., and H.S. Esfahani. 2001. Ownership Structure in Foreign Direct Investment Projects. Review of Economics and Statistics, 83(4): pp.647-662. Barro, R.J. 1979. On the Determination of Public Debt. Journal of Political Economy, 87: pp. 940-71. Barro, R.J. and J-W. Lee. 1996. International Measures of Schooling Years and Schooling Quality. American Economic Review, 86 (2): pp. 218-223. Campos, E., and S. Pradhan. 1996. Budgetary Institutions and Expenditure Outcomes: Binding Governments to Fiscal Performance. Policy Research Working Paper No. 1646. Washington DC: World Bank. Chatterton, I., and O.S. Puerto. 2006. Estimation of Infrastructure Investment Needs in the South Asia Region. Washington DC: World Bank. Chaudhury, N., and J.S. Hammer. 2003. Ghost Doctors: Absenteeism in Bangladeshi Health Facilities. Policy Research Working Paper No. 3065. Washington DC: World Bank. Computer Solutions Unlimited. 1998. Cross-national Time Series Data Archive. Computer file. Binghamton, NY. Deininger, K., and L. Squire. 1996. “A new data set measuring income inequality. World Bank Economic Review, 10 (3): pp. 565-591. de Oliveira, J.C., L. Tacca Jr., and P. Fontenele e Silva. 1995. Aspectos Relevantes do Orçamento da União. Revista ABOP, 37: pp. 89-131. Easterly, W., and R. Levine.1997. Africa's Growth Tragedy: Policies and Ethnic Divisions. Quarterly Journal of Economics, 112.4: pp. 1203-1250. Easterly, W., and K. Schmidt-Hebbel. 1993. Fiscal Deficits and Macroeconomic Performance in Developing Countries. World Bank Research Observer, 8 (2): pp. 211-237. Eichengreen, B., and T. Bayoumi. 1994. The Political Economy of Fiscal Restrictions: Implications for Europe from the United States. European Economic Review, 38(3-4): pp. 783-91. Esfahani, H.S. 1999a. Institutions, Budgetary Procedures, and Fiscal Performance: A New Institutional Framework for the Analysis of Budgeting Procedures. In Institutional Reform and Economic Development in the Middle East and North Africa, ed. I. Limam. Arab Planning Institute: Kuwait, pp. 149-180. Esfahani, H.S. 1999b. Budget Institutions and Fiscal Sustainability in Brazil.” Manuscript, Latin America and the Caribbean Region Department. Washington DC: World Bank. 165 Esfahani, H.S. 2000. Institutions and Government Controls. Journal of Development Economics, 63 (2): pp. 197-229. Fan, S., and N. Rao. 2003. Public Spending in Developing Countries: Trends, Determination, and Impact. EPTD Discussion Paper 99. Washington DC: International Food Policy Research Institute (IFPRI). Glinskaya, E. 2005. Education and Health Expenditures in Bangladesh: Benefit Incidence Analysis. Journal of Developing Societies, 21 (1-2): pp. 91-120. Grilli, V., D. Masciandaro and G. Tabellini. 1991. Political and Monetary Institutions and Public Financial Policies in the Industrial Countries. Economic Policy, vol 6: pp. 341-392. Gwatkin, D.R., S. Rutstein, K. Johnson, E. Suliman, A. Wagstaff, and A. Amouzou. 2007. Socio-economic Differences in Health, Nutrition and Population. Washington DC: World Bank. Hallerberg, M., and J. von Hagen. 1997. Electoral Institutions, Cabinet Negotiations, and Budget Deficits in the European Union. In Fiscal Institutions and Fiscal Performance, eds. J. Poterba and J. von Hagen. Chicago: University of Chicago Press, pp. 209-232. Humana, C. 1992. World Human Rights Guide, Third Edition. New York: Oxford University Press. IMF-World Bank. 2008. Policy Note on Debt Sustainability Analysis. Inter-American Development Bank. 1998. Economic and Social Progress Report 1998: Latin America After a Decade of Reforms. Washington DC: IDB. Jaggers, K., and T.R. Gurr. 1996. Polity III: Regime Type and Political Authority, 1800-1994. Inter- university Consortium for Political and Social Research, Ann Arbor MI. Krehbiel, K. 1991. Information and Legislative Organization. University of Michigan Press: Ann Arbor. Levine, R., and S. Zervos. 1998. Stock Markets, Banks, and Economic Growth. American Economic Review, 88 (3): pp. 537-58. Mercer, M.M., T.M. Murshid, M.Z. Kabir and A.R. Khan. 2005. Analysis of Madrasah in Bangladesh and Pakistan. SchlumbergerSema: Belgium. Önis, Z., and S.B. Webb, 1994. Turkey: Democratization and Adjustment from Above. In Voting For Reform: Democracy, Political Liberalization and Economic Adjustment, eds. S. Haggard and S.B. Webb. World Bank: Washington DC, pp. 128-184. Persson, T., and G. Tabellini. 2002. Political Economics and Public Finance. In Handbook of Public Economics: Volume 3, eds. A. Auerbach and M. Feldstein. North-Holland: Amsterdam. Political Risk Services. 2000. International Country Risk Guide Data, 1982-95. Computer file, Syracuse NY. Compiled by IRIS Center, University of Maryland: College Park MD. Poterba, J.M. 1994. State Responses to Fiscal Crises: The Effects of Budgetary Institutions and Politics. Journal of Political Economy, 102.4: pp. 799-821. Poterba, J.M. 1995. Capital Budgets, Borrowing Rules, and State Capital Spending. Journal of Public Economics, 56: pp. 165-187. Poterba, J.M., J. von Hagen, J.E. Rauch, and P.B. Evans. 1997. Bureaucratic Structure and Bureaucratic Performance in Less Developed Countries. Manuscript. The World Bank Policy Research Department. 166 Roubini, N., and J.D. Sachs. 1989. Political and Economic Determinants of Budget Deficits in the Industrial Democracies. European Economic Review, 33: pp. 903-938. Schick, A. 1995. The Federal Budget. The Brookings Institution: Washington DC. Scott, G. 1996. Government Reform in New Zealand. Washington DC: International Monetary Fund. Shepsle, K. 1979. Institutional Arrangement and Equilibrium in Multidimensional Voting Models. American Journal of Political Science, 23: pp. 27-60. Shepsle, K., and B.R. Weingast. 1994. Positive Theories of Congressional Institutions. Legislative Studies Quarterly, 19(2): pp. 149-179. Stein, E., E. Talvi, and A. Grisanti. 1998. Institutional Arrangements and Fiscal Performance: The Latin American Experience. NBER Working Paper No. 6358. Cambridge MA. Tanzi, V. and L. Schuknecht. 2000. Public Spending in the 20th Century: A Global Perspective. Cambridge: Cambridge University Press. von Hagen, J. 1992. Budgeting Procedures and Fiscal Performance in the European Communities. Economic Papers No. 96, Commission of the European Communities, Directorate General for Economic and Financial Affairs: Brussels. von Hagen, J., and I.J. Harden. 1994. National Budget Processes and Fiscal Performance. European Economy: Reports and Studies, 3: pp. 311-418. von Hagen, J., and I.J. Harden. 1995. Budget Processes and Commitment to Fiscal Discipline. European Economic Review, 39(3-4): pp. 771-779. Weingast, B.R., Shepsle, K. and C. Johnsen. 1981. The Political Economy of Benefits and Costs: A Neoclassical Approach to Distributive Politics. Journal of Political Economy, 89: pp. 642-64. Wilson, J.Q. 1989. Bureaucracy: What Government Agencies Do and Why They Do It. Basic Books: New York. World Bank. 2005. Growth and Competitiveness Report. World Bank. World Trade Indicators 2007. Zafarullah, H. 1998. National Administration in Bangladesh: An Analysis of Organisational Arrangements and Operating Methods. Asian Journal of Public Administration, 20 (1): pp. 79-112. Zafarullah, H., and A.S. Huque. 2001. Public Management for Good Governance: Reforms, Regimes and Reality in Bangladesh. International Journal of Public Administration, 24 (12): pp. 1379-1404. 167 Statistical Annex Table 1: Bangladesh Macroeconomic Indicators Description FY02 FY03 FY04 FY05 FY06 FY07 FY08 Growth Rates (%) GDP Growth 4.4 5.3 6.3 6.0 6.6 6.4 6.2 GDP Growth Per Capita 2.9 3.8 4.8 4.5 5.1 4.9 4.7 Per Capita GDP Atlas Method (US$ ) 372.0 389.0 418.3 447.7 469.0 493.7 532.1 Per Capita GDP Growth Atlas Method (US$ ) 0.2 4.6 7.5 7.0 4.8 5.3 7.8 Saving & Investment (% of GDP) Gross Domestic Saving 18.2 18.6 19.5 20.0 20.2 20.4 20.1 Gross National Saving 23.4 24.9 25.4 25.8 27.7 28.7 29.2 Private Investment 16.8 17.2 17.8 18.3 18.7 19.0 19.2 Public Investment 6.4 6.2 6.2 6.2 6.0 5.4 5.0 Central Govt. Budget (% of GDP) Total Revenue 10.1 10.3 10.2 10.5 10.7 10.4 11.4 Total Expenditure 14.7 13.7 13.1 14.0 14.1 13.5 15.1 Overall Budget Deficit 4.6 3.4 2.9 3.5 3.4 3.1 3.7 Balance of Payments (% of GDP) Trade 28.6 29.3 30.7 33.9 37.9 40.3 42.3 Exports(f.o.b.) 12.4 12.5 13.3 14.2 16.6 17.6 17.7 Imports(f.o.b.) 16.1 16.8 17.4 19.7 21.3 22.7 24.7 Services & Income (net) -1.9 -2.0 -2.2 -2.6 -2.8 -3.2 -3.2 Current Transfers 5.9 6.6 6.6 7.1 8.7 9.6 11.1 Current Account Balance( including transfers) 0.3 0.3 0.3 -0.9 1.3 1.4 0.9 External Indicators External Debt (US$ b.) 16.1 16.9 17.5 17.8 17.9 19.6 20.5 Ext. Debt as % of GDP 33.7 32.6 31.0 29.6 28.9 28.7 25.9 BB Gross Reserves (US$ b.) (end of period) 1.6 2.5 2.7 2.9 3.5 5.1 6.1 BB Gross Reserves (in months of imports) 2.1 2.9 2.8 2.5 2.7 3.3 3.4 External Debt Service Ratio (% of total foreign exchange earnings)* 6.3 5.8 4.8 4.5 4.1 3.7 3.4 Money and Credit M2 Growth (%, year-on-year) 13.1 15.6 13.8 16.7 19.5 17.0 17.6 Net Domestic Asset Growth (%, year-on-year) 11.6 12.2 13.5 17.1 19.7 12.5 18.2 Ratio of Private Sector Credit to GDP (%) 24.7 25.8 28.4 29.9 31.5 31.9 34.8 Exchange Rate Nominal Period Average (TK/US$) 57.4 57.9 58.9 61.5 67.2 69.1 68.6 Nominal End of Period (TK/US$) 57.9 58.5 60.4 63.7 69.7 68.8 68.5 Real Effective Exchange Rate Index 101.5 97.0 93.4 91.7 86.9 89.7 91.0 Rate of Inflation (%) (year on year)** 2.8 4.4 5.8 6.5 7.2 7.2 9.9 Total Public Debt (% of GDP) 52.9 51.1 49.1 47.5 46.9 46.5 44.4 Memorandum Items GDP at Current. Prices (Taka bill.) 2737.3 3005.8 3329.7 3707.1 4157.3 4724.8 5419.6 GDP at Current. Prices Atlas Method (US$ bill) 51.1 53.0 56.9 61.7 65.8 70.3 76.9 GDP at Current. Prices (US$ bill) 47.7 51.9 56.6 60.3 61.9 68.4 79.0 Population (mill.) 132.1 134.1 136.1 138.2 140.2 142.4 144.5 Population growth Rate 1.5 1.5 1.5 1.5 1.5 1.5 1.5 Source: Various publications of the World Bank, ADB and Bangladesh Bureau of Statistics 1/ balance of payment (% of GDP) refers to IMF's 2008 article IV consultation * Total foreign exchange earnings include commodity earnings, workers' remittances and invisible receipts. ** CPI was rebased from FY98 using FY96 weights The Atlas Method was not used for calculating average per capita GDP & GDP at current market prices in the 1980s, in FY04 and FY05 168 Table 2: Summary Macroeconomic Indicators Actual Provisional Projections FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 Output and Prices annual percent change Real GDP Growth 4.4 5.3 6.3 6.0 6.6 6.4 6.2 5.9 5.0 CPI 2.8 4.4 5.8 6.5 7.2 7.2 9.9 7.0 6.5 External Outlook Exports (f.o.b.) (billion $) 5.9 6.5 7.5 8.6 10.4 12.1 13.9 16.0 17.9 Annual % change -7.6 9.5 15.9 14.0 21.5 15.8 15.7 13.4 12.0 Imports (c.i.f.) (billion $) 7.7 8.7 9.8 11.9 13.3 15.5 19.5 22.2 25.3 Annual % change -8.7 13.1 13.0 20.6 12.1 16.6 25.6 14.1 14.0 Current account balance (% of GDP) 0.3 0.3 0.3 -0.9 1.3 1.4 0.9 0.8 -0.3 Gross official reserves (billion $) 1.6 2.5 2.7 2.9 3.5 5.1 6.1 6.1 5.8 In months of GNFS imports 2.1 2.9 2.8 2.5 2.7 3.3 3.4 2.9 2.8 Public Finance percent of GDP Total Revenue 10.1 10.3 10.2 10.5 10.7 10.4 11.4 10.9 11.0 Total Expenditures 14.7 13.7 13.1 14.0 14.1 13.5 15.1 14.4 16.1 Overall budget deficit (excluding grants) 4.6 3.4 2.9 3.5 3.4 3.1 3.7 3.5 5.1 Domestic financing 2.5 1.2 1.8 1.7 2.2 1.9 1.5 1.8 2.2 Public debt 52.9 51.1 49.1 47.5 46.9 46.5 44.4 43.6 42.8 Money and Credit End of year; percent change Net Domestic assets 11.6 12.2 13.5 17.1 19.7 12.5 18.2 18.1 17.5 Private sector 13.9 12.6 17.5 17.0 18.3 15.1 25.2 18.5 14.8 Broad money (M2) 13.1 15.6 13.8 16.7 19.5 17.0 17.6 17.5 17.0 Source: IMF & GOB 169 Table 3: Development and Revenue Expenditure by Ministry/Division (% of GDP) Ministry/Division FY02 FY03 FY04 FY05 FY06 FY07 FY08 a/ FY09 b/ Revised Revised Budget General Administration 3.3 3.3 3.1 3.3 3.1 3.4 3.7 4.4 General Public Services (GPS) 1.3 1.5 1.3 1.3 1.3 1.3 1.7 2.4 Defence 1.3 1.2 1.2 1.2 1.1 1.1 1.1 1.0 Public Order and Safety (POS) 0.7 0.7 0.7 0.8 0.7 0.9 0.9 0.9 Social Services 3.6 3.5 3.5 3.4 3.4 4.1 4.1 4.4 Education 2.1 2.1 1.9 1.8 2.1 2.3 2.1 2.0 Health 0.9 0.8 0.8 0.8 0.7 1.0 1.0 1.0 Social Security and Welfare (SSW) 0.5 0.5 0.6 0.7 0.5 0.6 0.9 1.3 Recreation, Culture and Religious Affairs (RCRA) 0.1 0.2 0.2 0.2 0.2 0.2 0.2 0.1 Economic Services 2.3 2.0 2.3 2.7 2.4 2.8 3.0 2.8 Agriculture, Fisheries and Livestock (AFL) 0.9 0.8 0.8 1.1 1.0 1.1 1.6 1.5 Mining, Manufacturing and Construction (MMC) 0.1 0.1 0.1 0.1 0.1 0.0 0.1 0.1 Rural Development & Cooperatives Division (RDC) 1/ 1.2 1.1 1.3 1.4 1.2 1.5 1.2 1.1 Ministry of Chittagong Hill Tracts Affairs (CHTs) 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 Ministry of Commerce, Labour & Employment (CLE) 0.0 0.0 0.0 0.0 0.1 0.1 0.1 0.0 Infrastructure Services 2.9 3.1 2.9 3.0 2.3 2.1 2.0 1.9 Ministry of Science & Technology (ST) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Fuel and Energy (FE) 0.7 0.9 1.1 1.1 0.9 0.6 0.7 0.7 Transport and Communication (TC) 2.0 2.0 1.5 1.8 1.2 1.3 1.1 1.0 Housing and Community Services (HCS) 0.1 0.1 0.2 0.2 0.1 0.1 0.1 0.2 Interest 1.7 1.9 1.7 1.8 1.9 1.9 2.2 2.0 Total 13.8 13.8 13.6 14.3 13.2 14.3 15.0 15.5 Source: Ministry Of Finance 1/ Local Government Division and Rural Development and Co-operatives are added together instead of putting Local Govt. Division under the Housing and Community Services 2/ In FY04-05, Ministry of Jute and Ministry of Textile, Ministry of Food and Ministry of Disaster Management and Relief have been merged and respectively renamed as Ministry of Textile and Jute and Ministry of Food and Disaster Management. a/ FY08(revised) Total does not include : Net Outlay for Food Account Operation -Tk. 8.09 bn, Loans and Advances Net -Tk. 18.01 bn, Structural Adjustment Expenditure -Tk. 17 bn, Non- ADP Employment Generation Programmes-Tk 5.49 bn & Liability of b/ FY09(budget) Total does not include : Net Outlay for Food Account Operation -Tk. 6.89 bn, Loans and Advances Net -Tk. 19.72 bn, Structural Adjustment Expenditure -Tk. 10 bn, Non- ADP Employment Generation Programmes-Tk 10.05 bn. Discrepancy may arise due to rounding up. Note:difference in total figure may arise due to rounding up. 170 Table 4: Bangladesh Central Government Operations Provisional FY02 FY03 FY04 FY05 FY06 2 FY07 FY08 (in billion Taka) Total Revenue 276.8 309.8 339.0 389.2 443.7 490.6 617.7 Tax 210.3 248.3 274.3 314.1 354.2 389.9 492.4 Non-Tax 66.5 61.5 64.7 75.1 89.5 100.7 125.3 Total Expenditure -403.5 -411.5 -435.7 -519.3 -585.3 -635.9 -819.0 Current expenditure -218.1 -244.5 -258.7 -312.5 -350.5 -424.7 -520.5 Other expenditures1 -2.1 -9.6 -11.3 -2.5 1.2 -3.1 -25.1 Annual Development Program -152.3 -163.0 -167.9 -185.8 -194.7 -180.2 -185.2 Non-ADP & Net Lending -17.2 -6.6 -16.0 -24.2 -41.3 -27.9 -88.2 Residual -13.8 12.2 18.2 5.7 0.0 0.0 0.0 Overall budget deficit * 126.7 101.7 96.7 130.0 141.6 145.3 201.3 Net foreign financing 58.1 64.2 36.8 66.3 49.3 55.9 120.0 Net domestic financing 68.1 37.5 59.9 63.7 92.3 89.4 81.3 Banking system 21.7 -10.7 13.5 34.9 64.7 54.7 61.3 Other domestic 46.4 48.2 46.4 28.8 27.6 34.7 20.0 Privatization Receipts 0.6 0.0 0.0 0.0 0.0 0.0 0.0 Total Central Government Debt 1449 1536 1635 1761 1950 2197 2406 GDP at current market prices 2737 3006 3330 3707 4157 4725 5420 (percent of GDP) Total Revenue 10.1 10.3 10.2 10.5 10.7 10.4 11.4 Tax 7.7 8.3 8.2 8.5 8.5 8.3 9.1 Non-Tax 2.4 2.0 1.9 2.0 2.2 2.1 2.3 Total Expenditure 14.7 13.7 13.1 14.0 14.1 13.5 15.1 Current expenditure 8.0 8.1 7.8 8.4 8.4 9.0 9.6 Other expenditures1 0.1 0.3 0.3 0.1 0.0 0.1 0.5 Annual Development Program 5.6 5.4 5.0 5.0 4.7 3.8 3.4 Non-ADP & Net Lending 0.6 0.2 0.5 0.7 1.0 0.6 1.6 Residual -0.5 0.4 0.5 0.2 0.0 0.0 0.0 Overall budget deficit 4.6 3.4 2.9 3.5 3.4 3.1 3.7 Net foreign financing 2.1 2.1 1.1 1.8 1.2 1.2 2.2 Net domestic financing 2.5 1.2 1.8 1.7 2.2 1.9 1.5 Banking system 0.8 -0.4 0.4 0.9 1.6 1.2 1.1 Other domestic 1.7 1.6 1.4 0.8 0.7 0.7 0.4 Privatization Receipts 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total Central Government Debt 52.9 51.1 49.1 47.5 46.9 46.5 44.4 Source: IMF and Staff estimates * The overall budget deficit for 2004/05 includes externally financed flood-related expenditure 1/ include food account surplus(-)/deficit(+) and extraordinary expenditures 2/ Non-ADP & Net Lending in FY06 includes bonds (10 billion taka) issued to a nationalized commercial bank to assume BPC's liabilities. 3/ Non-ADP & Net Lending includes bonds (84 billion taka) issued to three nationalized commercial banks to assume BPC's liabilities. Note: difference in total figure due to rounding up. 171 Table 5: Bangladesh - Revenue Trends (in billion Taka) FY02 FY03 FY04 FY05 FY06 FY07 FY08 (P) FY09 (B) Total revenue 276.8 309.8 339.0 389.2 443.7 490.6 617.7 693.8 Tax revenue 210.3 248.3 274.3 314.1 354.2 389.9 492.4 567.9 NBR tax revenue 202.2 237.6 261.9 299.9 338.9 372.2 472.0 545.0 Income Tax 37.9 43.7 47.1 56.7 71.4 87.2 115.8 130.5 Import based Tax 104.8 120.7 131.7 151.0 152.7 156.6 198.5 225.4 VAT (Domestic) 32.3 36.4 43.2 50.8 64.1 74.7 90.9 106.8 Other (residual) 27.2 36.7 40.0 41.3 50.6 53.7 66.8 82.3 Non-NBR tax revenue 8.1 10.7 12.4 14.2 15.3 17.7 20.4 22.9 Non-tax revenue 66.5 61.5 64.7 75.1 89.5 100.7 125.3 125.9 GDP Mkt. Prices 2,737 3,006 3330 3707 4157 4725 5420 6131 As % of GDP Total revenue 10.1 10.3 10.2 10.5 10.7 10.4 11.4 11.3 Tax revenue 7.7 8.3 8.2 8.5 8.5 8.3 9.1 9.3 NBR tax revenue 7.4 7.9 7.9 8.1 8.2 7.9 8.7 8.9 Income Tax 1.4 1.5 1.4 1.5 1.7 1.8 2.1 2.1 Import based Tax 3.8 4.0 4.0 4.1 3.7 3.3 3.7 3.7 VAT (Domestic) 1.2 1.2 1.3 1.4 1.5 1.6 1.7 1.7 Other (residual) 1.0 1.2 1.2 1.1 1.2 1.1 1.2 1.3 Non-NBR tax revenue 0.3 0.4 0.4 0.4 0.4 0.4 0.4 0.4 Non-tax revenue 2.4 2.0 1.9 2.0 2.2 2.1 2.3 2.1 Growth (%) Total revenue 20.9 11.9 9.4 14.8 14.0 10.6 25.9 12.3 Tax revenue 9.0 18.1 10.5 14.5 12.8 10.1 26.3 15.3 NBR tax revenue 7.5 17.5 10.2 14.5 13.0 9.8 26.8 15.5 Income Tax 7.8 15.4 7.7 20.5 25.9 22.1 32.8 12.7 Import based Tax 4.5 15.1 9.1 14.6 1.1 2.5 26.7 13.6 VAT (Domestic) 20.3 12.7 18.6 17.7 26.2 16.5 21.7 17.5 Other (residual) 5.8 35.0 8.8 3.4 22.4 6.0 24.6 23.1 Non-NBR tax revenue 63.7 32.8 15.9 14.5 7.7 15.7 15.4 12.1 Non-tax revenue 84.7 -7.5 5.2 16.1 19.1 12.5 24.4 0.5 Note: B = Budget, R = Revised Budget, P = Provisional Source: FY00-FY06 National Board of Revenue FY91-FY99: National Board of Revenue Annual Report 2002-03, pp. 49 Budget and revised budget figures from Annual Financial Statement, Ministry of Finance GDP figures are from Bangladesh Bureau of Statistics Residual include excise duty, domestic supplementary duty, turnover tax, travel tax and others 172 Table 6: Bangladesh - Key Economic Indicators FY02 FY03 FY04 FY05 FY06 FY07 FY08 National Accounts (as % of GDP) Gross Domestic Producta 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Agriculture 21.9 21.0 20.2 19.3 18.9 18.6 18.3 Industry 25.4 25.3 25.5 26.1 26.9 27.4 27.5 Services 49.0 50.1 50.2 50.5 50.6 50.6 50.6 Total Consumption 81.8 81.4 80.5 80.0 79.8 79.6 79.7 Gross Domestic Fixed Investment 23.1 23.4 24.0 24.5 24.7 24.5 24.2 Government Investment 6.4 6.2 6.2 6.2 6.0 5.4 5.0 Private Investment 16.8 17.2 17.8 18.3 18.7 19.0 19.2 Export (GNFS)b 14.3 14.2 14.9 16.2 19.0 19.8 20.0 Imports (GNFS) 19.0 19.8 20.6 23.1 25.3 26.7 29.0 Gross Domestic Savings 18.2 18.6 19.5 20.0 20.2 20.4 20.1 Gross National Savings 23.4 24.9 25.4 25.8 27.7 28.7 29.2 Memorandum Items GDP at Current. Prices (US$ million) 47,663 51,914 56,561 60,278 61,881 68,415 79,003 GDP per capita (US$, Atlas Method) 372.0 389.0 418.3 447.7 469.0 493.7 532.1 Real Annual Growth Rates (%) GDP at constant market prices 4.4 5.3 6.3 6.0 6.6 6.4 6.2 GNI at constant market prices 5.5 6.2 6.0 5.8 8.1 7.3 6.6 Balance of Payments (US$ millions) Exports (GNFS) 6,794.0 7,379.0 8,445.0 9,750.0 11,752.0 13,537.0 15,824.0 Merchandise f.o.b. 5,929.0 6,492.0 7,521.0 8,573.0 10,412.0 12,053.0 13,945.0 Imports (GNFS) 9,061.0 10,285.0 11,638.0 13,917.0 15,664.0 18,250.0 22,890.0 Merchandise c.i.f. 7,697.0 8,707.0 9,840.0 11,870.0 13,301.0 15,511.0 19,486.0 Resource Balance -2,267.0 -2,906.0 -3,193.0 -4,167.0 -3,912.0 -4,713.0 -7,066.0 Net Current Transfers 2,826.0 3,440.0 3,743.0 4,290.0 5,438.0 6,554.0 8,743.0 Current Account Balance 157.0 176.0 176.0 -557.0 824.0 936.0 672.0 Foreign Direct Investment 391.0 376.0 276.0 800.0 743.0 793.0 650.0 Net Aid Flow d 298.0 466.0 147.0 491.0 535.0 512.0 758.0 Aid Disbursement 733.0 918.0 544.0 940.0 1,023.0 1,037.0 1,338.0 Debt Amortization -435.0 -452.0 -397.0 -449.0 -488.0 -525.0 -580.0 Other Capital e -277.0 -482.0 -433.0 -261.0 -1,649.0 -523.0 -1,724.0 Change in Reserves -408.0 -815.0 -171.0 -67.0 -338.0 -1,493.0 605.0 Memorandum Items: Resource Balance (% of GDP) -4.8 -5.6 -5.6 -6.9 -6.3 -6.9 -8.9 Real Annual Growth Rates Merchandise Exports (f.o.b.) -7.6 9.5 15.9 14.0 21.5 15.8 15.7 Merchandise Imports (c.i.f.) -8.7 13.1 13.0 20.6 12.1 16.6 25.6 Indicator FY02 FY03 FY04 FY05 FY06 FY07 FY08(P) Public Finance ( % of GDP at mkt prices) Total Revenues 10.1 10.3 10.2 10.5 10.7 10.4 11.4 Current Expenditure 8.0 8.1 7.8 8.4 8.4 9.0 9.6 Overall Budget Deficit 4.6 3.4 2.9 3.5 3.4 3.1 3.7 Capital Expenditure 5.6 5.4 5.0 5.0 4.7 3.8 3.4 Net Foreign Financing 2.1 2.1 1.1 1.8 1.2 1.2 2.2 Monetary Indicators M2/GDP 36.1 37.9 38.9 40.8 43.4 44.9 46.0 Growth of M2 (year on year % change) 13.1 15.6 13.8 16.7 19.5 17.0 17.6 Private Sector Credit Growth (year on year % change) 13.9 12.6 17.5 17.0 18.3 15.1 25.2 Total Domestic Credit Growth (year on year % change) 12.9 9.5 20.7 17.5 20.5 14.9 21.2 Prices Indices (FY96 = 100) Merchandise Export Price Index 123.2 126.2 135.2 139.6 142.4 149.3 154.5 Merchandise Import Price Index 146.4 157.8 164.2 170.0 176.7 183.1 189.4 Merchandise Terms of Trade Index 84.1 80.0 82.4 82.1 80.6 81.5 81.6 Nominal Exchange Rate (US$/LCU) f 57.4 57.9 58.9 61.5 67.2 69.1 68.6 Consumer Price Index (% change) 2.8 4.4 5.8 6.5 7.2 7.2 9.9 GDP Deflator (% change) 3.1 4.4 4.1 5.1 5.2 6.8 8.1 Source: Bangladesh Bureau of Statistics, Bangladesh Bank and IMF a/ GDP at current market price includes agriculture, industry, service & import duty b/ "GNFS" denotes "goods and non-factor services" c/Loans only. For FY05 includes flood-related donor assistance of US$123 million, plus World Bank loans (DSC II and education) totaling another US$300 million that were originally planned for FY04. d/ excluding transfers e/ Other short-term loans + other assets + trade credit in the BOP f/ Nominal period average exchange rate "LCU" denotes "local currency units". An increase in US$/LCU denotes appreciation 173 Table 7: Bangladesh Social Indicators Latest Single Year Same Region/Income Group a 1991-92 1995-96 1999-2000 2002-03 2003-04 2004-05 South Asia Low-Income Population Total population end of the year (millions) 113.3 122.1 129.8 134.1 136.1 138.2 1425.0 2312.0 Growth rate (% annual average for period) 2.1 1.8 1.4 1.4 1.5 1.5 1.8 1.9 Urban population (% of population) 18.0 19.0 25.0 27.0 … 28.0 30.0 Total fertility rate (births per woman) 4.2 3.4 3.1 2.9 3.0 3.1 3.7 Poverty (Lower poverty line) (% of population) National Headcount Index 42.7 34.4 33.7 … 25.1 … … Urban Headcount Index 23.3 13.7 19.1 … 14.6 … … Rural Headcount Index 46.0 38.5 37.4 … 28.6 … … Income GNI per capita (US$)at constant prices 283.0 343.5 381.0 400.0 420.0 440.0 510.0 440.0 Consumer Price Index (1995/96=100) 100.0 124.3 136.0 143.9 153.2 136.0 142.0 Food Price Index (1995/96=100) 100.0 128.5 137.0 146.5 158.1 … … Income/Consumption Distribution Gini Co-efficient 0.4 0.4 0.5 … 0.5 … … Lowest Quintile (% of income or consumption) 6.5 5.7 6.2 … 5.3 … … Highest Quintile (% of income or consumption) 44.8 50.1 52.0 … 52.7 … … Public Expenditures Health (% of GDP) 0.6 0.7 1.0 0.8 0.8 0.8 0.9 1.2 Education (% of GDP) 1.7 2.1 2.2 2.1 1.9 1.8 3.1 3.3 Social Security and Welfare (% of GDP) … … 1.0 0.5 0.6 0.7 … … Gross Primary School Enrolment Rate (% of age group) Total 76.0 95.0 96.6 97.3 97.0 97.0 95.0 Male … 97.0 97.0 97.0 97.0 108.0 103.0 Female 70.0 93.0 97.0 98.0 98.0 89.0 87.0 Access to an Improved Water Source (% of population) Total … … 97.0 97.0 97.0 84.0 76.0 Urban … … … … … 92.0 88.0 Rural … … … … … 85.0 70.0 Immunization Rate (% under 12 months) b Measles 68.0 69.0 71.0 76.0 75.7 77.0 63.0 64.0 DPT … 77.0 … 83.0 87.0 75.0 70.0 Child Malnutrition (% of under 5 years) 73.0 66.0 49.1 48.0 47.5 47.0 … Life Expectancy at Birth (years) Total 56.3 58.9 61.5 62.1 65.1 63.0 59.0 Male 55.0 57.0 61.0 … 64.4 62.0 58.0 Female 56.0 59.0 62.0 … 65.7 63.0 60.0 Mortality Infant (per thousand live births) 88.0 77.0 … 66.7 65.0 56.0 74.0 77.0 Under 5 (per thousand live births) … 112.0 … 84.6 88.0 77.0 99.0 116.0 Adult (15-69) Male (per 1000 population) … … 278.0 262.0 … … Female (per 1000 population) … … 272.0 252.0 … … Maternal (per 100,000 live births) 4.7 4.4 … 3.2 3.2 300.0 … … Births attended by skilled health staff (%) 7.0 … 14.0 11.6 13.0 13.0 … … Source: Unlocking the Potential-Poverty Reduction Strategy Paper (GOB) ,Bangladesh Bureau of Statistics and World Bank a/ Some data correspond to 2000-01 and 2001-02 b/ The immunization rates in 2003/04 are for children 12-23 months old. 174 Table 8: Bangladesh: Key Exposure Indicators (in US$ million unless otherwise stated) Actual Estimate Projection 2002 2003 2004 2005 2006 2007 Total debt outstanding and 17,061.0 18,778.0 19,321.0 20,364.0 21,379.0 22,354.0 disbursed (TDO) Net disbursements 217.0 337.0 543.0 1,010.0 1,015.0 975.0 Total debt service (TDS)a 727.0 672.0 958.0 1,018.0 1,039.0 1,107.0 Debt and debt service indicators (%) TDO/XGSb 182.6 178.8 162.6 160.9 161.1 158.9 TDO/GDP 35.9 36.2 34.0 33.5 34.0 33.3 TDS/XGS 7.8 6.4 8.1 8.0 7.8 7.9 Concessional/TDO 92.8 93.2 … … … … IBRD exposure indicators (%) IBRD DS/Public DS 0.9 1.1 0.8 0.0 0.0 0.0 Preferred creditor DS/Public 55.2 63.9 39.5 41.7 45.0 48.6 DSc IBRD DS/XGS 0.1 0.1 0.1 0.0 0.0 0.0 IBRD TDOd (US$mill) 13.0 7.0 0.0 0.0 0.0 0.0 Share of IBRD portfolio 0.0 0.0 0.0 0.0 0.0 0.0 IDA TDOd (US$mill) 7,063.0 8,062.0 8,208.0 8,805.0 9,480.0 10,099.0 IFC Loans 88.0 94.0 116.0 98.0 … … Equity and quasi-equitye 13.0 13.0 12.0 12.0 … … MIGA MIGA guarantees 64.0 61.0 46.0 … … … a/ Includes public and publicly-guaranteed debt, private non-guaranteed, use of IMF credits and net short-term capital b/ "XGS" denotes exports of goods and services, including workers' remittances c/ Preferred creditors are defined as IBRD, IDA, the regional multilateral development banks, the IMF, and the Bank for International Settlements d/ Includes present value of guarantees e/ Includes equity and quasi-equity types of both loan and equity instruments 175 Table 9: Bangladesh - Progress Towards MDGs a d 1980 1990 2000 2002 2004-05 2015 PRSP MDG b target target c Poverty Headcount Rate … 58.8 49.8 … 40.8 25.0 29.4 Fertility Rate (children per woman) 5.0 4.3 3.0 3.0 3.0 Infant Mortality (per 1,000 live births) 101.4 94.0 66.3 53.0 56.0 18.0 31.0 Crude Birth Rate (per 1,000 population) 33.4 32.8 19.9 20.1 20.8 Crude Death Rate (per 1,000 population) 10.2 11.3 4.8 5.1 5.7 Life Expectancy (years) 56.9 56.0 60.6 62.0 63.5 73.0 73.0 Gross Primary Enrollment (%) 61.0 72.0 96.6 86.7 97.0 100.0 100.0 Gross Secondary Enrollment (%) 18.0 19.0 42.0 52.8 44.0 95.0 Adult Illiteracy (%) 71.0 65.0 55.0 50.4 41.1 10.0 Source: Unlocking the Potential-Poverty Reduction Strategy Paper (GOB) ,Bangladesh Bureau of Statistics a/ Some data are for 1999 b/ Based on the progress rate of 1990-02 c/ The poverty headcount rate refers to the upper poverty line d/ 2004 Crude Birth rate and Crude death rate figures 176 Table 9: Basic information on the formal education system in Bangladesh Level/ qualification Duration Indicative age Main providers and other information (yrs) group Pre-primary 3 5-Mar Mostly private and NGO provision. Primary (Ebtadayee) 5 10-Jun Government provides the majority of provision through its own schools and madrasahs as well as through subsidised registered non- government schools and madrasahs. Unregistered madrasahs and NGO schools also provide school places at this level. Secondary 5 15-Nov Largely provided through government subsidised secondary schools and madrasahs with some purely private provision. At the end of this cycle the Secondary School Certificate (SSC) is taken. SSC is prepared by 6 divisional education boards and the madrasah (Dakhil) education board. Higher secondary 2 16-17 Largely provided through government subsidised secondary schools and colleges, madrasahs and some purely private provision. At the (Alim) end of this cycle the Higher Secondary Certificate (HSC) is taken. SSC is prepared by 6 divisional education boards and the madrasah education board. Bachelor degree courses 4-Mar 18-21 Provided by government and non-government colleges, government and private universities and Fazil and Alim government and non- (Fazil/ Kamil) government madrasahs. Postgraduate study 5-Jan >21 Mostly undertaken in government and private universities although some 89 masters degree colleges (of which 61 are government run) also offer postgraduate study. Technical, vocational and professional education Basic 2 14-15 85% of institutions offering these courses in 2005 were run by the private sector although many receive government support. Students must have completed Class VIII of secondary. SSC (Voc) 5 15-Nov 90% of institutions offering these courses in 2005 were run by the private sector although many receive government support. HSC (Voc) 2 16-17 94% of institutions offering these courses in 2005 were run by the private sector although many receive government support. HSC (BM) Diploma 4 16-19 72% of institutions offering these courses in 2005 were run by the private sector although many receive government support. Generally of 4 years duration after SSC (e.g. diploma in engineering, textiles, commerce etc.). Certificate in education 1 16 Primary teacher training institutes (PTIs) run almost exclusively by government. Main primary teacher training certificate in education undertaken after SSC. Other teacher training 4-Mar 18+ Approximately 80% of institutions offering post-primary teaching courses in 2005 were run by the private sector. Diploma and B. ed for post-primary teachers. Notes: Information contained in this table provides basic information on the education system. For additional information on the relative sizes of the different providers and government financing of different providers see the main text and in particular Table 3 and Table 6. Names in parentheses correspond to the madrasah system equivalent of general education levels. Source: BANBEIS (2006) and World Bank (2006a). 177 Table 2: Primary and Secondary school enrolment rates , 2000-2005 Net enrolment rate Gross enrolment rate Male Female Total Male Female Total 2000 2005 2000 2005 2000 2005 2000 2005 2000 2005 2000 2005 Primary quintiles 1 52 55 53 62 53 58 71 71 78 83 75 77 2 62 64 65 65 63 65 89 91 90 91 90 91 3 66 71 73 73 69 72 94 98 100 98 97 98 4 70 74 78 77 74 76 98 101 105 102 101 101 5 81 81 75 79 78 80 109 102 99 96 104 99 poor 58 59 61 63 59 61 81 80 86 87 84 83 non-poor 73 75 75 76 74 75 101 100 103 99 102 100 rural 63 66 67 69 65 67 89 90 93 93 91 91 urban 67 71 68 73 67 72 94 92 93 93 93 93 total 64 67 67 70 65 68 89 90 93 93 91 92 Secondary quintiles 1 11 17 20 30 15 24 15 25 24 35 20 30 2 19 27 32 39 25 33 27 36 40 48 33 41 3 28 37 40 49 34 43 41 54 53 67 46 60 4 43 52 60 64 51 58 65 74 78 83 72 78 5 59 65 65 71 62 68 89 99 88 96 89 98 poor 18 22 28 34 23 28 25 31 36 41 30 36 non-poor 45 51 58 62 52 56 69 74 77 82 73 78 rural 30 38 43 49 36 43 45 55 56 63 50 59 urban 40 47 48 56 44 52 57 68 63 73 60 71 total 32 40 44 51 38 45 47 58 57 66 52 62 Source: HIES (2000 and 2005) 178 Table : Higher secondary and tertiary enrolment rates , 2000-2005 Net enrolment rate Gross enrolment rate Male Female Total Male Female Total 2000 2005 2000 2005 2000 2005 2000 2005 2000 2005 2000 2005 Higher secondary quintiles 1 3 2 0 3 2 2 9 9 0 7 6 8 2 5 3 4 3 4 3 16 15 12 13 15 14 3 4 4 6 5 5 4 24 24 17 15 21 20 4 14 11 12 6 13 9 75 45 54 37 67 42 5 20 18 24 23 22 20 119 101 100 95 111 98 poor 4 2 2 3 3 3 15 12 8 10 13 11 non-poor 15 11 17 12 16 12 85 56 70 53 79 55 rural 10 6 11 6 10 6 48 32 38 24 45 29 urban 15 15 15 18 15 16 90 73 75 75 83 74 total 11 8 12 9 11 9 57 42 48 39 54 41 Tertiary quintiles 1 2 1 0 0 1 0 2 1 0 0 1 0 2 1 2 0 0 0 1 1 2 0 0 0 1 3 3 3 1 1 2 2 3 3 1 1 2 2 4 7 6 3 3 5 5 9 7 3 3 6 5 5 20 17 12 12 18 15 23 18 13 13 18 16 poor 2 2 0 0 1 1 2 2 0 0 1 1 non-poor 12 9 6 6 9 7 14 10 7 6 10 8 rural 6 5 2 2 4 3 7 5 3 2 5 4 urban 14 12 8 8 11 10 17 14 9 8 13 11 total 8 7 4 4 6 5 10 7 4 4 7 6 Source: HIES (2000 and 2005) 179 Table 4: School age population shares across quintiles and poor/non-poor groupings Age quintiles of household per capita expenditure group 1 2 3 4 5 poor non-poor School age population Primary 6-10 27 23 20 17 14 50 50 Secondary 11-15 19 20 21 21 19 39 61 Higher secondary 16-17 14 17 21 23 24 31 69 Tertiary 18-25 15 18 20 23 24 33 67 Total 6-25 20 20 20 20 20 40 60 Students Primary - 23 23 21 18 15 59 41 Secondary - 10 13 20 26 31 55 45 Higher secondary - 3 6 11 24 57 58 42 Tertiary - 1 3 7 21 67 56 44 Total - 16 18 20 21 25 34 66 Per capita expenditure - 8 11 15 21 45 19 81 Household expenditure - 9 12 16 21 42 21 79 Source: (Al-Samarrai 2007b) 180